From ef601829dfc989e060edc0fe140061ea89220d5f Mon Sep 17 00:00:00 2001 From: ledger Date: Wed, 28 Oct 2026 03:22:28 +0000 Subject: [PATCH] ledger: arXiv cs.AI most-recent 20 submissions (2026-10-28) extraction table + reverify gate --- .../ledger/2026-10-28-arxiv-csai-recent/CHECK | 7 + .../2026-10-28-arxiv-csai-recent/README.md | 22 + .../arxiv_csai_recent.json | 222 +++ .../arxiv_csai_recent.md | 26 + .../arxiv_csai_recent.tsv | 21 + .../2026-10-28-arxiv-csai-recent/raw_api.xml | 1605 +++++++++++++++++ .../2026-10-28-arxiv-csai-recent/reverify.py | 153 ++ 7 files changed, 2056 insertions(+) create mode 100755 wizard-work/ledger/2026-10-28-arxiv-csai-recent/CHECK create mode 100644 wizard-work/ledger/2026-10-28-arxiv-csai-recent/README.md create mode 100644 wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.json create mode 100644 wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.md create mode 100644 wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.tsv create mode 100644 wizard-work/ledger/2026-10-28-arxiv-csai-recent/raw_api.xml create mode 100644 wizard-work/ledger/2026-10-28-arxiv-csai-recent/reverify.py diff --git a/wizard-work/ledger/2026-10-28-arxiv-csai-recent/CHECK b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/CHECK new file mode 100755 index 0000000..8e67dc9 --- /dev/null +++ b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/CHECK @@ -0,0 +1,7 @@ +#!/usr/bin/env bash +# CHECK -- arXiv cs.AI most-recent table (20 rows), pinned to raw_api.xml. +# Re-derives the TSV/JSON/MD tables from raw_api.xml and diffs against the +# committed files. Exit 0 only on an exact match. +set -u +cd "$(dirname "${BASH_SOURCE[0]}")" || exit 1 +exec python3 reverify.py verify diff --git a/wizard-work/ledger/2026-10-28-arxiv-csai-recent/README.md b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/README.md new file mode 100644 index 0000000..ee87c1b --- /dev/null +++ b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/README.md @@ -0,0 +1,22 @@ +# arXiv cs.AI — most recent 20 submissions (2026-10-28) + +Extraction + re-verification gate. 20 latest arXiv `cs.AI` submissions fetched +from the arXiv API on 2026-10-28, re-derived and diffed byte-for-byte from the +pinned `raw_api.xml` snapshot. + +## Files + +| File | Contents | +| --- | --- | +| `raw_api.xml` | pinned arXiv API Atom feed (source of truth; `updated` = 2026-10-28T02:33:55Z) | +| `arxiv_csai_recent.tsv` | tab-separated table (rank, arxiv_id, title, authors, published, updated, categories, primary_category, abstract) | +| `arxiv_csai_recent.json` | same rows as JSON | +| `arxiv_csai_recent.md` | markdown table (20 data rows) | +| `reverify.py` | emit/verify gate: `python3 reverify.py verify` exits 0 only on exact match | +| `CHECK` | verdict line: run `python3 reverify.py verify` | + +## Evidence + +- Fetch command: `curl -s "https://export.arxiv.org/api/query?search_query=cat:cs.AI&sortBy=submittedDate&sortOrder=descending&max_results=20"` +- API `` / fetch timestamp: `2026-10-28T02:33:55Z` +- `` values are pinned via `raw_api.xml`; no network fetch inside the gate. diff --git a/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.json b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.json new file mode 100644 index 0000000..ff05824 --- /dev/null +++ b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.json @@ -0,0 +1,222 @@ +[ + { + "arxiv_id": "2610.37668", + "title": "Honeycomb-Surfels: Compact Hexagonal Primitives for Efficient Scene Rendering and Reconstruction", + "authors": "Prakhar Kaushik; Soumava Paul; Alan Yuille", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.CV; cs.AI; cs.GR; cs.LG", + "primary_category": "cs.CV", + "abstract": "Novel view synthesis and 3D reconstruction have advanced rapidly with radiance fields and splatting, yet their memory and computational demands limit their use in resource-constrained settings. We introduce Hexels: novel planar hexagonal surface elements with a first-order spatial color field for ef", + "rank": 1 + }, + { + "arxiv_id": "2610.37666", + "title": "ThinkWorld: World Imagination Reinforces Chain-of-Dynamics Reasoning", + "authors": "Chaehyun Kim; Chi-Pin Huang; Jisoo Kim; Min-Hung Chen; Yu-Chiang Frank Wang; Seungryong Kim; Fu-En Yang", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.RO; cs.AI; cs.CV", + "primary_category": "cs.RO", + "abstract": "Vision-Language-Action (VLA) models have made substantial progress in robotic manipulation, but their generalization remains constrained by the diversity and coverage of action-labeled robot demonstrations. Human videos provide much broader behavioral and environmental experience, while learning fro", + "rank": 2 + }, + { + "arxiv_id": "2610.37656", + "title": "Attention as Lattice Gauge Theory: Relations Interact Only Through Wilson Loops", + "authors": "Aakash Sarkar; Ananya Mukherjee; Motoshige Sato", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.LG; cs.AI; physics.comp-ph", + "primary_category": "cs.LG", + "abstract": "Relations compose through the entity they share - Alice is Bob's mother and Bob is Carol's father, so Alice is Carol's grandmother - and the order matters, since a father's mother is not a mother's father; verifying a composed relation means closing the loop. We show that if each entity's coordinate", + "rank": 3 + }, + { + "arxiv_id": "2610.37655", + "title": "KPPax: preprocess differentiable kinetic ODE for machine learning in atmospheric chemistry", + "authors": "Wenqing Peng; Michael Boy; Zhi-Song Liu", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.AI", + "primary_category": "cs.AI", + "abstract": "Atmospheric chemical mechanisms are commonly distributed in the language of the Kinetic PreProcessor (KPP), which generates efficient Fortran or C code but does not fit into the differentiable programming frameworks in which machine-learning models are developed. We present KPPax, a kinetic preproce", + "rank": 4 + }, + { + "arxiv_id": "2610.37649", + "title": "Can LLMs Uncover Hidden Physical Laws through Free Exploration?", + "authors": "Zhen Yang; Wen Wang; Xander Xu; Ying-Cong Chen", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.CL; cs.AI; cs.MA", + "primary_category": "cs.CL", + "abstract": "Scientific discovery often begins with an unexpected observation that prompts a new research question. In a world with altered physical laws, can large language models (LLMs) notice unexpected behavior, such as heavier objects falling more slowly, and discover the law through experiments without bei", + "rank": 5 + }, + { + "arxiv_id": "2610.37647", + "title": "MuonGS: Pretraining LLMs with Relaxed Weight-Norm Control", + "authors": "Athanasios Glentis; Dawei Li; Mingyi Hong", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.LG; cs.AI; math.OC", + "primary_category": "cs.LG", + "abstract": "Weight decay and explicit weight-norm constraints improve optimizer performance in language model pretraining, yet a unified understanding of why such control helps and how much is needed remains incomplete. We analyze norm control in Muon over the entire training trajectory using the \\emph{accumula", + "rank": 6 + }, + { + "arxiv_id": "2610.37631", + "title": "What Improves During Pretraining? Tracking Knowledge Acquisition with Point-in-Time Models", + "authors": "Chang Su; Bin Wu; Emine Yilmaz; Yao Lu", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.AI", + "primary_category": "cs.AI", + "abstract": "Language models pretrained on web corpora collected over multiple years learn both general and domain-specific knowledge. How these different kinds of knowledge are learned as pretraining progresses remains unclear. Recent point-in-time models trained on web snapshots from different years provide a ", + "rank": 7 + }, + { + "arxiv_id": "2610.37624", + "title": "Rare, Relevant, and Still Wrong: How Memory Fails LLM Group Decisions", + "authors": "Anouar Imel; Imad Charradi; Abderrahman Sghir; Khalil El Khammar", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.AI; cs.CL; cs.MA; cs.SI", + "primary_category": "cs.AI", + "abstract": "When an LLM group's discussion is compressed into a short memory for a final decision maker, what should the memory keep? Two natural heuristics keep the rarest claims or the claims relevant to the decision criteria. On hidden-profile tasks both look attractive, because how widely a fact is held is ", + "rank": 8 + }, + { + "arxiv_id": "2610.37613", + "title": "VideoLVR: Latent Visual Reasoning over Grounded Video Evidence", + "authors": "Shilong Li; Xingyuan Bu; Hao Shi; Yun Xing; Weixin Feng; Qiangpeng Yang; Shilei Wen", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.CV; cs.AI", + "primary_category": "cs.CV", + "abstract": "Chain-of-thought reasoning has substantially advanced video understanding in multimodal large language models (MLLMs). However, language-based reasoning traces do not provide a reliable mechanism for models to revisit relevant visual evidence, leading to visual forgetting and impairing their ability", + "rank": 9 + }, + { + "arxiv_id": "2610.37605", + "title": "Beyond Answer Changes: Matched Same-Class Controls for Local Audio Evidence", + "authors": "Yunuo Wang; Maoyuan Chen; Yuanjin Huang; Xing Liu", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.SD; cs.AI", + "primary_category": "cs.SD", + "abstract": "For audio-language models, an answer change after a local sound-event edit can reflect a response to the event's class, sensitivity to waveform rewriting, or both. We compare matched same-class and different-class replacements of a queried event. Both use the same interval, level estimate, donor fit", + "rank": 10 + }, + { + "arxiv_id": "2610.37599", + "title": "MoI-Bench: A Benchmark for Image Generation and Editing on Mobile Devices", + "authors": "Kailai Feng; Haoxiang Rao; Yuxiang Wei; Bo Chen; Junsong Li; Yang Pan; Chengkun Cai; Jagmohan Chauhan", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.CV; cs.AI", + "primary_category": "cs.CV", + "abstract": "Developing on-device image models requires considering practical user needs, task types and device resource constraints. However, existing benchmarks provide limited guidance on identifying capability gaps between on-device models and server-side models. We introduce MoI-Bench, a benchmark grounded ", + "rank": 11 + }, + { + "arxiv_id": "2610.37598", + "title": "GAS-PFN: Graph-Aware Spatiotemporal Prior-Data Fitted Networks for Zero-Shot Traffic Forecasting", + "authors": "Joseph Natter; Long Vu; Yifan Zhang; Bobin Deng; Kun Suo; Rui Wu", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.LG; cs.AI; cs.NE", + "primary_category": "cs.LG", + "abstract": "Traffic forecasting requires modeling complex temporal dynamics and the spatial dependencies imposed by road topology. Supervised graph-based models achieve strong performance, but are often trained for a particular dataset and sensor graph, making deployment to new or evolving networks costly. Prio", + "rank": 12 + }, + { + "arxiv_id": "2610.37588", + "title": "Understanding the limits of latent world models", + "authors": "Rodrigo Gonzalez Laiz; Steffen Schneider", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.LG; cs.AI", + "primary_category": "cs.LG", + "abstract": "World models are widely proposed as core components for enabling agents to reason over long temporal and spatial horizons. Yet current evaluations often emphasize task performance while providing limited insight into whether learned world models acquire predictive latent dynamics that reflect the un", + "rank": 13 + }, + { + "arxiv_id": "2610.37585", + "title": "Timestral", + "authors": "Adrien Petralia; Alexandre Gavaudan; Artjom Joosen; Etienne Goffinet; Richard Kurle; Yassine El Ouahidi; Yassir Bendou; Abdelaziz Bounhar; Abhijeet Somani; Aditi Kabra; Adrian Valente; Adrien Sadé; Ahmed Djermani; Alan Jeffares; Albert Jiang; Alberto Bietti; Aleksandr Timashov; Alexandre Cahill; Alexandre Laval; Alexandre Sablayrolles; Alexandros Sanchez Vassopoulos; Amélie Héliou; Amos You; André Jonasson; Andreas Fürst; Andrew Bai; Andrew Ehrenberg; Andrew Zhao; Andy Lo; Angele Lenglemetz; Anmol Agarwal; Antonia Calvi; Arata Suzuki; Arjun Majumdar; Arthur Fournier; Ashwani Kumar; Augustin Lombard; Avinash Sooriyarachchi; Aylin Guliz Akkus; Aysenur Karaduman; Baptiste Bout; Baptiste Rozière; Barna Pásztor; Baudouin De Monicault; Benjamin Holzschuh; Benjamin Tibi; Benno Stein; Bernhard Stadlbauer; Bhargavi Bhaskar; Binbing Hou; Błażej Osiński; Boris van Breugel; Camille Le Scao; Chaoran Yu; Charlotte Cronjäger; Chen-Yo Sun; Chris Bamford; Christian Wallenwein; Christophe Renaudin; Clémence Lanfranchi; Colin Bonatti; Corentin Barreau; Corentin Sautier; Cristiana-Diana Diaconu; Cyprien Courtot; Daniel Lopes; Daniel Marczak; Darius Dabert; Diego de las Casas; Dominik Nuss; Dylan Rubini; Dzmitry Soupel; Elizaveta Demyanenko; Elliot Chane-Sane; Emilien Fugier; Emmanuel Gottlob; Erik Aas; Etienne Duchesne; Fabian Paischer; Fabian Schlager; Faruk Ahmed; Federico Baldassarre; Félix Rosseeuw; Filip Szatkowski; Filippo Vicentini; Florian Wiesner; Gabrielle Berrada; Gaëtan Ecrepont; Gaspard Blanchet; Gaspard Donada-Vidal; Gaurav Mahajan; Gauthier Guinet; Genevieve Hayes; Georgii Novikov; Georgios Christoglou; Giada Pistilli; Gonzague de Carpentier; Graeme Nail; Guillaume Breton; Guillaume Kunsch; Guillaume Lample; Guillaume Martin; Guillaume Raille; Guillaume Richard; Guillem Simeon; Gunjan Dhanuka; Gunshi Gupta; Hamza Senhaji Rhazi; Han Zhou; Hédi Hadiji; Himanshu Jaju; Hope McGovern; Hubert de Parseval; Hugo Cisneros; Hugo Germain; Hugo Thimonier; Hugo Venturini; Hyeonju Lee; Hyunwoo Oh; Iliass Lasri; Ilyas Oulkadda; Jacques Sun; Jan Ludziejewski; Jason Rute; Jean de Becdelièvre; Jean Quentin; Jean-Malo Delignon; Jessica Chudnovsky; Jianchao Li; Jie Zhang; Jin Bin Ou; Joachim Studnia; Joep Barmentlo; Johannes Brandstetter; John Harvill; Jonas Amar; Jonas Schweizer; Jonathan Roberts; Joséphine Delas; Josselin Somerville; Jules Depersin; Julien Denize; Julien Tauran; Julien Verlaguet; Kartik Khandelwal; Khyathi Raghavi Chandu; Kilian Tep; Konstantin Mishchenko; Kris Szczepaniak; Kush Jain; Larissa Laich; Laura Calem; Laurence Aitchison; Laurent Callot; Laurent Fainsin; Léo Cotteleer; Léonard Blier; Loïc Chadoutaud; Loïck Chambon; Lorenzo Proietti; Louis Faucon; Louis Martin; Louis Serrano; Lu Li; Luc Boudier; Lucas Fernandes Martins; Lucas Mebille; Lucile Saulnier; Ludovic Ho Fuh; Luis Montero; Łukasz Borchmann; Maarten Buyl; Malik Hassan Raza; Manon Chossegros; Marcin Możejko; Margaret Jennings; Markus Hennerbichler; Martin Alexandre; Mathieu Poirée; Mathieu Schmitt; Mathilde Guillaumin; Mathis Felardos; Matthieu André; Matthieu Dinot; Matthieu Futeral; Maurits Bleeker; Mauro Comi; Max Mynter; Maxim Berman; Maxime Darrin; Maxime Langelier; Maxime Louis; Maximilian Augustin; Maximilian Müller; Mehdi Bouaziz; Melina Jingting Laimon; Mert Unsal; Mia Chiquier; Michael Pilcer; Michał Pietruszka; Michał Zając; Mickaël Seznec; Mikhail Biriuchinskii; Minh-Quang Pham; Minwoo Kang; Mircea Lica; Morgane Rivière; Namit Katariya; Natasha Ørregaard Klingenbrunn; Nathan Bout; Nathan Grinsztajn; Nathan Simpson; Neeraj Aggarwal; Neil Ashton; Neil Mallinar; Nick Bührer; Nicolas Moreau; Nicolò Lucchesi; Ola Mysiak; Oliver Leicht; Olivier Bousquet; Olivier Koch; Parag Jain; Patricia Wang; Patrick Blies; Patrick von Platen; Patryk Saffer; Paul Jacob; Paul Teiletche; Paul Wambergue; Paula Kurylowicz; Pavan Kumar Reddy; Pavel Kuksa; Philippe Formont; Philippe Pinel; Philomène Chagniot; Pierre Glandon; Pierre Stock; Pierre-André Savalle; Piotr Miłoś; Poojan Shah; Prateek Gupta; Pravesh Agrawal; Pui Ki Chan; Quentin Desreumaux; Quentin Selle; Quentin Torroba; Quentin Wargnier; Quercus Hernandez; Ram Ramrakhya; Randall Isenhour; Ranjit Parva; Raphaël Balouka; Raphaël Rialland; Raul Perez Pelaez; Reinhard Sonnleitner; Remi Delacourt; Rishi Shah; Rob Romijnders; Rohan Shah; Rohin Arora; Romain Sauvestre; Roman Soletskyi; Ronan Riochet; Rosalie Millner; Rupert Menneer; Sagar Vaze; Samuel Barry; Samuel Belkadi; Samuel Humeau; Samuele Papa; Sanchit Gandhi; Sandeep Cranganore; Sandeep Subramanian; Sarthak Mittal; Saskia Adaime; Scott Matthews; Sean Cha; Sebastian Kaltenbach; Seungjae Jung; Shashwat Verma; Sherif Waly; Shrimai Prabhumoye; Siddhant Waghjale; Siddharth Gandhi; Simon Lepage; Simon Sorg; Soham Ghosh; Sophie Marbach; Srijan Mishra; Stanislas Lange; Steve Hong; Sumukh Aithal; Szymon Antoniak; Tarun Kumar Vangani; Teddy Turquet; Teven Le Scao; Théo Cachet; Thibaut Lavril; Thomas Chabal; Thomas Coste; Thomas Defard; Thomas Foubert; Thomas Loux; Thomas Robert; Thomas Wang; Tianyu Zhang; Tim Lawson; Timothée Lacroix; Timothy Wallis; Tobias Kronlachner; Tom Bewley; Tom Edwards; Tomas Hodan; Triantafyllos Afouras; Tuhin Das; Tyler Wang; Udit Saxena; Ulrick Blé; Umar Jamil; Umberto Tomasini; Valentin Macé; Van Phung; Venkata Reddy; Vibha Masti; Victor Jouault; Victor Letzelter; Victor Paltz; Victor Poucheret; Vincent Maladière; Vincent Pfister; Virgile Richard; Vladislav Bataev; Wassim Bouaziz; Wen Ding Li; William Flanagan; William Havard; William Marshall; Witold Jędrzejewski; Xinghui Li; Xingran Guo; Xinyu Yang; Xuanyu Zhang; Yann Dreze; Yann Viegas; Yannic Neuhaus; Yassine Turki; Yessin Moakher; Yi He; Yihan Wang; Yimu Pan; Yuying Zhu; Yves Martin des Taillades; Zaccharie Ramzi; Zhenlin Xu; Zsófia Csákány", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.AI", + "primary_category": "cs.AI", + "abstract": "We introduce Timestral, a family of four time series foundation models for probabilistic forecasting, with 5M, 28M, 123M, and 209M parameters. Timestral sets a new state of the art on GIFT-Eval, TIME, and fev-bench, where Timestral-209M and Timestral-123M take the top two places on every metric. We ", + "rank": 14 + }, + { + "arxiv_id": "2610.37584", + "title": "REGER: Modeling inter-item Relations with Semantic IDs for Generative Recommendation", + "authors": "Yuxi Huang; Silin Zhou; Chenhao Wang; Zihan Li; Gangming Guo; Pei Ke; Shen Gao; Shuo Shang", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.IR; cs.AI", + "primary_category": "cs.IR", + "abstract": "Personalized recommendation depends on learning semantic and collaborative relations among items from online content and user interactions. Generative recommendation (GR) offers a promising paradigm by autoregressively generating discrete semantic IDs (SIDs) of items. However, randomly initialized S", + "rank": 15 + }, + { + "arxiv_id": "2610.37571", + "title": "TeamGym: A Gym for Team Decision-Making by AI Agents with Distributed Information", + "authors": "Chanwoo Park; Jaywoong Jeong; Zachary Wojtowicz; Yubin Kim; Taekyun Lee; Kaiqing Zhang; Asuman Ozdaglar", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.AI", + "primary_category": "cs.AI", + "abstract": "Many real-world decisions require teams to integrate information distributed across participants. Yet AI agents are still primarily trained and benchmarked in isolation. We introduce TeamGym, an environment for evaluating and training language model-based agents on team decision-making problems cons", + "rank": 16 + }, + { + "arxiv_id": "2610.37567", + "title": "Gym-Anything-RT: Scaling Real-Time Environments for Vision-Language Agents", + "authors": "Pranjal Aggarwal; Sean Welleck", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.LG; cs.AI", + "primary_category": "cs.LG", + "abstract": "AI agents have made substantial progress in domains where they have a non-trivial amount of time to deliberate between actions, such as mathematics, software engineering, and performing tasks on desktop applications. It is less clear how such models fare in settings that demand real-time interaction", + "rank": 17 + }, + { + "arxiv_id": "2610.37561", + "title": "When uncertainty quantification fails: structural vulnerabilities and corrections for selective prediction under clinical class imbalance", + "authors": "Walter Endrizzi; Flavio Ragni; Raffaele Marchesi; Stefano Bovo; Monica Moroni; Giuseppe Jurman; Venet Osmani", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.LG; cs.AI", + "primary_category": "cs.LG", + "abstract": "Background. Selective prediction systems improve the safety of clinical artificial intelligence by deferring ambiguous cases to human experts while automating high-confidence decisions. However, this safety mechanism relies entirely on the premise that uncertainty scores accurately reflect clinical ", + "rank": 18 + }, + { + "arxiv_id": "2610.37550", + "title": "Kinematic Constraint-Referenced Evaluation for Articulated Object Manipulation", + "authors": "Baptiste Bonin; Caro Strickland; Audrey Durand", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.RO; cs.AI", + "primary_category": "cs.RO", + "abstract": "Manipulation policies are evaluated predominantly by binary task success, which cannot tell whether the commanded gesture is the one the object's kinematics admit. We introduce Constraint-Referenced Strategy (CRS), which projects the commanded end-effector twist onto the object's joint axis, separat", + "rank": 19 + }, + { + "arxiv_id": "2610.37547", + "title": "When, Where, and What to Correct: Human-Guided Adaptive Residual Correction for Robot Policies", + "authors": "Anukriti Singh; Maxime Caniot; Pratap Tokekar; Firas Abi-Farraj", + "published": "2026-10-27", + "updated": "2026-10-27", + "categories": "cs.RO; cs.AI", + "primary_category": "cs.RO", + "abstract": "Imitation policies that reliably execute most of a manipulation trajectory can still fail in localized regions requiring last-mile precision. Updating the full policy or exploring residual actions everywhere treats this local failure as a global learning problem. We introduce Adaptive Residual Corre", + "rank": 20 + } +] diff --git a/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.md b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.md new file mode 100644 index 0000000..b5e3613 --- /dev/null +++ b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.md @@ -0,0 +1,26 @@ +# arXiv cs.AI — most recent 20 submissions (fetched 2026-10-28) + +Source: `raw_api.xml` (API: cat:cs.AI, sortBy=submittedDate, sortOrder=descending, max_results=20) + +| Rank | arXiv | Title | Authors | Published | Updated | Categories | Primary | Abstract (first 50 chars) | +|---:|---|---|---|---|---|---|---|---| +| 1 | 2610.37668 | Honeycomb-Surfels: Compact Hexagonal Primitives for Efficient Scene Rendering and Reconstruction | Prakhar Kaushik; Soumava Paul; Alan Yuille | 2026-10-27 | 2026-10-27 | cs.CV; cs.AI; cs.GR; cs.LG | cs.CV | Novel view synthesis and 3D reconstruction have ad… | +| 2 | 2610.37666 | ThinkWorld: World Imagination Reinforces Chain-of-Dynamics Reasoning | Chaehyun Kim; Chi-Pin Huang; Jisoo Kim; Min-Hung Chen; Yu-Chiang Frank Wang; Seungryong Kim; Fu-En Yang | 2026-10-27 | 2026-10-27 | cs.RO; cs.AI; cs.CV | cs.RO | Vision-Language-Action (VLA) models have made subs… | +| 3 | 2610.37656 | Attention as Lattice Gauge Theory: Relations Interact Only Through Wilson Loops | Aakash Sarkar; Ananya Mukherjee; Motoshige Sato | 2026-10-27 | 2026-10-27 | cs.LG; cs.AI; physics.comp-ph | cs.LG | Relations compose through the entity they share - … | +| 4 | 2610.37655 | KPPax: preprocess differentiable kinetic ODE for machine learning in atmospheric chemistry | Wenqing Peng; Michael Boy; Zhi-Song Liu | 2026-10-27 | 2026-10-27 | cs.AI | cs.AI | Atmospheric chemical mechanisms are commonly distr… | +| 5 | 2610.37649 | Can LLMs Uncover Hidden Physical Laws through Free Exploration? | Zhen Yang; Wen Wang; Xander Xu; Ying-Cong Chen | 2026-10-27 | 2026-10-27 | cs.CL; cs.AI; cs.MA | cs.CL | Scientific discovery often begins with an unexpect… | +| 6 | 2610.37647 | MuonGS: Pretraining LLMs with Relaxed Weight-Norm Control | Athanasios Glentis; Dawei Li; Mingyi Hong | 2026-10-27 | 2026-10-27 | cs.LG; cs.AI; math.OC | cs.LG | Weight decay and explicit weight-norm constraints … | +| 7 | 2610.37631 | What Improves During Pretraining? Tracking Knowledge Acquisition with Point-in-Time Models | Chang Su; Bin Wu; Emine Yilmaz; Yao Lu | 2026-10-27 | 2026-10-27 | cs.AI | cs.AI | Language models pretrained on web corpora collecte… | +| 8 | 2610.37624 | Rare, Relevant, and Still Wrong: How Memory Fails LLM Group Decisions | Anouar Imel; Imad Charradi; Abderrahman Sghir; Khalil El Khammar | 2026-10-27 | 2026-10-27 | cs.AI; cs.CL; cs.MA; cs.SI | cs.AI | When an LLM group's discussion is compressed into … | +| 9 | 2610.37613 | VideoLVR: Latent Visual Reasoning over Grounded Video Evidence | Shilong Li; Xingyuan Bu; Hao Shi; Yun Xing; Weixin Feng; Qiangpeng Yang; Shilei Wen | 2026-10-27 | 2026-10-27 | cs.CV; cs.AI | cs.CV | Chain-of-thought reasoning has substantially advan… | +| 10 | 2610.37605 | Beyond Answer Changes: Matched Same-Class Controls for Local Audio Evidence | Yunuo Wang; Maoyuan Chen; Yuanjin Huang; Xing Liu | 2026-10-27 | 2026-10-27 | cs.SD; cs.AI | cs.SD | For audio-language models, an answer change after … | +| 11 | 2610.37599 | MoI-Bench: A Benchmark for Image Generation and Editing on Mobile Devices | Kailai Feng; Haoxiang Rao; Yuxiang Wei; Bo Chen; Junsong Li; Yang Pan; Chengkun Cai; Jagmohan Chauhan | 2026-10-27 | 2026-10-27 | cs.CV; cs.AI | cs.CV | Developing on-device image models requires conside… | +| 12 | 2610.37598 | GAS-PFN: Graph-Aware Spatiotemporal Prior-Data Fitted Networks for Zero-Shot Traffic Forecasting | Joseph Natter; Long Vu; Yifan Zhang; Bobin Deng; Kun Suo; Rui Wu | 2026-10-27 | 2026-10-27 | cs.LG; cs.AI; cs.NE | cs.LG | Traffic forecasting requires modeling complex temp… | +| 13 | 2610.37588 | Understanding the limits of latent world models | Rodrigo Gonzalez Laiz; Steffen Schneider | 2026-10-27 | 2026-10-27 | cs.LG; cs.AI | cs.LG | World models are widely proposed as core component… | +| 14 | 2610.37585 | Timestral | Adrien Petralia; Alexandre Gavaudan; Artjom Joosen; Etienne Goffinet; Richard Kurle; Yassine El Ouahidi; Yassir Bendou; Abdelaziz Bounhar; Abhijeet Somani; Aditi Kabra; Adrian Valente; Adrien Sadé; Ahmed Djermani; Alan Jeffares; Albert Jiang; Alberto Bietti; Aleksandr Timashov; Alexandre Cahill; Alexandre Laval; Alexandre Sablayrolles; Alexandros Sanchez Vassopoulos; Amélie Héliou; Amos You; André Jonasson; Andreas Fürst; Andrew Bai; Andrew Ehrenberg; Andrew Zhao; Andy Lo; Angele Lenglemetz; Anmol Agarwal; Antonia Calvi; Arata Suzuki; Arjun Majumdar; Arthur Fournier; Ashwani Kumar; Augustin Lombard; Avinash Sooriyarachchi; Aylin Guliz Akkus; Aysenur Karaduman; Baptiste Bout; Baptiste Rozière; Barna Pásztor; Baudouin De Monicault; Benjamin Holzschuh; Benjamin Tibi; Benno Stein; Bernhard Stadlbauer; Bhargavi Bhaskar; Binbing Hou; Błażej Osiński; Boris van Breugel; Camille Le Scao; Chaoran Yu; Charlotte Cronjäger; Chen-Yo Sun; Chris Bamford; Christian Wallenwein; Christophe Renaudin; Clémence Lanfranchi; Colin Bonatti; Corentin Barreau; Corentin Sautier; Cristiana-Diana Diaconu; Cyprien Courtot; Daniel Lopes; Daniel Marczak; Darius Dabert; Diego de las Casas; Dominik Nuss; Dylan Rubini; Dzmitry Soupel; Elizaveta Demyanenko; Elliot Chane-Sane; Emilien Fugier; Emmanuel Gottlob; Erik Aas; Etienne Duchesne; Fabian Paischer; Fabian Schlager; Faruk Ahmed; Federico Baldassarre; Félix Rosseeuw; Filip Szatkowski; Filippo Vicentini; Florian Wiesner; Gabrielle Berrada; Gaëtan Ecrepont; Gaspard Blanchet; Gaspard Donada-Vidal; Gaurav Mahajan; Gauthier Guinet; Genevieve Hayes; Georgii Novikov; Georgios Christoglou; Giada Pistilli; Gonzague de Carpentier; Graeme Nail; Guillaume Breton; Guillaume Kunsch; Guillaume Lample; Guillaume Martin; Guillaume Raille; Guillaume Richard; Guillem Simeon; Gunjan Dhanuka; Gunshi Gupta; Hamza Senhaji Rhazi; Han Zhou; Hédi Hadiji; Himanshu Jaju; Hope McGovern; Hubert de Parseval; Hugo Cisneros; Hugo Germain; Hugo Thimonier; Hugo Venturini; Hyeonju Lee; Hyunwoo Oh; Iliass Lasri; Ilyas Oulkadda; Jacques Sun; Jan Ludziejewski; Jason Rute; Jean de Becdelièvre; Jean Quentin; Jean-Malo Delignon; Jessica Chudnovsky; Jianchao Li; Jie Zhang; Jin Bin Ou; Joachim Studnia; Joep Barmentlo; Johannes Brandstetter; John Harvill; Jonas Amar; Jonas Schweizer; Jonathan Roberts; Joséphine Delas; Josselin Somerville; Jules Depersin; Julien Denize; Julien Tauran; Julien Verlaguet; Kartik Khandelwal; Khyathi Raghavi Chandu; Kilian Tep; Konstantin Mishchenko; Kris Szczepaniak; Kush Jain; Larissa Laich; Laura Calem; Laurence Aitchison; Laurent Callot; Laurent Fainsin; Léo Cotteleer; Léonard Blier; Loïc Chadoutaud; Loïck Chambon; Lorenzo Proietti; Louis Faucon; Louis Martin; Louis Serrano; Lu Li; Luc Boudier; Lucas Fernandes Martins; Lucas Mebille; Lucile Saulnier; Ludovic Ho Fuh; Luis Montero; Łukasz Borchmann; Maarten Buyl; Malik Hassan Raza; Manon Chossegros; Marcin Możejko; Margaret Jennings; Markus Hennerbichler; Martin Alexandre; Mathieu Poirée; Mathieu Schmitt; Mathilde Guillaumin; Mathis Felardos; Matthieu André; Matthieu Dinot; Matthieu Futeral; Maurits Bleeker; Mauro Comi; Max Mynter; Maxim Berman; Maxime Darrin; Maxime Langelier; Maxime Louis; Maximilian Augustin; Maximilian Müller; Mehdi Bouaziz; Melina Jingting Laimon; Mert Unsal; Mia Chiquier; Michael Pilcer; Michał Pietruszka; Michał Zając; Mickaël Seznec; Mikhail Biriuchinskii; Minh-Quang Pham; Minwoo Kang; Mircea Lica; Morgane Rivière; Namit Katariya; Natasha Ørregaard Klingenbrunn; Nathan Bout; Nathan Grinsztajn; Nathan Simpson; Neeraj Aggarwal; Neil Ashton; Neil Mallinar; Nick Bührer; Nicolas Moreau; Nicolò Lucchesi; Ola Mysiak; Oliver Leicht; Olivier Bousquet; Olivier Koch; Parag Jain; Patricia Wang; Patrick Blies; Patrick von Platen; Patryk Saffer; Paul Jacob; Paul Teiletche; Paul Wambergue; Paula Kurylowicz; Pavan Kumar Reddy; Pavel Kuksa; Philippe Formont; Philippe Pinel; Philomène Chagniot; Pierre Glandon; Pierre Stock; Pierre-André Savalle; Piotr Miłoś; Poojan Shah; Prateek Gupta; Pravesh Agrawal; Pui Ki Chan; Quentin Desreumaux; Quentin Selle; Quentin Torroba; Quentin Wargnier; Quercus Hernandez; Ram Ramrakhya; Randall Isenhour; Ranjit Parva; Raphaël Balouka; Raphaël Rialland; Raul Perez Pelaez; Reinhard Sonnleitner; Remi Delacourt; Rishi Shah; Rob Romijnders; Rohan Shah; Rohin Arora; Romain Sauvestre; Roman Soletskyi; Ronan Riochet; Rosalie Millner; Rupert Menneer; Sagar Vaze; Samuel Barry; Samuel Belkadi; Samuel Humeau; Samuele Papa; Sanchit Gandhi; Sandeep Cranganore; Sandeep Subramanian; Sarthak Mittal; Saskia Adaime; Scott Matthews; Sean Cha; Sebastian Kaltenbach; Seungjae Jung; Shashwat Verma; Sherif Waly; Shrimai Prabhumoye; Siddhant Waghjale; Siddharth Gandhi; Simon Lepage; Simon Sorg; Soham Ghosh; Sophie Marbach; Srijan Mishra; Stanislas Lange; Steve Hong; Sumukh Aithal; Szymon Antoniak; Tarun Kumar Vangani; Teddy Turquet; Teven Le Scao; Théo Cachet; Thibaut Lavril; Thomas Chabal; Thomas Coste; Thomas Defard; Thomas Foubert; Thomas Loux; Thomas Robert; Thomas Wang; Tianyu Zhang; Tim Lawson; Timothée Lacroix; Timothy Wallis; Tobias Kronlachner; Tom Bewley; Tom Edwards; Tomas Hodan; Triantafyllos Afouras; Tuhin Das; Tyler Wang; Udit Saxena; Ulrick Blé; Umar Jamil; Umberto Tomasini; Valentin Macé; Van Phung; Venkata Reddy; Vibha Masti; Victor Jouault; Victor Letzelter; Victor Paltz; Victor Poucheret; Vincent Maladière; Vincent Pfister; Virgile Richard; Vladislav Bataev; Wassim Bouaziz; Wen Ding Li; William Flanagan; William Havard; William Marshall; Witold Jędrzejewski; Xinghui Li; Xingran Guo; Xinyu Yang; Xuanyu Zhang; Yann Dreze; Yann Viegas; Yannic Neuhaus; Yassine Turki; Yessin Moakher; Yi He; Yihan Wang; Yimu Pan; Yuying Zhu; Yves Martin des Taillades; Zaccharie Ramzi; Zhenlin Xu; Zsófia Csákány | 2026-10-27 | 2026-10-27 | cs.AI | cs.AI | We introduce Timestral, a family of four time seri… | +| 15 | 2610.37584 | REGER: Modeling inter-item Relations with Semantic IDs for Generative Recommendation | Yuxi Huang; Silin Zhou; Chenhao Wang; Zihan Li; Gangming Guo; Pei Ke; Shen Gao; Shuo Shang | 2026-10-27 | 2026-10-27 | cs.IR; cs.AI | cs.IR | Personalized recommendation depends on learning se… | +| 16 | 2610.37571 | TeamGym: A Gym for Team Decision-Making by AI Agents with Distributed Information | Chanwoo Park; Jaywoong Jeong; Zachary Wojtowicz; Yubin Kim; Taekyun Lee; Kaiqing Zhang; Asuman Ozdaglar | 2026-10-27 | 2026-10-27 | cs.AI | cs.AI | Many real-world decisions require teams to integra… | +| 17 | 2610.37567 | Gym-Anything-RT: Scaling Real-Time Environments for Vision-Language Agents | Pranjal Aggarwal; Sean Welleck | 2026-10-27 | 2026-10-27 | cs.LG; cs.AI | cs.LG | AI agents have made substantial progress in domain… | +| 18 | 2610.37561 | When uncertainty quantification fails: structural vulnerabilities and corrections for selective prediction under clinical class imbalance | Walter Endrizzi; Flavio Ragni; Raffaele Marchesi; Stefano Bovo; Monica Moroni; Giuseppe Jurman; Venet Osmani | 2026-10-27 | 2026-10-27 | cs.LG; cs.AI | cs.LG | Background. Selective prediction systems improve t… | +| 19 | 2610.37550 | Kinematic Constraint-Referenced Evaluation for Articulated Object Manipulation | Baptiste Bonin; Caro Strickland; Audrey Durand | 2026-10-27 | 2026-10-27 | cs.RO; cs.AI | cs.RO | Manipulation policies are evaluated predominantly … | +| 20 | 2610.37547 | When, Where, and What to Correct: Human-Guided Adaptive Residual Correction for Robot Policies | Anukriti Singh; Maxime Caniot; Pratap Tokekar; Firas Abi-Farraj | 2026-10-27 | 2026-10-27 | cs.RO; cs.AI | cs.RO | Imitation policies that reliably execute most of a… | diff --git a/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.tsv b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.tsv new file mode 100644 index 0000000..95e3872 --- /dev/null +++ b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/arxiv_csai_recent.tsv @@ -0,0 +1,21 @@ +rank arxiv_id title authors published updated categories primary_category abstract +1 2610.37668 Honeycomb-Surfels: Compact Hexagonal Primitives for Efficient Scene Rendering and Reconstruction Prakhar Kaushik; Soumava Paul; Alan Yuille 2026-10-27 2026-10-27 cs.CV; cs.AI; cs.GR; cs.LG cs.CV Novel view synthesis and 3D reconstruction have advanced rapidly with radiance fields and splatting, yet their memory and computational demands limit their use in resource-constrained settings. We introduce Hexels: novel planar hexagonal surface elements with a first-order spatial color field for ef +2 2610.37666 ThinkWorld: World Imagination Reinforces Chain-of-Dynamics Reasoning Chaehyun Kim; Chi-Pin Huang; Jisoo Kim; Min-Hung Chen; Yu-Chiang Frank Wang; Seungryong Kim; Fu-En Yang 2026-10-27 2026-10-27 cs.RO; cs.AI; cs.CV cs.RO Vision-Language-Action (VLA) models have made substantial progress in robotic manipulation, but their generalization remains constrained by the diversity and coverage of action-labeled robot demonstrations. Human videos provide much broader behavioral and environmental experience, while learning fro +3 2610.37656 Attention as Lattice Gauge Theory: Relations Interact Only Through Wilson Loops Aakash Sarkar; Ananya Mukherjee; Motoshige Sato 2026-10-27 2026-10-27 cs.LG; cs.AI; physics.comp-ph cs.LG Relations compose through the entity they share - Alice is Bob's mother and Bob is Carol's father, so Alice is Carol's grandmother - and the order matters, since a father's mother is not a mother's father; verifying a composed relation means closing the loop. We show that if each entity's coordinate +4 2610.37655 KPPax: preprocess differentiable kinetic ODE for machine learning in atmospheric chemistry Wenqing Peng; Michael Boy; Zhi-Song Liu 2026-10-27 2026-10-27 cs.AI cs.AI Atmospheric chemical mechanisms are commonly distributed in the language of the Kinetic PreProcessor (KPP), which generates efficient Fortran or C code but does not fit into the differentiable programming frameworks in which machine-learning models are developed. We present KPPax, a kinetic preproce +5 2610.37649 Can LLMs Uncover Hidden Physical Laws through Free Exploration? Zhen Yang; Wen Wang; Xander Xu; Ying-Cong Chen 2026-10-27 2026-10-27 cs.CL; cs.AI; cs.MA cs.CL Scientific discovery often begins with an unexpected observation that prompts a new research question. In a world with altered physical laws, can large language models (LLMs) notice unexpected behavior, such as heavier objects falling more slowly, and discover the law through experiments without bei +6 2610.37647 MuonGS: Pretraining LLMs with Relaxed Weight-Norm Control Athanasios Glentis; Dawei Li; Mingyi Hong 2026-10-27 2026-10-27 cs.LG; cs.AI; math.OC cs.LG Weight decay and explicit weight-norm constraints improve optimizer performance in language model pretraining, yet a unified understanding of why such control helps and how much is needed remains incomplete. We analyze norm control in Muon over the entire training trajectory using the \emph{accumula +7 2610.37631 What Improves During Pretraining? Tracking Knowledge Acquisition with Point-in-Time Models Chang Su; Bin Wu; Emine Yilmaz; Yao Lu 2026-10-27 2026-10-27 cs.AI cs.AI Language models pretrained on web corpora collected over multiple years learn both general and domain-specific knowledge. How these different kinds of knowledge are learned as pretraining progresses remains unclear. Recent point-in-time models trained on web snapshots from different years provide a +8 2610.37624 Rare, Relevant, and Still Wrong: How Memory Fails LLM Group Decisions Anouar Imel; Imad Charradi; Abderrahman Sghir; Khalil El Khammar 2026-10-27 2026-10-27 cs.AI; cs.CL; cs.MA; cs.SI cs.AI When an LLM group's discussion is compressed into a short memory for a final decision maker, what should the memory keep? Two natural heuristics keep the rarest claims or the claims relevant to the decision criteria. On hidden-profile tasks both look attractive, because how widely a fact is held is +9 2610.37613 VideoLVR: Latent Visual Reasoning over Grounded Video Evidence Shilong Li; Xingyuan Bu; Hao Shi; Yun Xing; Weixin Feng; Qiangpeng Yang; Shilei Wen 2026-10-27 2026-10-27 cs.CV; cs.AI cs.CV Chain-of-thought reasoning has substantially advanced video understanding in multimodal large language models (MLLMs). However, language-based reasoning traces do not provide a reliable mechanism for models to revisit relevant visual evidence, leading to visual forgetting and impairing their ability +10 2610.37605 Beyond Answer Changes: Matched Same-Class Controls for Local Audio Evidence Yunuo Wang; Maoyuan Chen; Yuanjin Huang; Xing Liu 2026-10-27 2026-10-27 cs.SD; cs.AI cs.SD For audio-language models, an answer change after a local sound-event edit can reflect a response to the event's class, sensitivity to waveform rewriting, or both. We compare matched same-class and different-class replacements of a queried event. Both use the same interval, level estimate, donor fit +11 2610.37599 MoI-Bench: A Benchmark for Image Generation and Editing on Mobile Devices Kailai Feng; Haoxiang Rao; Yuxiang Wei; Bo Chen; Junsong Li; Yang Pan; Chengkun Cai; Jagmohan Chauhan 2026-10-27 2026-10-27 cs.CV; cs.AI cs.CV Developing on-device image models requires considering practical user needs, task types and device resource constraints. However, existing benchmarks provide limited guidance on identifying capability gaps between on-device models and server-side models. We introduce MoI-Bench, a benchmark grounded +12 2610.37598 GAS-PFN: Graph-Aware Spatiotemporal Prior-Data Fitted Networks for Zero-Shot Traffic Forecasting Joseph Natter; Long Vu; Yifan Zhang; Bobin Deng; Kun Suo; Rui Wu 2026-10-27 2026-10-27 cs.LG; cs.AI; cs.NE cs.LG Traffic forecasting requires modeling complex temporal dynamics and the spatial dependencies imposed by road topology. Supervised graph-based models achieve strong performance, but are often trained for a particular dataset and sensor graph, making deployment to new or evolving networks costly. Prio +13 2610.37588 Understanding the limits of latent world models Rodrigo Gonzalez Laiz; Steffen Schneider 2026-10-27 2026-10-27 cs.LG; cs.AI cs.LG World models are widely proposed as core components for enabling agents to reason over long temporal and spatial horizons. Yet current evaluations often emphasize task performance while providing limited insight into whether learned world models acquire predictive latent dynamics that reflect the un +14 2610.37585 Timestral Adrien Petralia; Alexandre Gavaudan; Artjom Joosen; Etienne Goffinet; Richard Kurle; Yassine El Ouahidi; Yassir Bendou; Abdelaziz Bounhar; Abhijeet Somani; Aditi Kabra; Adrian Valente; Adrien Sadé; Ahmed Djermani; Alan Jeffares; Albert Jiang; Alberto Bietti; Aleksandr Timashov; Alexandre Cahill; Alexandre Laval; Alexandre Sablayrolles; Alexandros Sanchez Vassopoulos; Amélie Héliou; Amos You; André Jonasson; Andreas Fürst; Andrew Bai; Andrew Ehrenberg; Andrew Zhao; Andy Lo; Angele Lenglemetz; Anmol Agarwal; Antonia Calvi; Arata Suzuki; Arjun Majumdar; Arthur Fournier; Ashwani Kumar; Augustin Lombard; Avinash Sooriyarachchi; Aylin Guliz Akkus; Aysenur Karaduman; Baptiste Bout; Baptiste Rozière; Barna Pásztor; Baudouin De Monicault; Benjamin Holzschuh; Benjamin Tibi; Benno Stein; Bernhard Stadlbauer; Bhargavi Bhaskar; Binbing Hou; Błażej Osiński; Boris van Breugel; Camille Le Scao; Chaoran Yu; Charlotte Cronjäger; Chen-Yo Sun; Chris Bamford; Christian Wallenwein; Christophe Renaudin; Clémence Lanfranchi; Colin Bonatti; Corentin Barreau; Corentin Sautier; Cristiana-Diana Diaconu; Cyprien Courtot; Daniel Lopes; Daniel Marczak; Darius Dabert; Diego de las Casas; Dominik Nuss; Dylan Rubini; Dzmitry Soupel; Elizaveta Demyanenko; Elliot Chane-Sane; Emilien Fugier; Emmanuel Gottlob; Erik Aas; Etienne Duchesne; Fabian Paischer; Fabian Schlager; Faruk Ahmed; Federico Baldassarre; Félix Rosseeuw; Filip Szatkowski; Filippo Vicentini; Florian Wiesner; Gabrielle Berrada; Gaëtan Ecrepont; Gaspard Blanchet; Gaspard Donada-Vidal; Gaurav Mahajan; Gauthier Guinet; Genevieve Hayes; Georgii Novikov; Georgios Christoglou; Giada Pistilli; Gonzague de Carpentier; Graeme Nail; Guillaume Breton; Guillaume Kunsch; Guillaume Lample; Guillaume Martin; Guillaume Raille; Guillaume Richard; Guillem Simeon; Gunjan Dhanuka; Gunshi Gupta; Hamza Senhaji Rhazi; Han Zhou; Hédi Hadiji; Himanshu Jaju; Hope McGovern; Hubert de Parseval; Hugo Cisneros; Hugo Germain; Hugo Thimonier; Hugo Venturini; Hyeonju Lee; Hyunwoo Oh; Iliass Lasri; Ilyas Oulkadda; Jacques Sun; Jan Ludziejewski; Jason Rute; Jean de Becdelièvre; Jean Quentin; Jean-Malo Delignon; Jessica Chudnovsky; Jianchao Li; Jie Zhang; Jin Bin Ou; Joachim Studnia; Joep Barmentlo; Johannes Brandstetter; John Harvill; Jonas Amar; Jonas Schweizer; Jonathan Roberts; Joséphine Delas; Josselin Somerville; Jules Depersin; Julien Denize; Julien Tauran; Julien Verlaguet; Kartik Khandelwal; Khyathi Raghavi Chandu; Kilian Tep; Konstantin Mishchenko; Kris Szczepaniak; Kush Jain; Larissa Laich; Laura Calem; Laurence Aitchison; Laurent Callot; Laurent Fainsin; Léo Cotteleer; Léonard Blier; Loïc Chadoutaud; Loïck Chambon; Lorenzo Proietti; Louis Faucon; Louis Martin; Louis Serrano; Lu Li; Luc Boudier; Lucas Fernandes Martins; Lucas Mebille; Lucile Saulnier; Ludovic Ho Fuh; Luis Montero; Łukasz Borchmann; Maarten Buyl; Malik Hassan Raza; Manon Chossegros; Marcin Możejko; Margaret Jennings; Markus Hennerbichler; Martin Alexandre; Mathieu Poirée; Mathieu Schmitt; Mathilde Guillaumin; Mathis Felardos; Matthieu André; Matthieu Dinot; Matthieu Futeral; Maurits Bleeker; Mauro Comi; Max Mynter; Maxim Berman; Maxime Darrin; Maxime Langelier; Maxime Louis; Maximilian Augustin; Maximilian Müller; Mehdi Bouaziz; Melina Jingting Laimon; Mert Unsal; Mia Chiquier; Michael Pilcer; Michał Pietruszka; Michał Zając; Mickaël Seznec; Mikhail Biriuchinskii; Minh-Quang Pham; Minwoo Kang; Mircea Lica; Morgane Rivière; Namit Katariya; Natasha Ørregaard Klingenbrunn; Nathan Bout; Nathan Grinsztajn; Nathan Simpson; Neeraj Aggarwal; Neil Ashton; Neil Mallinar; Nick Bührer; Nicolas Moreau; Nicolò Lucchesi; Ola Mysiak; Oliver Leicht; Olivier Bousquet; Olivier Koch; Parag Jain; Patricia Wang; Patrick Blies; Patrick von Platen; Patryk Saffer; Paul Jacob; Paul Teiletche; Paul Wambergue; Paula Kurylowicz; Pavan Kumar Reddy; Pavel Kuksa; Philippe Formont; Philippe Pinel; Philomène Chagniot; Pierre Glandon; Pierre Stock; Pierre-André Savalle; Piotr Miłoś; Poojan Shah; Prateek Gupta; Pravesh Agrawal; Pui Ki Chan; Quentin Desreumaux; Quentin Selle; Quentin Torroba; Quentin Wargnier; Quercus Hernandez; Ram Ramrakhya; Randall Isenhour; Ranjit Parva; Raphaël Balouka; Raphaël Rialland; Raul Perez Pelaez; Reinhard Sonnleitner; Remi Delacourt; Rishi Shah; Rob Romijnders; Rohan Shah; Rohin Arora; Romain Sauvestre; Roman Soletskyi; Ronan Riochet; Rosalie Millner; Rupert Menneer; Sagar Vaze; Samuel Barry; Samuel Belkadi; Samuel Humeau; Samuele Papa; Sanchit Gandhi; Sandeep Cranganore; Sandeep Subramanian; Sarthak Mittal; Saskia Adaime; Scott Matthews; Sean Cha; Sebastian Kaltenbach; Seungjae Jung; Shashwat Verma; Sherif Waly; Shrimai Prabhumoye; Siddhant Waghjale; Siddharth Gandhi; Simon Lepage; Simon Sorg; Soham Ghosh; Sophie Marbach; Srijan Mishra; Stanislas Lange; Steve Hong; Sumukh Aithal; Szymon Antoniak; Tarun Kumar Vangani; Teddy Turquet; Teven Le Scao; Théo Cachet; Thibaut Lavril; Thomas Chabal; Thomas Coste; Thomas Defard; Thomas Foubert; Thomas Loux; Thomas Robert; Thomas Wang; Tianyu Zhang; Tim Lawson; Timothée Lacroix; Timothy Wallis; Tobias Kronlachner; Tom Bewley; Tom Edwards; Tomas Hodan; Triantafyllos Afouras; Tuhin Das; Tyler Wang; Udit Saxena; Ulrick Blé; Umar Jamil; Umberto Tomasini; Valentin Macé; Van Phung; Venkata Reddy; Vibha Masti; Victor Jouault; Victor Letzelter; Victor Paltz; Victor Poucheret; Vincent Maladière; Vincent Pfister; Virgile Richard; Vladislav Bataev; Wassim Bouaziz; Wen Ding Li; William Flanagan; William Havard; William Marshall; Witold Jędrzejewski; Xinghui Li; Xingran Guo; Xinyu Yang; Xuanyu Zhang; Yann Dreze; Yann Viegas; Yannic Neuhaus; Yassine Turki; Yessin Moakher; Yi He; Yihan Wang; Yimu Pan; Yuying Zhu; Yves Martin des Taillades; Zaccharie Ramzi; Zhenlin Xu; Zsófia Csákány 2026-10-27 2026-10-27 cs.AI cs.AI We introduce Timestral, a family of four time series foundation models for probabilistic forecasting, with 5M, 28M, 123M, and 209M parameters. Timestral sets a new state of the art on GIFT-Eval, TIME, and fev-bench, where Timestral-209M and Timestral-123M take the top two places on every metric. We +15 2610.37584 REGER: Modeling inter-item Relations with Semantic IDs for Generative Recommendation Yuxi Huang; Silin Zhou; Chenhao Wang; Zihan Li; Gangming Guo; Pei Ke; Shen Gao; Shuo Shang 2026-10-27 2026-10-27 cs.IR; cs.AI cs.IR Personalized recommendation depends on learning semantic and collaborative relations among items from online content and user interactions. Generative recommendation (GR) offers a promising paradigm by autoregressively generating discrete semantic IDs (SIDs) of items. However, randomly initialized S +16 2610.37571 TeamGym: A Gym for Team Decision-Making by AI Agents with Distributed Information Chanwoo Park; Jaywoong Jeong; Zachary Wojtowicz; Yubin Kim; Taekyun Lee; Kaiqing Zhang; Asuman Ozdaglar 2026-10-27 2026-10-27 cs.AI cs.AI Many real-world decisions require teams to integrate information distributed across participants. Yet AI agents are still primarily trained and benchmarked in isolation. We introduce TeamGym, an environment for evaluating and training language model-based agents on team decision-making problems cons +17 2610.37567 Gym-Anything-RT: Scaling Real-Time Environments for Vision-Language Agents Pranjal Aggarwal; Sean Welleck 2026-10-27 2026-10-27 cs.LG; cs.AI cs.LG AI agents have made substantial progress in domains where they have a non-trivial amount of time to deliberate between actions, such as mathematics, software engineering, and performing tasks on desktop applications. It is less clear how such models fare in settings that demand real-time interaction +18 2610.37561 When uncertainty quantification fails: structural vulnerabilities and corrections for selective prediction under clinical class imbalance Walter Endrizzi; Flavio Ragni; Raffaele Marchesi; Stefano Bovo; Monica Moroni; Giuseppe Jurman; Venet Osmani 2026-10-27 2026-10-27 cs.LG; cs.AI cs.LG Background. Selective prediction systems improve the safety of clinical artificial intelligence by deferring ambiguous cases to human experts while automating high-confidence decisions. However, this safety mechanism relies entirely on the premise that uncertainty scores accurately reflect clinical +19 2610.37550 Kinematic Constraint-Referenced Evaluation for Articulated Object Manipulation Baptiste Bonin; Caro Strickland; Audrey Durand 2026-10-27 2026-10-27 cs.RO; cs.AI cs.RO Manipulation policies are evaluated predominantly by binary task success, which cannot tell whether the commanded gesture is the one the object's kinematics admit. We introduce Constraint-Referenced Strategy (CRS), which projects the commanded end-effector twist onto the object's joint axis, separat +20 2610.37547 When, Where, and What to Correct: Human-Guided Adaptive Residual Correction for Robot Policies Anukriti Singh; Maxime Caniot; Pratap Tokekar; Firas Abi-Farraj 2026-10-27 2026-10-27 cs.RO; cs.AI cs.RO Imitation policies that reliably execute most of a manipulation trajectory can still fail in localized regions requiring last-mile precision. Updating the full policy or exploring residual actions everywhere treats this local failure as a global learning problem. We introduce Adaptive Residual Corre diff --git a/wizard-work/ledger/2026-10-28-arxiv-csai-recent/raw_api.xml b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/raw_api.xml new file mode 100644 index 0000000..940fe18 --- /dev/null +++ b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/raw_api.xml @@ -0,0 +1,1605 @@ + + + https://arxiv.org/api/vLCmJh+cSjpRTe+CD7vJBA8w+qc + arXiv Query: search_query=cat:cs.AI&id_list=&start=0&max_results=20 + 2026-10-28T02:33:55Z + + 20 + 209696 + 0 + + http://arxiv.org/abs/2610.37668v1 + Honeycomb-Surfels: Compact Hexagonal Primitives for Efficient Scene Rendering and Reconstruction + 2026-10-27T17:59:59Z + + + Novel view synthesis and 3D reconstruction have advanced rapidly with radiance fields and splatting, yet their memory and computational demands limit their use in resource-constrained settings. We introduce Hexels: novel planar hexagonal surface elements with a first-order spatial color field for efficient 3D scene and 2D image representation. Hexels exploit locally planar regions with smoothly varying appearance using fewer, larger primitives. Their hexagonal support provides an area- and boundary-efficient domain, while the local color field captures appearance variation within each primitive. Across 34 scenes, Hexels improve PSNR over memory-matched 2DGS by 4.53-7.38 dB across datasets while requiring lower training VRAM. We further ablate Hexels by showing that per primitive, Hexels outperform 2DGS by upto 4.74 dB while using lower training VRAM and having 2.31 times faster FPS. Even in unconstrained settings, Hexels offers complementary quality-memory trade-offs. For efficient 3D surface reconstruction, Hexels reduce error by 28.9% compared with memory-matched 2DGS across 15 DTU scans. On 6 Tanks and Temples surface scenes, Hexels achieves the highest mean F1 among compared methods (34.38%, versus 31.72% for 2DGS). We further also show that Hexels, as a general primitive, improves 2D image representation over Image-GS. Hexels push the Pareto frontier on the quality-efficiency trade-off of surface-aligned primitives. + + + + + 2026-10-27T17:59:59Z + Project page at https://honeycomb-surfels.github.io/ + + + Prakhar Kaushik + + + Soumava Paul + + + Alan Yuille + + + + http://arxiv.org/abs/2610.37666v1 + ThinkWorld: World Imagination Reinforces Chain-of-Dynamics Reasoning + 2026-10-27T17:59:52Z + + + Vision-Language-Action (VLA) models have made substantial progress in robotic manipulation, but their generalization remains constrained by the diversity and coverage of action-labeled robot demonstrations. Human videos provide much broader behavioral and environmental experience, while learning from them faces a fundamental trade-off between scalable data and explicit action grounding. We introduce ThinkWorld, a VLA learning framework that grounds action reasoning in world dynamics and strengthens it through reinforcement learning with world imagination. ThinkWorld represents each action thought as a Chain-of-Dynamics (CoD), a sequence of discrete motion tokens that captures the intermediate dynamics connecting the current world to a future state and serves as a common action representation across embodiments. We first initialize CoD with action-labeled demonstrations, then predict the future induced by each CoD and reinforce those whose predictions better match the observed outcome. This enables videos without action annotations to provide direct supervision from observed state transitions, allowing human and robot experience to refine the same reasoning space. Experiments in simulation and on real robots demonstrate effective human-to-robot transfer and strong generalization to unseen tasks and environments. + + + + 2026-10-27T17:59:52Z + Project Page: https://thinkworld-project.github.io/ + + + Chaehyun Kim + + + Chi-Pin Huang + + + Jisoo Kim + + + Min-Hung Chen + + + Yu-Chiang Frank Wang + + + Seungryong Kim + + + Fu-En Yang + + + + http://arxiv.org/abs/2610.37656v1 + Attention as Lattice Gauge Theory: Relations Interact Only Through Wilson Loops + 2026-10-27T17:58:48Z + + + Relations compose through the entity they share - Alice is Bob's mother and Bob is Carol's father, so Alice is Carol's grandmother - and the order matters, since a father's mother is not a mother's father; verifying a composed relation means closing the loop. We show that if each entity's coordinate frame may be rotated independently (per-entity frame freedom), three things follow. In the pure-gauge sector, relations that share no cycle cannot interact at all. The only gauge-invariant score of a triangle is a class function of its holonomy $H = T_{ij} T_{jk} T_{ki}$ - a Wilson loop. And the only equivariant update of an edge is $U = Ψ(H)\, T_{ij}$ with $Ψ$ conjugation-equivariant. Edge attention by Wilson loops is therefore not one design among many but the unique form the symmetry permits. Three communities hold pieces of this form: the commuting-operator embeddings of knowledge graphs are its abelian (commuting) sector; AlphaFold's triangle update and the Edge Transformer have its shape - an edge updated through the third vertex - without its constraints; and lattice QCD builds it equivariantly on lattices. We instantiate the equivariant sub-family on general relational graphs as holonomy attention. With no training and composition alone, it infers unobserved $S_4$ relations on sparse graphs at 0.88 with a fifth of the edges corrupted, at or slightly below specialist synchronization algorithms, where every commutative architecture is capped at 0.083; its holonomy score flags corrupted triangles at AUC 0.99 where commuting scores are capped at 0.76 by a second selection rule, and it separates clusters without labels. + + + + 2026-10-27T17:58:48Z + 7 pages, 1 figure. Accepted at the 9th Workshop on Machine Learning and the Physical Sciences (ML4PS 2026), co-located with NeurIPS 2026 + + + Aakash Sarkar + + + Ananya Mukherjee + + + Motoshige Sato + + + + http://arxiv.org/abs/2610.37655v1 + KPPax: preprocess differentiable kinetic ODE for machine learning in atmospheric chemistry + 2026-10-27T17:58:40Z + + + Atmospheric chemical mechanisms are commonly distributed in the language of the Kinetic PreProcessor (KPP), which generates efficient Fortran or C code but does not fit into the differentiable programming frameworks in which machine-learning models are developed. We present KPPax, a kinetic preprocessor that generates differentiable, just-in-time compiled, and vectorisable chemistry code in JAX from mechanisms written in the KPP language. KPPax evaluates the rate law with sparse operations in log space, which keeps it differentiable, and integrates the stiff system with the implicit solvers and the forward- and reverse-mode adjoints of the \texttt{diffrax} library, so that exact sensitivities are obtained without hand-derived adjoint code. We demonstrate two applications with the POLLU benchmark mechanism. In the first, the Fisher information matrix computed from an automatically differentiated Jacobian predicts, before any fit, which of the 25 rate coefficients a flow-tube-like design can identify, and noisy synthetic estimation confirms the predicted uncertainties of the nine identifiable rate coefficients to within about 30\% while revealing clusters of fast reactions of which only ratios are constrained. In the second, a neural-network correction of an emission inventory is trained offline on per-scenario 4D-Var retrievals and then fine-tuned online through the chemical model, which reduces the error of the biogenic correction by a factor of 2.4--5 and recovers information on low-temperature emissions that the individual retrievals lose. KPPax makes KPP mechanisms directly usable for parameter estimation and hybrid physics--machine-learning modelling. + + 2026-10-27T17:58:40Z + + + Wenqing Peng + + + Michael Boy + + + Zhi-Song Liu + + + + http://arxiv.org/abs/2610.37649v1 + Can LLMs Uncover Hidden Physical Laws through Free Exploration? + 2026-10-27T17:57:00Z + + + Scientific discovery often begins with an unexpected observation that prompts a new research question. In a world with altered physical laws, can large language models (LLMs) notice unexpected behavior, such as heavier objects falling more slowly, and discover the law through experiments without being told what to investigate? To evaluate this ability, we introduce LongChang, an automated framework that creates simulated worlds with altered physics, where one or more laws differ from ordinary physics. The evaluated LLMs act as explorers, manipulating objects and building tools to investigate their own questions without research tasks or hints about these laws. Separate LLM agents generate and review scenes, review tool designs, and evaluate reports for correctness and experimental support using the altered physical laws and exploration records. These records guide scene evolution for further exploration, allowing the full cycle to run without human intervention. Among 12 LLMs, GPT-6-Astra achieves a discovery rate of 16.7%, while the other 11 score 0%. The exploration records reveal uninvestigated anomalies and only partial identification of the altered physical laws. + + + + 2026-10-27T17:57:00Z + https://zhenyangcs.github.io/LongChang-website/ + + + Zhen Yang + + + Wen Wang + + + Xander Xu + + + Ying-Cong Chen + + + + http://arxiv.org/abs/2610.37647v1 + MuonGS: Pretraining LLMs with Relaxed Weight-Norm Control + 2026-10-27T17:56:40Z + + + Weight decay and explicit weight-norm constraints improve optimizer performance in language model pretraining, yet a unified understanding of why such control helps and how much is needed remains incomplete. We analyze norm control in Muon over the entire training trajectory using the \emph{accumulated effective learning rate}, $C_T=\sum_{t<T}η_t/\|W_t\|_F$, where $η_t$ is the nominal learning rate and $η_t/\|W_t\|_F$ is the effective learning rate relative to the weight norm. For bounded update directions, accumulation over any training interval bounds relative weight displacement. We show that vanilla Muon can accumulate only logarithmically, potentially limiting exploration and helping explain its slower empirical progress, whereas MuonW and MuonH guarantee at least linear accumulation under the stated conditions. Going beyond bounded-norm control, we show that sufficiently slow norm growth guarantees faster accumulation than the logarithmic worst case, opening up less restrictive optimizer designs. + Based on this insight, we propose MuonGS, which uses Gram--Schmidt (GS) projection to remove the radial component of each Muon update without fixing the weight norm. With bounded update directions and a constant nominal learning rate, we prove that the weight norm grows at most proportionally to $\sqrt{t}$ after $t$ steps. Consequently, MuonGS guarantees $C_T=Ω(\sqrt{T})$, with accumulation sustained even in later training. Further, for smooth scale-invariant objectives we establish its $O(T^{-1/2})$ stationarity guarantee. In pretraining experiments with 160M- and 400M-parameter dense models as well as a 135M active- and 1B total- parameter MoE model, MuonGS achieves lower validation perplexity than MuonH at long training horizons. In particular, for the dense model it achieves an estimated token-equivalent speedup up to about $15\%$ at matched validation loss to MuonH. + + + + 2026-10-27T17:56:40Z + 21 pages, 9 figures + + + Athanasios Glentis + + + Dawei Li + + + Mingyi Hong + + + + http://arxiv.org/abs/2610.37631v1 + What Improves During Pretraining? Tracking Knowledge Acquisition with Point-in-Time Models + 2026-10-27T17:52:38Z + + + Language models pretrained on web corpora collected over multiple years learn both general and domain-specific knowledge. How these different kinds of knowledge are learned as pretraining progresses remains unclear. Recent point-in-time models trained on web snapshots from different years provide a way to probe how knowledge prevalence varies over time. Using these models, we propose a temporal variability score (TVS), defined as the coefficient of variation of an item's negative log-likelihood across models. A higher score suggests that the knowledge may be less consistently represented across years of web text. We evaluate six benchmarks covering diverse topics using six models from multiple open-source model families. We find that low- and high-TVS items are learned similarly early in training, but high-TVS items improve faster later on. At the end of the pretraining run, high-TVS items outperform low-TVS items in 34 of 36 model-benchmark settings, with an average advantage of 7.3 points of chance-normalized accuracy. This pattern is most robust on knowledge-intensive tasks, even when comparing items of similar difficulty. Our findings reveal differences in how knowledge is learned over the course of pretraining, with temporal variability providing a diagnostic of these learning dynamics beyond aggregate benchmark scores. + + 2026-10-27T17:52:38Z + Accepted to AACL 2026. This version corrects minor errors in the appendix + + + Chang Su + + + Bin Wu + + + Emine Yilmaz + + + Yao Lu + + + + http://arxiv.org/abs/2610.37624v1 + Rare, Relevant, and Still Wrong: How Memory Fails LLM Group Decisions + 2026-10-27T17:49:39Z + + + When an LLM group's discussion is compressed into a short memory for a final decision maker, what should the memory keep? Two natural heuristics keep the rarest claims or the claims relevant to the decision criteria. On hidden-profile tasks both look attractive, because how widely a fact is held is confounded with whether it helps. We introduce CrossProfile, a diagnostic benchmark of 60 templates with automatically verified answer keys, in which the ownership of supporting, decoy, and irrelevant facts is crossed while their text is held fixed (480 tasks). Across Qwen2.5-7B, Llama-3.1-8B, and two DeepSeek models, both heuristics fail in predictable places. Selecting by relevance readmits decoys, because decoys are on topic: even a perfect relevance filter loses 23 to 31 points to rarest-first selection in the classic hidden-profile configuration. Rarest-first selection is no better than random selection on average and wins only where the rare facts happen to be the useful ones. In both cases accuracy follows the balance of evidence the memory keeps, not the policy that kept it. The decision model adds a third failure, a form of social proof: it gives extra weight to facts attributed to many agents. Removing the attribution tags, or reassigning them at random with line order fixed, raises accuracy in the classic configuration by 13 to 24 points. In natural discussions the evidence is lost before any memory acts: the complete transcript yields lower accuracy than no notes at all. + + + + + 2026-10-27T17:49:39Z + + + Anouar Imel + + + Imad Charradi + + + Abderrahman Sghir + + + Khalil El Khammar + + + + http://arxiv.org/abs/2610.37613v1 + VideoLVR: Latent Visual Reasoning over Grounded Video Evidence + 2026-10-27T17:41:56Z + + + Chain-of-thought reasoning has substantially advanced video understanding in multimodal large language models (MLLMs). However, language-based reasoning traces do not provide a reliable mechanism for models to revisit relevant visual evidence, leading to visual forgetting and impairing their ability to reason over video content. Existing approaches maintain visual grounding during reasoning in two main ways: they either refer to evidence using textual coordinates, which specify symbolic locations rather than the visual content itself, or reinsert cropped regions or frames into the multimodal context, preserving visual details at the cost of substantial context expansion. To address these limitations, we introduce VideoLVR, which constructs latent visual evidence from local regions at relevant video moments, thereby providing visual support for intermediate reasoning steps. To supervise latent visual evidence generation, we introduce a visual token prediction objective that formulates visual token reconstruction as prediction over a vocabulary derived from the input video, thereby unifying visual evidence reconstruction and text generation within the same token prediction framework. This shared prediction framework enables MLLMs to reason over interleaved text tokens and latent visual states while integrating evidence from different moments. We further extend VideoLVR to reinforcement learning by treating sampled visual tokens as actions alongside text tokens and jointly optimizing the policy using final-answer rewards. By efficiently revisiting task-relevant visual information during reasoning, VideoLVR improves video understanding without substantially expanding the multimodal context. Experiments on multiple video benchmarks demonstrate its effectiveness, especially on tasks requiring fine-grained visual evidence. + + + 2026-10-27T17:41:56Z + 13 pages + + + Shilong Li + + + Xingyuan Bu + + + Hao Shi + + + Yun Xing + + + Weixin Feng + + + Qiangpeng Yang + + + Shilei Wen + + + + http://arxiv.org/abs/2610.37605v1 + Beyond Answer Changes: Matched Same-Class Controls for Local Audio Evidence + 2026-10-27T17:36:06Z + + + For audio-language models, an answer change after a local sound-event edit can reflect a response to the event's class, sensitivity to waveform rewriting, or both. We compare matched same-class and different-class replacements of a queried event. Both use the same interval, level estimate, donor fitting, and crossfade, while the same-class replacement preserves the correct answer. Across 600 synthetic soundscapes and three models from two architecture families, different-class replacements produce 49.7 to 60.9 percentage points more answer changes than same-class replacements. Same-class accuracy remains within 5.2 percentage points of original accuracy. Among trials with correct original and same-class answers, soundscape-weighted selective accuracy for the inserted class is 73.7 to 81.9%, averaging eligible trials within each soundscape and then weighting eligible soundscapes equally. This protocol tests how model answers respond to local changes in event class. + + + 2026-10-27T17:36:06Z + Under review at ICASSP 2027. Code and frozen evidence: https://github.com/huaruoji/beyond-answer-changes + + + Yunuo Wang + + + Maoyuan Chen + + + Yuanjin Huang + + + Xing Liu + + + + http://arxiv.org/abs/2610.37599v1 + MoI-Bench: A Benchmark for Image Generation and Editing on Mobile Devices + 2026-10-27T17:31:40Z + + + Developing on-device image models requires considering practical user needs, task types and device resource constraints. However, existing benchmarks provide limited guidance on identifying capability gaps between on-device models and server-side models. We introduce MoI-Bench, a benchmark grounded in mobile user needs, comprising 64 subtasks and 1150 samples. It defines 16 operations, such as removal and replacement, and eight visual target categories covering content such as portraits and text. These form two orthogonal axes whose applicable combinations define specific tasks, such as removing facial makeup. To better guide model improvement, we further propose MoI-Eval, which introduces task-specific criteria covering instruction following, visual quality and detail preservation. We further construct sample-level evaluation questions and aggregate their scores into overall quality and capability profiles along the operation, target and criteria axes. We evaluate representative on-device and server-side models and measure mobile deployment costs. Results show that this decoupled Operation x Target design provides finer-grained guidance for improving compact models. For example, it highlights the need to prioritize specific capability gaps (e.g., text-replacement) over broad application-level goals (e.g., improving text editing). Unlike conventional benchmarks that rank models using a single score, we introduce a two-dimensional Quality-Deployability capability map to assess models for practical mobile use. The map shows that current on-device models can achieve efficient generation or editing but expanding their task coverage deserves greater attention than further improving deployment efficiency. Further progress requires improving editing quality, particularly identity preservation and text rendering, while broadening task coverage. + + + 2026-10-27T17:31:40Z + Project page: https://carlofkl.github.io/MoI-Bench-project/ + + + Kailai Feng + + + Haoxiang Rao + + + Yuxiang Wei + + + Bo Chen + + + Junsong Li + + + Yang Pan + + + Chengkun Cai + + + Jagmohan Chauhan + + + + http://arxiv.org/abs/2610.37598v1 + GAS-PFN: Graph-Aware Spatiotemporal Prior-Data Fitted Networks for Zero-Shot Traffic Forecasting + 2026-10-27T17:31:26Z + + + Traffic forecasting requires modeling complex temporal dynamics and the spatial dependencies imposed by road topology. Supervised graph-based models achieve strong performance, but are often trained for a particular dataset and sensor graph, making deployment to new or evolving networks costly. Prior-Data Fitted Networks (PFNs) offer a zero-shot alternative by learning in-context inference from synthetic task priors, which provide an explicit mechanism for encoding domain-specific assumptions. Existing forecasting PFNs, however, do not explicitly represent the graph-structured dependencies governing traffic dynamics. To combine graph-aware traffic modeling with zero-shot adaptation, we introduce the Graph-Aware Spatiotemporal Prior-Data Fitted Network (GAS-PFN). GAS-PFN incorporates sampled road topology into its synthetic task prior through random-walk diffusion, teaching the model to learn graph-coupled dynamics during pretraining. Its co-designed inference architecture uses two mask-conditioned passes through a shared Transformer backbone to separately model sensor-specific temporal dynamics and spatiotemporal interactions among neighboring sensors before integrating their representations for prediction. This design enables GAS-PFN to use target-network structure and context without parameter updates. Across six benchmarks at the 90-minute horizon, GAS-PFN reduces MSE by 34.0% on average relative to TimePFN, its closest PFN-based predecessor. It also reduces MSE by 27.7% on average relative to four recent industry-developed general-purpose zero-shot forecasters. In this comparison, all models receive 90 minutes of contiguous recent history, and any additional context is drawn from each dataset's chronological training split. These results indicate that aligning PFN priors and inference architectures with graph structure improves in-context learning for zero-shot traffic forecasting. + + + + 2026-10-27T17:31:26Z + + + Joseph Natter + + + Long Vu + + + Yifan Zhang + + + Bobin Deng + + + Kun Suo + + + Rui Wu + + + + http://arxiv.org/abs/2610.37588v1 + Understanding the limits of latent world models + 2026-10-27T17:25:50Z + + + World models are widely proposed as core components for enabling agents to reason over long temporal and spatial horizons. Yet current evaluations often emphasize task performance while providing limited insight into whether learned world models acquire predictive latent dynamics that reflect the underlying structure of the environment. Here we introduce the concept of faithful latent world models as an equivariance condition between the environment's transition map and the learned latent dynamics. We propose latent rollout consistency (LRC) as a metric to evaluate action-conditioned latent prediction. Strikingly, we find that across the studied state of the art latent world models, none solves the task by actually faithfully modeling the dynamics of the environment. + + + 2026-10-27T17:25:50Z + + + Rodrigo Gonzalez Laiz + + + Steffen Schneider + + + + http://arxiv.org/abs/2610.37585v1 + Timestral + 2026-10-27T17:25:43Z + + + We introduce Timestral, a family of four time series foundation models for probabilistic forecasting, with 5M, 28M, 123M, and 209M parameters. Timestral sets a new state of the art on GIFT-Eval, TIME, and fev-bench, where Timestral-209M and Timestral-123M take the top two places on every metric. We achieve this with three ingredients: Decoupled Horizon Forecasting (DHF), which separates prediction tokens from the causally encoded context, so that longer horizons are trained without masking context and each observed value is both a forecast target and context for later forecasts; a covariate framework that natively supports known-future covariates within a causal architecture; and a two-stage pipeline that pretrains exclusively on a diverse, large-scale synthetic corpus before post-training on a mix of real and synthetic data with extended context. This report describes the model design and its benchmark results. Weights for all four models are available on Hugging Face and inference code on GitHub. + + 2026-10-27T17:25:43Z + Webpage: https://mistral.ai/news/timestral ; Weights: https://huggingface.co/collections/mistralai/timestral ; Code: https://github.com/mistralai/timestral + + + Adrien Petralia + + + Alexandre Gavaudan + + + Artjom Joosen + + + Etienne Goffinet + + + Richard Kurle + + + Yassine El Ouahidi + + + Yassir Bendou + + + Abdelaziz Bounhar + + + Abhijeet Somani + + + Aditi Kabra + + + Adrian Valente + + + Adrien Sadé + + + Ahmed Djermani + + + Alan Jeffares + + + Albert Jiang + + + Alberto Bietti + + + Aleksandr Timashov + + + Alexandre Cahill + + + Alexandre Laval + + + Alexandre Sablayrolles + + + Alexandros Sanchez Vassopoulos + + + Amélie Héliou + + + Amos You + + + André Jonasson + + + Andreas Fürst + + + Andrew Bai + + + Andrew Ehrenberg + + + Andrew Zhao + + + Andy Lo + + + Angele Lenglemetz + + + Anmol Agarwal + + + Antonia Calvi + + + Arata Suzuki + + + Arjun Majumdar + + + Arthur Fournier + + + Ashwani Kumar + + + Augustin Lombard + + + Avinash Sooriyarachchi + + + Aylin Guliz Akkus + + + Aysenur Karaduman + + + Baptiste Bout + + + Baptiste Rozière + + + Barna Pásztor + + + Baudouin De Monicault + + + Benjamin Holzschuh + + + Benjamin Tibi + + + Benno Stein + + + Bernhard Stadlbauer + + + Bhargavi Bhaskar + + + Binbing Hou + + + Błażej Osiński + + + Boris van Breugel + + + Camille Le Scao + + + Chaoran Yu + + + Charlotte Cronjäger + + + Chen-Yo Sun + + + Chris Bamford + + + Christian Wallenwein + + + Christophe Renaudin + + + Clémence Lanfranchi + + + Colin Bonatti + + + Corentin Barreau + + + Corentin Sautier + + + Cristiana-Diana Diaconu + + + Cyprien Courtot + + + Daniel Lopes + + + Daniel Marczak + + + Darius Dabert + + + Diego de las Casas + + + Dominik Nuss + + + Dylan Rubini + + + Dzmitry Soupel + + + Elizaveta Demyanenko + + + Elliot Chane-Sane + + + Emilien Fugier + + + Emmanuel Gottlob + + + Erik Aas + + + Etienne Duchesne + + + Fabian Paischer + + + Fabian Schlager + + + Faruk Ahmed + + + Federico Baldassarre + + + Félix Rosseeuw + + + Filip Szatkowski + + + Filippo Vicentini + + + Florian Wiesner + + + Gabrielle Berrada + + + Gaëtan Ecrepont + + + Gaspard Blanchet + + + Gaspard Donada-Vidal + + + Gaurav Mahajan + + + Gauthier Guinet + + + Genevieve Hayes + + + Georgii Novikov + + + Georgios Christoglou + + + Giada Pistilli + + + Gonzague de Carpentier + + + Graeme Nail + + + Guillaume Breton + + + Guillaume Kunsch + + + Guillaume Lample + + + Guillaume Martin + + + Guillaume Raille + + + Guillaume Richard + + + Guillem Simeon + + + Gunjan Dhanuka + + + Gunshi Gupta + + + Hamza Senhaji Rhazi + + + Han Zhou + + + Hédi Hadiji + + + Himanshu Jaju + + + Hope McGovern + + + Hubert de Parseval + + + Hugo Cisneros + + + Hugo Germain + + + Hugo Thimonier + + + Hugo Venturini + + + Hyeonju Lee + + + Hyunwoo Oh + + + Iliass Lasri + + + Ilyas Oulkadda + + + Jacques Sun + + + Jan Ludziejewski + + + Jason Rute + + + Jean de Becdelièvre + + + Jean Quentin + + + Jean-Malo Delignon + + + Jessica Chudnovsky + + + Jianchao Li + + + Jie Zhang + + + Jin Bin Ou + + + Joachim Studnia + + + Joep Barmentlo + + + Johannes Brandstetter + + + John Harvill + + + Jonas Amar + + + Jonas Schweizer + + + Jonathan Roberts + + + Joséphine Delas + + + Josselin Somerville + + + Jules Depersin + + + Julien Denize + + + Julien Tauran + + + Julien Verlaguet + + + Kartik Khandelwal + + + Khyathi Raghavi Chandu + + + Kilian Tep + + + Konstantin Mishchenko + + + Kris Szczepaniak + + + Kush Jain + + + Larissa Laich + + + Laura Calem + + + Laurence Aitchison + + + Laurent Callot + + + Laurent Fainsin + + + Léo Cotteleer + + + Léonard Blier + + + Loïc Chadoutaud + + + Loïck Chambon + + + Lorenzo Proietti + + + Louis Faucon + + + Louis Martin + + + Louis Serrano + + + Lu Li + + + Luc Boudier + + + Lucas Fernandes Martins + + + Lucas Mebille + + + Lucile Saulnier + + + Ludovic Ho Fuh + + + Luis Montero + + + Łukasz Borchmann + + + Maarten Buyl + + + Malik Hassan Raza + + + Manon Chossegros + + + Marcin Możejko + + + Margaret Jennings + + + Markus Hennerbichler + + + Martin Alexandre + + + Mathieu Poirée + + + Mathieu Schmitt + + + Mathilde Guillaumin + + + Mathis Felardos + + + Matthieu André + + + Matthieu Dinot + + + Matthieu Futeral + + + Maurits Bleeker + + + Mauro Comi + + + Max Mynter + + + Maxim Berman + + + Maxime Darrin + + + Maxime Langelier + + + Maxime Louis + + + Maximilian Augustin + + + Maximilian Müller + + + Mehdi Bouaziz + + + Melina Jingting Laimon + + + Mert Unsal + + + Mia Chiquier + + + Michael Pilcer + + + Michał Pietruszka + + + Michał Zając + + + Mickaël Seznec + + + Mikhail Biriuchinskii + + + Minh-Quang Pham + + + Minwoo Kang + + + Mircea Lica + + + Morgane Rivière + + + Namit Katariya + + + Natasha Ørregaard Klingenbrunn + + + Nathan Bout + + + Nathan Grinsztajn + + + Nathan Simpson + + + Neeraj Aggarwal + + + Neil Ashton + + + Neil Mallinar + + + Nick Bührer + + + Nicolas Moreau + + + Nicolò Lucchesi + + + Ola Mysiak + + + Oliver Leicht + + + Olivier Bousquet + + + Olivier Koch + + + Parag Jain + + + Patricia Wang + + + Patrick Blies + + + Patrick von Platen + + + Patryk Saffer + + + Paul Jacob + + + Paul Teiletche + + + Paul Wambergue + + + Paula Kurylowicz + + + Pavan Kumar Reddy + + + Pavel Kuksa + + + Philippe Formont + + + Philippe Pinel + + + Philomène Chagniot + + + Pierre Glandon + + + Pierre Stock + + + Pierre-André Savalle + + + Piotr Miłoś + + + Poojan Shah + + + Prateek Gupta + + + Pravesh Agrawal + + + Pui Ki Chan + + + Quentin Desreumaux + + + Quentin Selle + + + Quentin Torroba + + + Quentin Wargnier + + + Quercus Hernandez + + + Ram Ramrakhya + + + Randall Isenhour + + + Ranjit Parva + + + Raphaël Balouka + + + Raphaël Rialland + + + Raul Perez Pelaez + + + Reinhard Sonnleitner + + + Remi Delacourt + + + Rishi Shah + + + Rob Romijnders + + + Rohan Shah + + + Rohin Arora + + + Romain Sauvestre + + + Roman Soletskyi + + + Ronan Riochet + + + Rosalie Millner + + + Rupert Menneer + + + Sagar Vaze + + + Samuel Barry + + + Samuel Belkadi + + + Samuel Humeau + + + Samuele Papa + + + Sanchit Gandhi + + + Sandeep Cranganore + + + Sandeep Subramanian + + + Sarthak Mittal + + + Saskia Adaime + + + Scott Matthews + + + Sean Cha + + + Sebastian Kaltenbach + + + Seungjae Jung + + + Shashwat Verma + + + Sherif Waly + + + Shrimai Prabhumoye + + + Siddhant Waghjale + + + Siddharth Gandhi + + + Simon Lepage + + + Simon Sorg + + + Soham Ghosh + + + Sophie Marbach + + + Srijan Mishra + + + Stanislas Lange + + + Steve Hong + + + Sumukh Aithal + + + Szymon Antoniak + + + Tarun Kumar Vangani + + + Teddy Turquet + + + Teven Le Scao + + + Théo Cachet + + + Thibaut Lavril + + + Thomas Chabal + + + Thomas Coste + + + Thomas Defard + + + Thomas Foubert + + + Thomas Loux + + + Thomas Robert + + + Thomas Wang + + + Tianyu Zhang + + + Tim Lawson + + + Timothée Lacroix + + + Timothy Wallis + + + Tobias Kronlachner + + + Tom Bewley + + + Tom Edwards + + + Tomas Hodan + + + Triantafyllos Afouras + + + Tuhin Das + + + Tyler Wang + + + Udit Saxena + + + Ulrick Blé + + + Umar Jamil + + + Umberto Tomasini + + + Valentin Macé + + + Van Phung + + + Venkata Reddy + + + Vibha Masti + + + Victor Jouault + + + Victor Letzelter + + + Victor Paltz + + + Victor Poucheret + + + Vincent Maladière + + + Vincent Pfister + + + Virgile Richard + + + Vladislav Bataev + + + Wassim Bouaziz + + + Wen Ding Li + + + William Flanagan + + + William Havard + + + William Marshall + + + Witold Jędrzejewski + + + Xinghui Li + + + Xingran Guo + + + Xinyu Yang + + + Xuanyu Zhang + + + Yann Dreze + + + Yann Viegas + + + Yannic Neuhaus + + + Yassine Turki + + + Yessin Moakher + + + Yi He + + + Yihan Wang + + + Yimu Pan + + + Yuying Zhu + + + Yves Martin des Taillades + + + Zaccharie Ramzi + + + Zhenlin Xu + + + Zsófia Csákány + + + + http://arxiv.org/abs/2610.37584v1 + REGER: Modeling inter-item Relations with Semantic IDs for Generative Recommendation + 2026-10-27T17:25:12Z + + + Personalized recommendation depends on learning semantic and collaborative relations among items from online content and user interactions. Generative recommendation (GR) offers a promising paradigm by autoregressively generating discrete semantic IDs (SIDs) of items. However, randomly initialized SID token embeddings and GR models trained from scratch do not directly inherit the semantic relations inherent in items, while inter-item collaborative relations are learned largely implicitly through token-level attention and next-token supervision. To address this limitation, we propose REGER, a unified framework that explicitly incorporates inter-item relations into semantic pretraining, history encoding, and SID generation. We pretrain the GR model through generative semantic retrieval, enabling it to inherit inter-item semantic similarity by learning to generate semantic neighbors' SIDs. We combine global attention with local attention aligned with SID boundaries to capture historical context and directed transitions between adjacent items. We further develop a theoretical framework for hierarchical preference decomposition over the SID prefix tree, proving an exact factorization of collaborative preference distributions and a weighted decomposition of item-level KL divergence into prefix-conditional KL divergences. Building on this analysis, we distill a discriminative teacher's full conditional distributions over all legal next tokens at ground-truth prefixes along the target SID path, jointly optimizing this objective with cross-entropy. Extensive experiments on real-world datasets demonstrate consistent improvements over existing GR baselines and support the effectiveness of the proposed preference distillation. + + + 2026-10-27T17:25:12Z + 13 pages, 6 figures + + + Yuxi Huang + + + Silin Zhou + + + Chenhao Wang + + + Zihan Li + + + Gangming Guo + + + Pei Ke + + + Shen Gao + + + Shuo Shang + + + + http://arxiv.org/abs/2610.37571v1 + TeamGym: A Gym for Team Decision-Making by AI Agents with Distributed Information + 2026-10-27T17:15:54Z + + + Many real-world decisions require teams to integrate information distributed across participants. Yet AI agents are still primarily trained and benchmarked in isolation. We introduce TeamGym, an environment for evaluating and training language model-based agents on team decision-making problems constructed from real-world legal, finance, medical, and software engineering corpora. TeamGym introduces a method to identify a minimal set of facts, each element of which is necessary and all of which are collectively sufficient for a correct final decision, which we call pivotal information. This enables TeamGym to both control the distribution of pivotal information and track how well agents identify and integrate pivotal information when making team decisions. Examining 8 language models on 12 source corpora, we find that current AI agents often fail to effectively share information: Distributing pivotal information across participants lowers success rate by 22.9 percentage points on average relative to giving the same information to a single model (a communication gap). We also document that participants do not effectively retrieve decision-relevant information on their own: A team's success rate is lowered by a further 19.9 points when participants must identify and extract pivotal information from source material rather than receiving it directly (a retrieval gap). To help remedy these deficits, we use TeamGym to train policies through reinforcement learning on team decision loss. Across two policy-training experiments, we find that training improves final-decision quality and induces emergent coordination in participation, speaker selection, and communication, even though the reward evaluates only the final decision. + + 2026-10-27T17:15:54Z + + + Chanwoo Park + + + Jaywoong Jeong + + + Zachary Wojtowicz + + + Yubin Kim + + + Taekyun Lee + + + Kaiqing Zhang + + + Asuman Ozdaglar + + + + http://arxiv.org/abs/2610.37567v1 + Gym-Anything-RT: Scaling Real-Time Environments for Vision-Language Agents + 2026-10-27T17:13:05Z + + + AI agents have made substantial progress in domains where they have a non-trivial amount of time to deliberate between actions, such as mathematics, software engineering, and performing tasks on desktop applications. It is less clear how such models fare in settings that demand real-time interaction, such as playing a video game or operating a physical device. In these settings, the environment state may evolve as the agent is reasoning, or even as the underlying model is performing a forward pass. To gain insight into the real-time interaction capabilities of vision-language agents, we introduce Gym-Anything-RT. Our key idea is to construct visual minigames that demand a portfolio of capabilities required for interactive real-time settings. Gym-Anything-RT uses an agentic pipeline to automatically generate such environments based on a taxonomy of six capabilities required for real-time interaction, including the ability to perceive the environment and act under an evolving environment state. A key feature of Gym-Anything-RT is that it allows for controlled comparison of real-time interaction capabilities: we can run the same minigame in real time, or pause or slow down the environment to relax the requirement for real-time actions. We use Gym-Anything-RT to construct 100 real-time minigames and evaluate AI agents on them. We find large capability gaps in real-time interaction. For example, GPT-6 Astra (xHigh) achieves a pass rate of only 9.0% on the real-time minigames, which rises to 75.0% when the environment is paused. We also identify intriguing behaviors by the most powerful frontier agents: for instance, when given the ability to write code, these agents write code that bypasses the need to perceive and act in real time. We publicly release Gym-Anything-RT and the generated minigames to spur further research on AI agents in real-time interactive settings. + + + 2026-10-27T17:13:05Z + + + Pranjal Aggarwal + + + Sean Welleck + + + + http://arxiv.org/abs/2610.37561v1 + When uncertainty quantification fails: structural vulnerabilities and corrections for selective prediction under clinical class imbalance + 2026-10-27T17:08:31Z + + + Background. Selective prediction systems improve the safety of clinical artificial intelligence by deferring ambiguous cases to human experts while automating high-confidence decisions. However, this safety mechanism relies entirely on the premise that uncertainty scores accurately reflect clinical ambiguity. In prognostic tasks characterized by severe class imbalance, where the clinical outcome of interest is a minority, this assumption breaks down. Standard uncertainty measures cease to capture clinically aligned ambiguity; instead, they become a proxy for class membership, systematically penalizing minority patients regardless of their individual predictive clarity. + Results. Across 56 synthetic conditions and three neurodegenerative cohorts spanning severe class imbalance to near-balanced conditions (multiple sclerosis, Parkinson's Disease, Alzheimer's Disease), our framework consistently maintained bounded, balanced risk across trajectories. In contrast, global metrics either collapsed minority Sensitivity or protected the minority at the expense of majority patients. Applying CCR with standard entropy fails to protect the minority because it breaks within-class monotonicity when tau=0.5. Combining CCR with threshold-aware metrics restores strict monotonicity, correctly identifies ambiguous patients, and safely mitigates minority exclusion. + Conclusions. Under clinical class imbalance, standard uncertainty quantification measures dataset sparsity rather than clinical ambiguity, creating a systemic illusion of safety. This bias is a correctable structural flaw. Clinical pipelines deploying selective prediction under imbalance should adopt threshold-aligned uncertainty measures and class-stratified evaluation. This reduces class-specific risk imbalance, supporting safer selective prediction under the evaluated conditions without silently excluding the most vulnerable patients. + + + 2026-10-27T17:08:31Z + https://github.com/endrwalter/uncertainty_ml + + + Walter Endrizzi + + + Flavio Ragni + + + Raffaele Marchesi + + + Stefano Bovo + + + Monica Moroni + + + Giuseppe Jurman + + + Venet Osmani + + + + http://arxiv.org/abs/2610.37550v1 + Kinematic Constraint-Referenced Evaluation for Articulated Object Manipulation + 2026-10-27T17:00:01Z + + + Manipulation policies are evaluated predominantly by binary task success, which cannot tell whether the commanded gesture is the one the object's kinematics admit. We introduce Constraint-Referenced Strategy (CRS), which projects the commanded end-effector twist onto the object's joint axis, separating progress along the joint from off-constraint waste. CRS requires a joint model and a joint-following scripted controller derived from it, but no recorded demonstration or learned judge. On 74 simulated articulated objects, success registers three transfer failures as lower rates but attributes none, while the two CRS channels give each a distinct signature. A door-trained behavior-cloning policy opens drawers in 59 of 110 episodes, and in these passes it commands about $40\times$ the off-axis rotation of a drawer-trained policy's passes ($1.27$ vs.\ $0.032$~rad/s). Relative to a scripted controller scored with the same axis, CRS verdicts are unchanged under joint-axis errors up to $30^\circ$; absolute thresholds tolerate about $8^\circ$. + + + 2026-10-27T17:00:01Z + + + Baptiste Bonin + + + Caro Strickland + + + Audrey Durand + + + + http://arxiv.org/abs/2610.37547v1 + When, Where, and What to Correct: Human-Guided Adaptive Residual Correction for Robot Policies + 2026-10-27T16:58:54Z + + + Imitation policies that reliably execute most of a manipulation trajectory can still fail in localized regions requiring last-mile precision. Updating the full policy or exploring residual actions everywhere treats this local failure as a global learning problem. We introduce Adaptive Residual Correction (ARC), a lightweight post-training module that uses reinforcement learning (RL) to learn a residual correction on top of a frozen policy while using sparse human corrections to structure autonomous exploration. Rather than treating human corrections as imitation targets, ARC uses pre-intervention states to determine 'when' residual control should become active and correction geometry to determine 'where' in residual action space autonomous refinement should explore; task reward determines 'what' residual is ultimately executed. Across three precision-manipulation tasks on a real robot, ARC improves average task success from 33.3% to 75.0%. The human corrections define task-dependent 2-3 dimensional subspaces of the seven-joint residual space that retain 96.0-97.7% of squared correction magnitude. Together, these results show that sparse human corrections can localize and constrain autonomous post-training without requiring the observed correction itself to serve as the final action target. + + + 2026-10-27T16:58:54Z + + + Anukriti Singh + + + Maxime Caniot + + + Pratap Tokekar + + + Firas Abi-Farraj + + + diff --git a/wizard-work/ledger/2026-10-28-arxiv-csai-recent/reverify.py b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/reverify.py new file mode 100644 index 0000000..1a52ba2 --- /dev/null +++ b/wizard-work/ledger/2026-10-28-arxiv-csai-recent/reverify.py @@ -0,0 +1,153 @@ +#!/usr/bin/env python3 +"""arXiv cs.AI recent-papers extraction + re-verification gate. + +Extract latest arXiv cs.AI submissions from the pinned raw API response +(raw_api.xml) into TSV / JSON / Markdown tables, then re-derive from the same +XML and diff against the committed tables. Exit 0 only on an exact match. + +Usage: + python3 reverify.py emit # write tsv / json / md from raw_api.xml + python3 reverify.py verify # re-derive and diff against committed files +""" +import csv +import json +import re +import sys +import xml.etree.ElementTree as ET + +HERE = "/root/troupe-standup/wizard-work/ledger/2026-10-28-arxiv-csai-recent" +NS_ATOM = "http://www.w3.org/2005/Atom" + +ABSTRACT_LIMIT = 300 +TITLE_LIMIT = 90 +ABSTRACT_SHOW = 50 + +FIELDS = ["rank", "arxiv_id", "title", "authors", "published", "updated", + "categories", "primary_category", "abstract"] + + +def text(el, tag): + child = el.find(f"{{{NS_ATOM}}}{tag}") + return child.text.strip() if child is not None and child.text else "" + + +def derive(root): + """Re-derive the row list from the raw Atom feed. Pure function of XML.""" + rows = [] + for e in root.findall(f"{{{NS_ATOM}}}entry"): + rid = text(e, "id") + m = re.search(r"arxiv\.org/abs/(\d+\.\d+)(?:v\d+)?", rid) + arxiv_id = m.group(1) if m else rid + title = text(e, "title") + published = text(e, "published")[:10] + updated = text(e, "updated")[:10] + authors = [text(a, "name") for a in e.findall(f"{{{NS_ATOM}}}author")] + cats = [c.get("term") for c in e.findall(f"{{{NS_ATOM}}}category")] + abstract = text(e, "summary") + rows.append({ + "arxiv_id": arxiv_id, + "title": title, + "authors": "; ".join(authors), + "published": published, + "updated": updated, + "categories": "; ".join(cats), + "primary_category": cats[0] if cats else "", + "abstract": abstract[:ABSTRACT_LIMIT], + }) + # arXiv API returns descending by submitted date. + rows.sort(key=lambda r: r["published"], reverse=True) + for i, r in enumerate(rows, 1): + r["rank"] = i + return rows + + +def fmt_tsv(rows): + out = io.StringIO() + w = csv.DictWriter(out, fieldnames=FIELDS, delimiter="\t", lineterminator="\n") + w.writeheader() + w.writerows(rows) + return out.getvalue() + + +def fmt_json(rows): + return json.dumps(rows, indent=2, ensure_ascii=False) + "\n" + + +def fmt_md(rows): + lines = [ + "# arXiv cs.AI — most recent 20 submissions (fetched 2026-10-28)", + "", + "Source: `raw_api.xml` (API: cat:cs.AI, sortBy=submittedDate, sortOrder=descending, max_results=20)", + "", + "| Rank | arXiv | Title | Authors | Published | Updated | Categories | Primary | Abstract (first 50 chars) |", + "|---:|---|---|---|---|---|---|---|---|", + ] + for r in rows: + title = r["title"].replace("|", "\\|") + abstract = r["abstract"][:ABSTRACT_SHOW] + ("…" if len(r["abstract"]) > ABSTRACT_SHOW else "") + lines.append(f"| {r['rank']} | {r['arxiv_id']} | {title} | {r['authors']} | " + f"{r['published']} | {r['updated']} | {r['categories']} | {r['primary_category']} | {abstract} |") + return "\n".join(lines) + "\n" + + +def emit(rows, out_dir): + tsv = out_dir + "/arxiv_csai_recent.tsv" + js = out_dir + "/arxiv_csai_recent.json" + md = out_dir + "/arxiv_csai_recent.md" + with open(tsv, "w", encoding="utf-8", newline="") as f: + f.write(fmt_tsv(rows)) + with open(js, "w", encoding="utf-8") as f: + f.write(fmt_json(rows)) + with open(md, "w", encoding="utf-8") as f: + f.write(fmt_md(rows)) + return tsv, js, md + + +def verify(out_dir): + root = ET.parse(out_dir + "/raw_api.xml").getroot() + expected = derive(root) + expected_tsv = fmt_tsv(expected) + expected_json = fmt_json(expected) + expected_md = fmt_md(expected) + files = { + "tsv": out_dir + "/arxiv_csai_recent.tsv", + "json": out_dir + "/arxiv_csai_recent.json", + "md": out_dir + "/arxiv_csai_recent.md", + } + ok = True + for name, path in files.items(): + with open(path, encoding="utf-8") as f: + actual = f.read() + if name == "tsv": + want = expected_tsv + elif name == "json": + want = expected_json + else: + want = expected_md + if actual != want: + ok = False + print(f"FAIL {name}: committed content differs from re-derivation") + if name in ("tsv", "json", "md"): + # Show first differing line for diagnosis. + a_lines = actual.splitlines() + w_lines = want.splitlines() + for i in range(max(len(a_lines), len(w_lines))): + a = a_lines[i] if i < len(a_lines) else "" + w = w_lines[i] if i < len(w_lines) else "" + if a != w: + print(f" first diff at line {i+1}:\n actual: {a[:100]}\n expected: {w[:100]}") + break + if ok: + print(f"PASS re-derived from raw_api.xml; table has {len(expected)} rows") + return 0 if ok else 1 + + +if __name__ == "__main__": + import io + mode = sys.argv[1] if len(sys.argv) > 1 else "emit" + if mode == "emit": + rows = derive(ET.parse(HERE + "/raw_api.xml").getroot()) + for f in emit(rows, HERE): + print(f"WROTE {f}") + else: + sys.exit(verify(HERE))