Pith. sign in

REVIEW 3 major objections 5 minor 131 references

ABot-AgentOS claims that a general agent operating system layer above low-level robot controllers—with hierarchical planning, skill delegation, verification, and source-grounded graph memory—improves long-horizon embodied execution and prov

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-02 07:16 UTC pith:MYSKTYLO

load-bearing objection Serious agent-OS paper with a carefully gated self-evolution protocol; the embodied-execution headline is currently unfalsifiable from the paper alone. the 3 major comments →

arxiv 2607.10350 v3 pith:MYSKTYLO submitted 2026-07-11 cs.AI cs.RO

ABot-AgentOS: A General Robotic Agent OS with Lifelong Multi-modal Memory

Jiayi Tian , Shiao Liu , Yuting Xu , Jia Lu , Zihao Guan , Honglin Han , Di Yang , Minqi Gu
show 25 more authors
Yifei Qian Tianlin Zhang Yanqing Zhu Zeqian Ye Menglin Yang Fei Wang Xu Hu Xiuxian Li Wei Zhang Shihui Su Yiyan Ji Jingbo Wang Ziteng Feng Jiaheng Liu Zhaoxiang Zhang Xiaolong Wu Zixiao Tang Zhining Gu Yang Cai Linbo Zheng Jingjing Ma Mingyang Yin Zedong Chu Wenbin Tang Mu Xu
This is my paper
classification cs.AI cs.RO
keywords agent operating systemembodied AImulti-modal memorygraph memorylong-horizon tasksself-evolutionembodied benchmarkrobot skills
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper is trying to establish that long-horizon embodied intelligence needs a general runtime layer—an "agent operating system"—sitting above both vision-language models and low-level robot controllers. ABot-AgentOS supplies that layer: scene-conditioned planning, context-isolated skill execution, multi-stage verification, edge-cloud routing, and a persistent multi-modal memory built from typed, source-grounded graph nodes and edges. The authors argue this improves task success and goal completion over a single-controller baseline on a new executable benchmark, while the memory graph achieves strong scores on conversational, embodied-QA, multi-modal, and video recall benchmarks. A failure-driven self-evolution loop converts diagnosed memory failures into gated runtime assets that only affect later evaluation splits, turning the system into a lifelong learner. If right, the work matters because it separates cognition from embodiment, promising reusable planning and memory across different robot bodies.

Core claim

At its center, ABot-AgentOS claims that the hard part of long-horizon embodied tasks is not perception or action prediction but the runtime glue between them. Rather than emitting actions directly, the system's main LLM plans against the scene, delegates procedural subtasks to a Skill Runner with an isolated context, and uses a Verifier to check progress, skill outcomes, and finish conditions, closing a reasoning–execution–verification loop. Alongside this harness, the Universal Multi-modal Graph Memory writes dialogue, visual observations, spatial and temporal context, and task traces into typed nodes and edges with provenance, then retrieves local evidence subgraphs for grounded answering.

What carries the argument

Universal Multi-modal Graph Memory: a typed graph of entities, events, places, sessions, visual evidence, spatial/temporal relations, and provenance, written by adapters and queried by hybrid seed selection followed by typed-edge subgraph expansion. It is what makes memory persistent, relational, multi-modal, and auditable. The other load-bearing piece is the three-role Agent Harness—main LLM, Skill Runner, Verifier—that turns ReAct-style reasoning into a closed loop and context-isolates procedural execution.

Load-bearing premise

The agent's view of the world flows through a vision-language tool that converts first-person images into text; if those descriptions are unreliable or the filtered semantic map quietly reveals hidden target states, the measured gains could come from easier observation rather than better planning and memory.

What would settle it

Run the same EmbodiedWorldBench agents with raw visual input (or a second, independent perception encoder) instead of the VLM-generated text observations, keeping all planning, memory, and verification modules fixed; if the TSR/GCR advantage over the single-controller baseline shrinks or disappears, the architecture's benefit is an artifact of perception-text format. A second test: scramble or remove the filtered semantic map and check whether TSR/GCR drop dramatically, which would indicate hidden-state leakage.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Task success and goal completion improve over a single-controller baseline under the same base model, indicating the OS layer itself contributes, not just model scale.
  • The architecture is model-agnostic: substituting a stronger main LLM improves results further, and the skill/tool interface is designed to swap across embodiments.
  • Memory becomes an auditable evidence chain rather than a raw dump: answers are traceable to source nodes, frames, and provenance.
  • Self-evolution can improve later splits without current-split leakage, so a deployed robot can improve from interaction feedback over time.
  • The new executable benchmark with indoor/outdoor/hybrid scenes, dynamic events, and trace-grounded scores offers a reproducible way to measure long-horizon embodied agents.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The split-wise gated self-evolution protocol is a transferable recipe: any memory pipeline could be improved by diagnosing trace failures and promoting only regression-checked, versioned policy assets—independent of the underlying retriever.
  • If the source-grounded graph survives real deployment, it could underpin user-facing explanations ('I answered this because I saw it in frame X at time T'), addressing trust and auditing in household robotics.
  • A direct extension would benchmark whether the same OS layer transfers across embodiments—quadrupeds, arms, mobile manipulators—without retraining the cognitive stack, since that is the paper's stated goal but is not yet measured.
  • The paper leaves open how far the architecture's benefit extends when perception is strong: the listed failures (confusing people/objects, indoor-outdoor misjudgment) suggest that as observation quality improves, gains may shift from textual coping to genuine spatial reasoning.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes ABot-AgentOS, a robotic Agent Operating System layer that sits above low-level controllers and provides a deliberative agent loop, multi-modal graph memory, skill execution, verification, and edge-cloud collaboration. It also introduces EmbodiedWorldBench, a 16-scene executable benchmark, and a failure-driven self-evolution mechanism that turns memory failures into gated JSON-DSL assets promoted only to later evaluation splits. Experiments report that ABot-AgentOS improves task success rate and goal completion over a single-controller baseline on an unspecified subset of the new benchmark, and report strong memory results on LoCoMo, OpenEQA, Mem-Gallery, NExT-QA, and EgoLife, with further gains from self-evolution.

Significance. If the central claims hold, the system would be a meaningful step toward general embodied agent runtimes: the OS-style separation of planning, skill execution, and verification is plausible, and the source-grounded graph memory with auditable retrieval traces is a useful design direction. The self-evolution protocol deserves credit for explicitly preventing same-split leakage, for requiring gated validation and regression checks, and for documenting rejected candidate assets in Appendix B. The benchmark, if released, could fill a real gap in executable multi-scene embodied evaluation. However, the paper's headline embodied-execution result is currently supported only by a small, undisclosed subset with no statistical characterization, and the memory results rely on judge protocols whose leniency is not calibrated against official metrics. The architectural ideas are promising, but the empirical evidence as presented is not yet sufficient to establish the claimed improvements.

major comments (3)
  1. [§5.1.1, Table 1, §6] The central claim that ABot-AgentOS improves long-horizon embodied execution over a single-controller baseline rests entirely on Table 1, which reports three aggregate numbers (TSR 49.97/61.96/68.18; GCR 57.95/68.79/74.62) with no number of tasks, no episode count, no variance or confidence intervals, no per-scene breakdown, and no released benchmark or evaluation code. Section 5.1.1 states the results are on 'a subset of the current benchmark' and are 'intended as an initial system validation rather than a complete benchmark leaderboard'; Section 6 defers full benchmark release to future work. With an unknown and possibly small task sample, the 11.99-point TSR improvement could be noise, task-selective, or an artifact of how the subset was chosen. This is the paper's headline claim, so the evidence is currently unfalsifiable from the manuscript alone.
  2. [§5.2.2, Appendix A.3] The LoCoMo result of 87.5 — presented as 'approaching the human overall score of 87.9' and beating Mem0 by 1.9 — is obtained with the judge prompt in Appendix A.3, which is substantially more lenient than the standard LoCoMo evaluation. Rule 1 grants CORRECT for any one correct item from a multi-item gold list; Rule 4 tolerates dates within 14 days and durations within 50%; Rule 6 accepts answers that merely mention the same named entity; and the optional evidence variant only loosens acceptance further. Even if all baselines were scored with the same prompt, the absolute scores are not comparable to published LoCoMo numbers, and the near-human claim is not supported. The authors should state whether this is exactly the Mem0 judge protocol or a modification, report agreement with official LoCoMo labels on a sample, and give the score delta under the official judge.
  3. [§5.2.3, Eqs. (2), (6), Appendix B] The lifelong self-evolution claim rests on point improvements (NExT-QA +4.1, LoCoMo +1.2, OpenEQA +1.2, Mem-Gallery +0.4, EgoLife +0.8) with no number of questions per split, no confidence intervals, and no significance tests. Appendix B shows that in the OpenEQA trace most proposed assets are rejected or later deprecated, so the reported net gains may not be statistically distinguishable from noise. In addition, the retrieval score Eq. (2) depends on four unspecified weights (λ_sem, λ_lex, λ_meta, λ_type), and the asset gate Eq. (6) depends on τ_gain and τ_reg; no values or sensitivity analyses are given. Without split sizes, variance, per-split results, and the exact hyperparameters, the claim that assets 'improve later splits' is not testable.
minor comments (5)
  1. [§2.3.4] The sentence 'No evo-asset generated from split t is used during inference on split t; accepted assets are promoted only for later splits' appears twice nearly verbatim in this subsection; the duplication should be removed.
  2. [Figure 6 caption] Typo: 'Self-evoluation Meta-judge' should be 'Self-evolution Meta-judge'.
  3. [Table 3] The ABot-AgentOS Static rows are misformatted: '862.852.3 59.2' and '24 61.955.7 59.9' should be split into separate columns (frames 8/24, ScanNet 62.8/61.9, HM3D 52.3/55.7, Overall 59.2/59.9).
  4. [§3.3.2] The semantic-condition evaluator uses an LLM judge whose rubric is not validated for inter-annotator agreement or calibration against human verdicts; a short analysis of judge reliability would strengthen the benchmark's trace-grounded scoring claim.
  5. [§5.2.1] The statement that 'the most controlled comparison is therefore between reproduced ABot-AgentOS Static and ABot-AgentOS + Self-evo runs' is useful, but the OpenEQA setup uses GPT-5.4 as writer, answerer, and judge; the possibility of judge self-agreement inflating scores should be discussed.

Circularity Check

0 steps flagged

No significant circularity: central claims are external benchmark measurements and a no-leakage self-evolution protocol; no prediction reduces to its inputs by construction.

full rationale

The paper is primarily an empirical systems report rather than a derivation chain. Its headline execution claim (Section 5.1, Table 1) is a measured comparison on an EmbodiedWorldBench subset, with the paper itself cautioning that the results are 'intended as an initial system validation rather than a complete benchmark leaderboard' (Section 5.1.1). No equation in the paper transforms an input into these outputs: Eq. (2) is a retrieval scoring definition and Eq. (6) is an asset-acceptance rule; neither produces the reported TSR/GCR or memory scores by construction. The self-evolution loop is explicitly split-wise: failures are diagnosed and gated only after a split is evaluated, and 'No evo-asset generated from split t is used during inference on split t' (Section 2.3.4). Appendix B documents rejected candidates and stack-level rollback, so the reported later-split gains are out-of-sample measurements rather than fitted predictions. Memory results are evaluated against external benchmarks (LoCoMo, OpenEQA, Mem-Gallery, NExT-QA, EgoLife) under a fixed retrieval backbone and fixed judge protocols; even where the same model writes and judges, the protocol is held constant across compared methods. Author self-citations such as ABot-N0 and ABot-M0 appear as background/interface references, not as load-bearing support for the paper's central claims. Section 6's limitations concern real-world validation, benchmark release, and perception noise, which are evidence-quality issues, not circular reductions. Concerns about undisclosed hyperparameters, subset size, and missing confidence intervals are statistical robustness concerns, not circularity. No circular step could be identified by quoting a specific reduction of a claimed result to its own input.

Axiom & Free-Parameter Ledger

5 free parameters · 5 axioms · 2 invented entities

The central results depend on undisclosed weights and thresholds in retrieval, gating, and rewards, on the assumed fidelity of perception, information isolation, and LLM judges, and on the authors' enforcement of the no-leakage self-evolution protocol.

free parameters (5)
  • Retrieval similarity weights λ_sem, λ_lex, λ_meta, λ_type = not disclosed
    Eq. (2) defines hybrid seed selection as a weighted sum; the weights must be hand-tuned or fitted, and their values are not reported.
  • Gate thresholds τ_gain, τ_reg = not disclosed
    Eq. (6) acceptance criterion requires ΔS_target ≥ τ_gain and ΔS_reg ≥ -τ_reg; thresholds are not reported.
  • Meta-Judge weights w_k and threshold θ = not disclosed
    Eqs. (10)-(11) aggregate five dimensions with weights and threshold; no values reported.
  • Episode reward weights λ_eff, λ_cons, λ_comp = not disclosed
    Eq. (8) defines episode reward as weighted combination; no values reported.
  • LoCoMo judge tolerances = 14 days; 50% duration; 1-of-N partial credit
    Appendix A.3 sets date tolerance to 14 days, duration tolerance to 50%, and partial credit to any single correct item; these calibrated tolerances inflate absolute recall scores.
axioms (5)
  • domain assumption The VLM observation tool (Qwen3-VL-Plus) produces text descriptions faithful to the visual scene.
    Section 5.1.1 says visual observations are obtained through the VLM tool; Section 5.1.3 concedes perception errors.
  • domain assumption The filtered semantic map given to the agent does not leak hidden task states.
    Section 3.2.2 describes a 'strict information visibility boundary' but provides no leakage audit results.
  • domain assumption LLM judges for EmbodiedWorldBench semantic conditions and for memory benchmarks are unbiased and accurate.
    Section 3.3.2 relies on an LLM judge for semantic success; Appendix A.3's lenient rules suggest evaluation-bias risk.
  • domain assumption The memory-writer adapters faithfully convert raw sensory/dialogue streams into typed graph records without systematic loss.
    Section 2.3.2 describes graph construction but no writer accuracy measurement is reported.
  • ad hoc to paper Self-evolution assets are promoted only to later splits and never affect the split they are diagnosed on.
    Sections 2.3.4 and 5.2.3 assert this no-leakage constraint; it is enforced by the authors' own bookkeeping and cannot be externally verified.
invented entities (2)
  • evo-assets no independent evidence
    purpose: Gated, versioned JSON-DSL runtime policies that encode improvements to memory writing, retrieval, frame selection, or answering across splits
    They are software artifacts with no falsifiable handle outside the paper; acceptance is judged by the authors' own gate with undisclosed thresholds.
  • Universal Multi-modal Graph Memory no independent evidence
    purpose: Typed source-grounded graph replacing raw transcripts/videos for persistent memory
    A system component, not a physically testable entity; its 'universal' scope is asserted, not demonstrated beyond five benchmarks.

pith-pipeline@v1.3.0-alltime-deepseek · 35189 in / 14555 out tokens · 140526 ms · 2026-08-02T07:16:09.121070+00:00 · methodology

0 comments
read the original abstract

Recent VLM and VLA systems have improved robotic perception and action prediction, yet long-horizon embodied agents still require a general runtime layer for reasoning, memory, tool use, verification, and cross-embodiment execution. We present ABot-AgentOS, a general robotic Agent Operating System that sits above low-level controllers and provides a deliberative agent layer for scene-conditioned planning, context-isolated skill execution, multi-stage verification, multi-modal memory, and edge-cloud collaboration. To evaluate such systems, we introduce EmbodiedWorldBench, an executable benchmark with 16 indoor, outdoor, and hybrid scenes, four difficulty levels, and over 200 tasks involving navigation, object search, NPC dialogue, dynamic events, and trace-grounded scoring. ABot-AgentOS further introduces Universal Multi-modal Graph Memory, a persistent source-grounded substrate that converts dialogue, visual observations, spatial context, temporal relations, and task traces into typed nodes and edges. A failure-driven self-evolution loop converts diagnosed memory failures into gated runtime evo-assets that are promoted only to later evaluation splits, preventing current-split ground-truth leakage while enabling continual improvement. On an initial EmbodiedWorldBench subset, ABot-AgentOS improves over a single-controller baseline in both task success and goal completion. Across memory benchmarks, ABot-AgentOS Static achieves 87.5 on LoCoMo, 59.9 on OpenEQA EM-EQA, 88.6 on Mem-Gallery, and 76.5 Acc@All on NExT-QA; self-evolution further improves LoCoMo to 88.7, OpenEQA to 60.4, and Mem-Gallery to 89.0. These results suggest that a general Agent OS layer can improve long-horizon embodied execution while providing persistent, auditable memory for continual interaction.

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

131 extracted references · 61 linked inside Pith

  1. [1]

    Michael Ahn, Anthony Brohan, Noah Brown, Yevgen Chebotar, Omar Cortes, Byron David, Chelsea Finn, Chuyuan Fu, Keerthana Gopalakrishnan, Karol Hausman, Alex Herzog, Daniel Ho, Jasmine Hsu, Julian Ibarz, Brian Ichter, Alex Irpan, Eric Jang, Rosario Jauregui Ruano, Kyle Jeffrey, Sally Jesmonth, Nikhil J. Joshi, Ryan Julian, Dmitry Kalashnikov, Yuheng Kuang, ...

    Pith/arXiv arXiv 2022

  2. [2]

    Graphpad: Inference- time 3d scene graph updates for embodied question answering.arXiv preprint arXiv:2506.01174, 2025

    Muhammad Qasim Ali, Saeejith Nair, Alexander Wong, Yuchen Cui, and Yuhao Chen. Graphpad: Inference- time 3d scene graph updates for embodied question answering.arXiv preprint arXiv:2506.01174, 2025. URL https://arxiv.org/abs/2506.01174

    Pith/arXiv arXiv 2025

  3. [3]

    Introducing computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haiku.https://www.anthropic

    Anthropic. Introducing computer use, a new Claude 3.5 Sonnet, and Claude 3.5 Haiku.https://www.anthropic. com/news/3-5-models-and-computer-use, 2024. Accessed: 2026-06-26

    2024

  4. [4]

    Advisor tool

    Anthropic. Advisor tool. https://platform.claude.com/docs/en/agents-and-tools/tool-use/ advisor-tool, 2026. Accessed: 2026-06-26

    2026

  5. [5]

    Eve-agent: Evidence-verifiable self-evolving agents

    Yamato Arai and Yuma Ichikawa. Eve-agent: Evidence-verifiable self-evolving agents. arXiv preprint arXiv:2605.22905, 2026

    Pith/arXiv arXiv 2026

  6. [6]

    Glance and focus: Memory prompting for multi-event video question answering

    Ziyi Bai, Ruiping Wang, and Xilin CHEN. Glance and focus: Memory prompting for multi-event video question answering. In Thirty-seventh Conference on Neural Information Processing Systems, 2023. URL https://openreview.net/forum?id=J6Niv3yrMq

    2023

  7. [7]

    Mem-gallery: Benchmarking multimodal long-term conversational memory for mllm agents

    Yuanchen Bei, Tianxin Wei, Xuying Ning, Yanjun Zhao, Zhining Liu, Xiao Lin, Yada Zhu, Hendrik Hamann, Jingrui He, and Hanghang Tong. Mem-gallery: Benchmarking multimodal long-term conversational memory for mllm agents. arXiv preprint arXiv:2601.03515, 2026

    arXiv 2026

  8. [8]

    Revisiting the “Video” in Video-Language Understanding

    Shyamal Buch, Cristobal Eyzaguirre, Adrien Gaidon, Jiajun Wu, Li Fei-Fei, and Juan Carlos Niebles. Revisiting the “Video” in Video-Language Understanding. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022

    2022

  9. [9]

    Remember me, refine me: A dynamic procedural memory framework for experience-driven agent evolution.arXiv preprint arXiv:2512.10696, 2025

    Zouying Cao, Jiaji Deng, Li Yu, Weikang Zhou, Zhaoyang Liu, Bolin Ding, and Hai Zhao. Remember me, refine me: A dynamic procedural memory framework for experience-driven agent evolution.arXiv preprint arXiv:2512.10696, 2025

    Pith/arXiv arXiv 2025

  10. [10]

    Astranav-world: World model for foresight control and consistency, 2025

    Jintao Chen, Junjun Hu, Haochen Bai, Minghua Luo, Xinda Xue, Botao Ren, Chengyu Bai, Shichao Xie, Ziyi Chen, Fei Liu, et al. Astranav-world: World model for foresight control and consistency, 2025

    2025

  11. [11]

    FrugalGPT: How to use large language models while reducing cost and improving performance.Transactionson Machine Learning Research, 2024

    Lingjiao Chen, Matei Zaharia, and James Zou. FrugalGPT: How to use large language models while reducing cost and improving performance.Transactionson Machine Learning Research, 2024. URLhttps://openreview. net/forum?id=cSimKw5p6R

    2024

  12. [12]

    Wenhu Chen, Hexiang Hu, Xi Chen, Pat Verga, and William W. Cohen. Murag: Multimodal retrieval-augmented generator for open question answering over images and text. InProceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 5558–5570. Association for Computational Linguistics, 2022. URLhttps://aclanthology.org/2022.emnl...

    2022

  13. [13]

    Explore like humans: Autonomous exploration with online sg-memo construction for embodied agents,

    Xu Chen, Shichao Xie, Zhining Gu, Lu Jia, Minghua Luo, Fei Liu, Zedong Chu, Yanfen Shen, Xiaolong Wu, and Mu Xu. Explore like humans: Autonomous exploration with online sg-memo construction for embodied agents,

  14. [14]

    Socialnav: Training human-inspired foundation model for socially-aware embodied navigation

    Ziyi Chen, Yingnan Guo, Zedong Chu, Minghua Luo, Yanfen Shen, Mingchao Sun, Junjun Hu, Shichao Xie, Yang Kuan, Pei Shi, et al. Socialnav: Training human-inspired foundation model for socially-aware embodied navigation. pages 28796–28806, 2026

    2026

  15. [15]

    Mem0: Building production- ready ai agents with scalable long-term memory.arXiv preprint arXiv:2504.19413, 2025

    Prateek Chhikara, Dev Khant, Saket Aryan, Taranjeet Singh, and Deshraj Yadav. Mem0: Building production- ready ai agents with scalable long-term memory.arXiv preprint arXiv:2504.19413, 2025

    Pith/arXiv arXiv 2025

  16. [16]

    Graphvideoagent: Enhancing long-form video understanding with entity relation graphs

    Meng Chu, Yicong Li, and Tat-Seng Chua. Graphvideoagent: Enhancing long-form video understanding with entity relation graphs. InProceedings of the 33rd ACM International Conference on Multimedia, MM ’25, page 30 4639–4648, New York, NY, USA, 2025. Association for Computing Machinery. ISBN 9798400720352. doi: 10.1145/3746027.3755537. URLhttps://doi.org/10....

    arXiv 2025

  17. [18]

    Abot-n0: Technical report on the vla foundation model for versatile embodied navigation

    Zedong Chu, Shichao Xie, Xiaolong Wu, Yanfen Shen, Minghua Luo, Zhengbo Wang, Fei Liu, Xiaoxu Leng, Junjun Hu, Mingyang Yin, et al. Abot-n0: Technical report on the vla foundation model for versatile embodied navigation. arXiv preprint arXiv:2602.11598, 2026

    arXiv 2026

  18. [19]

    Videoagent: A memory-augmented multimodal agent for video understanding

    Yue Fan, Xiaojian Ma, Rujie Wu, Yuntao Du, Jiaqi Li, Zhi Gao, and Qing Li. Videoagent: A memory-augmented multimodal agent for video understanding. InEuropean Conference on Computer Vision, pages 75–92. Springer, 2025

    2025

  19. [20]

    Group-in-group policy optimization for llm agent training

    Lang Feng, Zhenghai Xue, Tingcong Liu, and Bo An. Group-in-group policy optimization for llm agent training. arXiv preprint arXiv:2505.10978, 2025. URLhttps://arxiv.org/abs/2505.10978

    Pith/arXiv arXiv 2025

  20. [21]

    Neural graph memory: A structured approach to long-term memory in multimodal agents

    Matthew Fisher. Neural graph memory: A structured approach to long-term memory in multimodal agents. Zenodo preprint, 2025. URLhttps://zenodo.org/records/16809331

    arXiv 2025

  21. [22]

    EmbodiedCity: A benchmark platform for embodied agent in real-world city environment.arXiv preprint arXiv:2410.09604, 2024

    Chen Gao, Baining Zhao, Weichen Zhang, Jinzhu Mao, Jun Zhang, Zhiheng Zheng, Fanhang Man, Jianjie Fang, Zile Zhou, Jinqiang Cui, Xinlei Chen, and Yong Li. EmbodiedCity: A benchmark platform for embodied agent in real-world city environment.arXiv preprint arXiv:2410.09604, 2024

    Pith/arXiv arXiv 2024

  22. [23]

    Assistgpt: A general multi-modal assistant that can plan, execute, inspect, and learn, 2023

    Difei Gao, Lei Ji, Luowei Zhou, Kevin Qinghong Lin, Joya Chen, Zihan Fan, and Mike Zheng Shou. Assistgpt: A general multi-modal assistant that can plan, execute, inspect, and learn, 2023. URLhttps://arxiv.org/abs/ 2306.08640

    Pith/arXiv arXiv 2023

  23. [24]

    Mist : Multi-modal iterative spatial- temporal transformer for long-form video question answering

    Difei Gao, Luowei Zhou, Lei Ji, Linchao Zhu, Yi Yang, and Mike Zheng Shou. Mist : Multi-modal iterative spatial- temporal transformer for long-form video question answering. In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14773–14783, 2023. doi: 10.1109/CVPR52729.2023.01419

    arXiv 2023

  24. [25]

    A survey of self-evolving agents: What, when, how, and where to evolve on the path to artificial super intelligence.arXiv preprint arXiv:2507.21046, 2025

    Huan-ang Gao, Jiayi Geng, Wenyue Hua, Mengkang Hu, Xinzhe Juan, Hongzhang Liu, Shilong Liu, Jiahao Qiu, Xuan Qi, Yiran Wu, et al. A survey of self-evolving agents: What, when, how, and where to evolve on the path to artificial super intelligence.arXiv preprint arXiv:2507.21046, 2025

    Pith/arXiv arXiv 2025

  25. [26]

    Openfly: A comprehensive platform for aerial vision-language navigation

    Yunpeng Gao, Chenhui Li, Zhongrui You, Junli Liu, Zhen Li, Pengan Chen, Qizhi Chen, Zhonghan Tang, Liansheng Wang, Penghui Yang, et al. Openfly: A comprehensive platform for aerial vision-language navigation. arXiv preprint arXiv:2502.18041, 2025

    arXiv 2025

  26. [27]

    Poinav: Benchmarking and enhancing final-meters arrival in real-world vision-language navigation, 2026

    Ruiyan Gong, Meisheng Zhang, Yuxiang Zhao, Mingchao Sun, Yanfen Shen, Zedong Chu, Zhining Gu, Wei Guo, Xiaolong Cheng, Qiming Li, Kangning Niu, Yanqing Zhu, Xiaolong Wu, Tianlun Li, and Mu Xu. Poinav: Benchmarking and enhancing final-meters arrival in real-world vision-language navigation, 2026. URLhttps: //arxiv.org/abs/2605.28237

    Pith/arXiv arXiv 2026

  27. [28]

    Google gemini 2.5 pro

    Google. Google gemini 2.5 pro. https://docs.cloud.google.com/gemini-enterprise-agent-platform/ models/gemini/2-5-pro?hl=zh-cn, 2025. Gemini Enterprise Agent Platform

    2025

  28. [29]

    Gemini Robotics brings AI into the physical world

    Google DeepMind. Gemini Robotics brings AI into the physical world. https://deepmind.google/blog/ gemini-robotics-brings-ai-into-the-physical-world/, 2025. Accessed: 2026-06-26

    2025

  29. [30]

    Swe-trace: Optimizing long-horizon swe agents through rubric process reward models and heuristic test-time scaling.arXiv preprint arXiv:2604.14820, 2026

    Hao Han, Jin Xie, Xuehao Ma, et al. Swe-trace: Optimizing long-horizon swe agents through rubric process reward models and heuristic test-time scaling.arXiv preprint arXiv:2604.14820, 2026

    Pith/arXiv arXiv 2026

  30. [31]

    Learning to evolve: A self-improving framework for multi-agent systems via textual parameter graph optimization.arXiv preprint arXiv:2604.20714, 2026

    Shan He, Runze Wang, Zhuoyun Du, Huiyu Bai, Zouying Cao, Yu Cheng, and Bo Zheng. Learning to evolve: A self-improving framework for multi-agent systems via textual parameter graph optimization.arXiv preprint arXiv:2604.20714, 2026

    Pith/arXiv arXiv 2026

  31. [32]

    Embodied web agents: Bridging physical-digital realms for integrated agent intelligence.arXiv preprint arXiv:2506.15677, 2025

    Yining Hong, Rui Sun, Bingxuan Li, Xingcheng Yao, Maxine Wu, Alexander Chien, Da Yin, Ying Nian Wu, Zhecan James Wang, and Kai-Wei Chang. Embodied web agents: Bridging physical-digital realms for integrated agent intelligence.arXiv preprint arXiv:2506.15677, 2025. 31

    Pith/arXiv arXiv 2025

  32. [33]

    Inner monologue: Embodied reasoning through planning with language models

    Wenlong Huang, Fei Xia, Ted Xiao, Harris Chan, Jacky Liang, Pete Florence, Andy Zeng, Jonathan Tompson, Igor Mordatch, Yevgen Chebotar, Pierre Sermanet, Noah Brown, Tomas Jackson, Linda Luu, Sergey Levine, Karol Hausman, and Brian Ichter. Inner monologue: Embodied reasoning through planning with language models. arXiv preprint arXiv:2207.05608, 2022

    Pith/arXiv arXiv 2022

  33. [34]

    Augustus: An llm-driven multimodal agent system with contextualized user memory

    Jitesh Jain, Shubham Maheshwari, Ning Yu, Wen-mei Hwu, and Humphrey Shi. Augustus: An llm-driven multimodal agent system with contextualized user memory. arXiv preprint arXiv:2510.15261, 2025. URL https://arxiv.org/abs/2510.15261

    arXiv 2025

  34. [35]

    Robobrain: A unified brain model for robotic manipulation from abstract to concrete

    Yuheng Ji, Huajie Tan, Jiayu Shi, Xiaoshuai Hao, Yuan Zhang, Hengyuan Zhang, Pengwei Wang, Mengdi Zhao, Yao Mu, Pengju An, et al. Robobrain: A unified brain model for robotic manipulation from abstract to concrete. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1724–1734, 2025

    2025

  35. [36]

    Galaxea open-world dataset and g0 dual-system vla model.arXiv preprint arXiv:2509.00576, 2025

    Tao Jiang, Tianyuan Yuan, Yicheng Liu, Chenhao Lu, Jianning Cui, Xiao Liu, Shuiqi Cheng, Jiyang Gao, Huazhe Xu, and Hang Zhao. Galaxea open-world dataset and g0 dual-system vla model.arXiv preprint arXiv:2509.00576, 2025

    Pith/arXiv arXiv 2025

  36. [37]

    Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, and Karthik Narasimhan

    Carlos E. Jimenez, John Yang, Alexander Wettig, Shunyu Yao, Kexin Pei, Ofir Press, and Karthik Narasimhan. Swe-bench: Can language models resolve real-world github issues?arXiv preprint arXiv:2310.06770, 2023

    Pith/arXiv arXiv 2023

  37. [38]

    Memory os of ai agent

    Jiazheng Kang, Mingming Ji, Zhe Zhao, and Ting Bai. Memory os of ai agent. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 25961–25970. Association for Computational Linguistics, 2025. URLhttps://aclanthology.org/2025.emnlp-main.1318

    2025

  38. [39]

    Joshi, Hanna Moazam, et al

    Omar Khattab, Arnav Singhvi, Paridhi Maheshwari, Zhiyuan Zhang, Keshav Santhanam, Sri Vardhamanan, Saiful Haq, Ashutosh Sharma, Thomas T. Joshi, Hanna Moazam, et al. Dspy: Compiling declarative language model calls into self-improving pipelines.arXiv preprint arXiv:2310.03714, 2023

    Pith/arXiv arXiv 2023

  39. [40]

    Semi-parametric video-grounded text generation,

    Sungdong Kim, Jin-Hwa Kim, Jiyoung Lee, and Minjoon Seo. Semi-parametric video-grounded text generation,

  40. [41]

    Gaussexplorer: 3d gaussian splatting for embodied exploration and reasoning

    Yu-Ji Kim, Dahye Lee, Jun-Seong Kim, GeonU Kim, Hyeon-Woo Nam, Yongjin Kwon, Yu-Chiang Frank Wang, Jaesung Choe, and Tae-Hyun Oh. Gaussexplorer: 3d gaussian splatting for embodied exploration and reasoning. arXiv preprint arXiv:2601.13132, 2026. URLhttps://arxiv.org/abs/2601.13132

    Pith/arXiv arXiv 2026

  41. [42]

    Citynav: A large-scale dataset for real-world aerial navigation

    Jungdae Lee, Taiki Miyanishi, Shuhei Kurita, Koya Sakamoto, Daichi Azuma, Yutaka Matsuo, and Nakamasa Inoue. Citynav: A large-scale dataset for real-world aerial navigation. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 5912–5922, 2025

    2025

  42. [43]

    Robomemory: A brain-inspired multi-memory agentic framework for interactive environmental learning in physical embodied systems.arXiv preprint arXiv:2508.01415, 2025

    Mingcong Lei, Honghao Cai, Yuyuan Yang, Yimou Wu, Jinke Ren, Zezhou Cui, Liangchen Tan, Junkun Hong, Gehan Hu, Shuangyu Zhu, et al. Robomemory: A brain-inspired multi-memory agentic framework for interactive environmental learning in physical embodied systems.arXiv preprint arXiv:2508.01415, 2025

    arXiv 2025

  43. [44]

    Behavior-1k: A benchmark for embodied ai with 1,000 everyday activities and realistic simulation

    Chengshu Li, Ruohan Zhang, Josiah Wong, Cem Gokmen, Sanjana Srivastava, Roberto Martín-Martín, Chen Wang, Gabrael Levine, Michael Lingelbach, Jiankai Sun, et al. Behavior-1k: A benchmark for embodied ai with 1,000 everyday activities and realistic simulation. pages 80–93, 2023

    2023

  44. [45]

    Leveraging llms as meta-judges: A multi-agent framework for evaluating llm judgments.arXiv preprint arXiv:2504.17087, 2025

    Yuran Li, Jama Hussein Mohamud, Chongren Sun, Di Wu, and Benoit Boulet. Leveraging llms as meta-judges: A multi-agent framework for evaluating llm judgments.arXiv preprint arXiv:2504.17087, 2025

    Pith/arXiv arXiv 2025

  45. [46]

    Evimem: Evidence-gap-driven iterative retrieval for long-term conversational memory.arXiv preprint arXiv:2604.27695, 2026

    Yuyang Li, Yime He, Zeyu Zhang, and Dong Gong. Evimem: Evidence-gap-driven iterative retrieval for long-term conversational memory.arXiv preprint arXiv:2604.27695, 2026

    Pith/arXiv arXiv 2026

  46. [47]

    Memma: Coordinating the memory cycle through multi-agent reasoning and in-situ self-evolution.arXiv preprint arXiv:2603.18718, 2026

    Minhua Lin, Zhiwei Zhang, Hanqing Lu, Hui Liu, Xianfeng Tang, Qi He, Xiang Zhang, and Suhang Wang. Memma: Coordinating the memory cycle through multi-agent reasoning and in-situ self-evolution.arXiv preprint arXiv:2603.18718, 2026

    arXiv 2026

  47. [48]

    Navforesee: A unified vision-language world model for hierarchical planning and dual-horizon navigation prediction

    Fei Liu, Shichao Xie, Minghua Luo, Zedong Chu, Junjun Hu, Xiaolong Wu, and Mu Xu. Navforesee: A unified vision-language world model for hierarchical planning and dual-horizon navigation prediction. pages 32431–32440, 2026

    2026

  48. [49]

    Simplemem: Efficient lifelong memory for llm agents.arXiv preprint arXiv:2601.02553, 2026

    Jiaqi Liu, Yaofeng Su, Peng Xia, Siwei Han, Zeyu Zheng, Cihang Xie, Mingyu Ding, and Huaxiu Yao. Simplemem: Efficient lifelong memory for llm agents.arXiv preprint arXiv:2601.02553, 2026. 32

    Pith/arXiv arXiv 2026

  49. [50]

    Evolvemem: Self-evolving memory architecture via autoresearch for llm agents.arXiv preprint arXiv:2605.13941, 2026

    Jiaqi Liu, Xinyu Ye, Peng Xia, Zeyu Zheng, Cihang Xie, Mingyu Ding, and Huaxiu Yao. Evolvemem: Self-evolving memory architecture via autoresearch for llm agents.arXiv preprint arXiv:2605.13941, 2026

    Pith/arXiv arXiv 2026

  50. [51]

    Wanderland: Geometrically grounded simulation for open-world embodied ai

    Xinhao Liu, Jiaqi Li, Youming Deng, Ruxin Chen, Yingjia Zhang, Yifei Ma, Li Guo, Yiming Li, Jing Zhang, and Chen Feng. Wanderland: Geometrically grounded simulation for open-world embodied ai. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1041–1052, 2026

    2026

  51. [52]

    Evaluating very long-term conversational memory of llm agents.arXiv preprint arXiv:2402.17753, 2024

    Adyasha Maharana, Dong-Ho Lee, Sergey Tulyakov, Mohit Bansal, Francesco Barbieri, and Yuwei Fang. Evaluating very long-term conversational memory of llm agents.arXiv preprint arXiv:2402.17753, 2024

    Pith/arXiv arXiv 2024

  52. [53]

    Openeqa: Embodied question answering in the era of foundation models

    Arjun Majumdar, Anurag Ajay, Xiaohan Zhang, Pranav Putta, Sriram Yenamandra, Mikael Henaff, Sneha Silwal, Paul McVay, Oleksandr Maksymets, Sergio Arnaud, et al. Openeqa: Embodied question answering in the era of foundation models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024

    2024

  53. [55]

    R-eqa: Retrieval-augmented generation for embodied question answering, 2025

    Hyobin Ong and Minsu Jang. R-eqa: Retrieval-augmented generation for embodied question answering, 2025. URL https://embodied-ai.org/papers/2025/6_R_EQA_Retrieval_Augmented_Ge.pdf. Embodied AI Workshop

    2025

  54. [56]

    Gonzalez, M

    Isaac Ong, Amjad Almahairi, Vincent Wu, Wei-Lin Chiang, Tianhao Wu, Joseph E. Gonzalez, M. Waleed Kadous, and Ion Stoica. RouteLLM: Learning to route LLMs with preference data.arXiv preprint arXiv:2406.18665, 2024

    Pith/arXiv arXiv 2024

  55. [57]

    Openai gpt-4.1

    OpenAI. Openai gpt-4.1. https://developers.openai.com/api/docs/models/gpt-4.1, 2025. OpenAI Devel- opers

    2025

  56. [58]

    Introducing ChatGPT agent: Bridging research and action

    OpenAI. Introducing ChatGPT agent: Bridging research and action. https://openai.com/index/ introducing-chatgpt-agent/, 2025. Accessed: 2026-06-26

    2025

  57. [59]

    Introducing Operator

    OpenAI. Introducing Operator. https://openai.com/index/introducing-operator/, 2025. Accessed: 2026- 06-26

    2025

  58. [60]

    Openai gpt-5.4 mini

    OpenAI. Openai gpt-5.4 mini. https://developers.openai.com/api/docs/models/gpt-5.4-mini, 2026. Ope- nAI Developers

    2026

  59. [61]

    Patil, Ion Stoica, and Joseph E

    Charles Packer, Sarah Wooders, Kevin Lin, Vivian Fang, Shishir G. Patil, Ion Stoica, and Joseph E. Gonzalez. Memgpt: Towards llms as operating systems.arXiv preprint arXiv:2310.08560, 2023

    Pith/arXiv arXiv 2023

  60. [62]

    O’Brien, Carrie J

    Joon Sung Park, Joseph C. O’Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. Generative agents: Interactive simulacra of human behavior. InProceedings of the 36th Annual ACM Symposium on User Interface Software and Technology, 2023

    2023

  61. [63]

    Toolllm: Facilitating large language models to master 16000+ real-world apis.arXiv preprint arXiv:2307.16789, 2023

    Yujia Qin, Shihao Liang, Yining Ye, Kunlun Zhu, Lan Yan, Yaxi Lu, Yankai Lin, Xin Cong, Xiangru Tang, Bill Qian, Sihan Zhao, Lauren Hong, Runchu Tian, Ruobing Xie, Jie Zhou, Mark Gerstein, Dahai Li, Zhiyuan Liu, and Maosong Sun. Toolllm: Facilitating large language models to master 16000+ real-world apis.arXiv preprint arXiv:2307.16789, 2023

    Pith/arXiv arXiv 2023

  62. [64]

    Agentic very long video understanding.arXiv preprint arXiv:2601.18157, 2026

    Aniket Rege, Arka Sadhu, Yuliang Li, Kejie Li, Ramya Korlakai Vinayak, Yuning Chai, Yong Jae Lee, and Hyo Jin Kim. Agentic very long video understanding.arXiv preprint arXiv:2601.18157, 2026

    Pith/arXiv arXiv 2026

  63. [65]

    Meminsight: Autonomous memory augmentation for llm agents.arXiv preprint arXiv:2503.21760, 2025

    Rana Salama, Jason Cai, Michelle Yuan, Anna Currey, Monica Sunkara, Yi Zhang, and Yassine Benajiba. Meminsight: Autonomous memory augmentation for llm agents.arXiv preprint arXiv:2503.21760, 2025. URL https://arxiv.org/abs/2503.21760

    Pith/arXiv arXiv 2025

  64. [66]

    Toolformer: Language models can teach themselves to use tools

    Timo Schick, Jane Dwivedi-Yu, Roberto Dessì, Roberta Raileanu, Maria Lomeli, Eric Hambro, Luke Zettlemoyer, Nicola Cancedda, and Thomas Scialom. Toolformer: Language models can teach themselves to use tools. In Advances in Neural Information Processing Systems, 2023. URL https://proceedings.neurips.cc/paper_ files/paper/2023/hash/d842425e4bf79ba039352da0f...

    2023

  65. [67]

    Hi robot: Open-ended instruction following with hierarchical vision-language-action models

    Lucy Xiaoyang Shi, Brian Ichter, Michael Equi, Liyiming Ke, Karl Pertsch, Quan Vuong, James Tanner, Anna Walling, Haohuan Wang, Niccolo Fusai, et al. Hi robot: Open-ended instruction following with hierarchical vision-language-action models. arXiv preprint arXiv:2502.19417, 2025. 33

    Pith/arXiv arXiv 2025

  66. [68]

    Meta-judging with large language models: Concepts, methods, and challenges.arXiv preprint arXiv:2601.17312, 2026

    Hugo Silva, Mateus Mendes, and Hugo Gonçalo Oliveira. Meta-judging with large language models: Concepts, methods, and challenges.arXiv preprint arXiv:2601.17312, 2026

    arXiv 2026

  67. [69]

    Envscaler: Scaling tool-interactive environments for llm agent via programmatic synthesis.arXiv preprint arXiv:2601.05808, 2026

    Xiaoshuai Song, Haofei Chang, Guanting Dong, Yutao Zhu, Zhicheng Dou, and Ji-Rong Wen. Envscaler: Scaling tool-interactive environments for llm agent via programmatic synthesis.arXiv preprint arXiv:2601.05808, 2026

    Pith/arXiv arXiv 2026

  68. [70]

    TongSIM: A general platform for simulating intelligent machines.arXiv preprint arXiv:2512.20206, 2025

    Zhe Sun, Kunlun Wu, Chuanjian Fu, Zeming Song, Langyong Shi, et al. TongSIM: A general platform for simulating intelligent machines.arXiv preprint arXiv:2512.20206, 2025

    arXiv 2025

  69. [72]

    Robobrain 2.0 technical report.arXiv preprint arXiv:2507.02029, 2025

    BAAI RoboBrain Team, Mingyu Cao, Huajie Tan, Yuheng Ji, Xiansheng Chen, Minglan Lin, Zhiyu Li, Zhou Cao, Pengwei Wang, Enshen Zhou, et al. Robobrain 2.0 technical report.arXiv preprint arXiv:2507.02029, 2025

    Pith/arXiv arXiv 2025

  70. [73]

    Gemini robotics 1.5: Pushing the frontier of generalist robots with advanced embodied reasoning, thinking, and motion transfer

    Gemini Robotics Team, Abbas Abdolmaleki, Saminda Abeyruwan, Joshua Ainslie, Jean-Baptiste Alayrac, Montserrat Gonzalez Arenas, Ashwin Balakrishna, Nathan Batchelor, Alex Bewley, Jeff Bingham, et al. Gemini robotics 1.5: Pushing the frontier of generalist robots with advanced embodied reasoning, thinking, and motion transfer. arXiv preprint arXiv:2510.03342, 2025

    Pith/arXiv arXiv 2025

  71. [74]

    Qwen3.5-omni technical report.arXiv preprint arXiv:2604.15804, 2026

    Qwen Team. Qwen3.5-omni technical report.arXiv preprint arXiv:2604.15804, 2026

    Pith/arXiv arXiv 2026

  72. [75]

    Outcome accuracy is not enough: Aligning the reasoning process of reward models

    Binghai Wang, Yantao Liu, Yuxuan Liu, Tianyi Tang, Shenzhi Wang, Chang Gao, Chujie Zheng, Yichang Zhang, Le Yu, Shixuan Liu, et al. Outcome accuracy is not enough: Aligning the reasoning process of reward models. arXiv preprint arXiv:2602.04649, 2026

    arXiv 2026

  73. [76]

    Voyager: An open-ended embodied agent with large language models.arXiv preprint arXiv:2305.16291, 2023

    Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, and Anima Anand- kumar. Voyager: An open-ended embodied agent with large language models.arXiv preprint arXiv:2305.16291, 2023

    Pith/arXiv arXiv 2023

  74. [77]

    Genie: A generalizable navigation system for in-the-wild environments.IEEE Robotics and Automation Letters, 2025

    Jiaming Wang, Diwen Liu, Jizhuo Chen, Jiaxuan Da, Nuowen Qian, Minh Man Tram, and Harold Soh. Genie: A generalizable navigation system for in-the-wild environments.IEEE Robotics and Automation Letters, 2025

    2025

  75. [78]

    ODYSSEY: Open-world quadrupeds exploration and manipulation for long-horizon tasks

    Kaijun Wang, Liqin Lu, Mingyu Liu, Jianuo Jiang, Zeju Li, Bolin Zhang, Wancai Zheng, Xinyi Yu, Hao Chen, and Chunhua Shen. ODYSSEY: Open-world quadrupeds exploration and manipulation for long-horizon tasks. arXiv preprint arXiv:2508.08240, 2025

    Pith/arXiv arXiv 2025

  76. [79]

    Internvideo: General video foundation models via generative and discriminative learning, 2022

    Yi Wang, Kunchang Li, Yizhuo Li, Yinan He, Bingkun Huang, Zhiyu Zhao, Hongjie Zhang, Jilan Xu, Yi Liu, Zun Wang, Sen Xing, Guo Chen, Junting Pan, Jiashuo Yu, Yali Wang, Limin Wang, and Yu Qiao. Internvideo: General video foundation models via generative and discriminative learning, 2022. URLhttps://arxiv.org/ abs/2212.03191

    Pith/arXiv arXiv 2022

  77. [81]

    Ragen: Understanding self-evolution in llm agents via multi-turn reinforcement learning

    Zihan Wang, Kangrui Wang, Qineng Wang, Pingyue Zhang, Linjie Li, Zhengyuan Yang, Kefan Yu, Minh Nhat Nguyen, Licheng Liu, Eli Gottlieb, Monica Lam, Yiping Lu, Kyunghyun Cho, Jiajun Wu, Li Fei-Fei, Lijuan Wang, Yejin Choi, and Manling Li. Ragen: Understanding self-evolution in llm agents via multi-turn reinforcement learning. arXiv preprint arXiv:2504.2007...

    Pith/arXiv arXiv 2025

  78. [82]

    Reinforcing multi-turn reasoning in llm agents via turn-level reward design

    Quan Wei, Siliang Zeng, Chenliang Li, et al. Reinforcing multi-turn reasoning in llm agents via turn-level reward design. arXiv preprint arXiv:2505.11821, 2025

    arXiv 2025

  79. [83]

    Chi, et al

    Tianxin Wei, Noveen Sachdeva, Benjamin Coleman, Zhankui He, Yuanchen Bei, Xuying Ning, Mengting Ai, Yunzhe Li, Jingrui He, Ed H. Chi, et al. Evo-memory: Benchmarking llm agent test-time learning with self-evolving memory.arXiv preprint arXiv:2511.20857, 2025

    Pith/arXiv arXiv 2025

  80. [84]

    Webagent-r1: Training web agents via end-to-end multi-turn reinforcement learning

    Zhepei Wei, Wenlin Yao, Yao Liu, Weizhi Zhang, Qin Lu, Liang Qiu, Changlong Yu, Puyang Xu, Chao Zhang, Bing Yin, Hyokun Yun, and Lihong Li. Webagent-r1: Training web agents via end-to-end multi-turn reinforcement learning. arXiv preprint arXiv:2505.16421, 2025. URLhttps://arxiv.org/abs/2505.16421

    arXiv 2025

Showing first 80 references.