Good Papers: read the good ones

PriorVLA freezes a prior expert and trains an adaptation expert via expert queries to preserve pretrained vision-language-action priors, updating only 25% of full fine-tuning parameters while outperforming baselines on OOD and few-shot robot manipulation.

Xinyu Guo, Bin Xie, Wei Chai, Xianchi Deng and 3 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

16/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 16 of 20 reviewers recommend it

lenient 5/5

medium 8/10

strict 3/5

72%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026Meituan U Hong Kong -kuaishou tech Institute of automation, Chinese Efficient inference & serving

SnapMLA: Efficient Long-Context MLA Decoding via Hardware-Aware FP8 Quantized Pipelining

SnapMLA improves long-context MLA decoding throughput up to 1.91x via hardware-aware FP8 quantization and pipeline optimization while preserving benchmark quality.

Yifan Zhang, Zunhai Su, Shuhao Hu, YangRui and 4 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

8/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 8 of 20 reviewers recommend it

lenient 2/5

medium 6/10

strict 0/5

86%Must read

?Must readVote to see the score

NeurIPS 2026Meituan Fudan Tongji HKUST Institute of automation, Chinese Web & GUI agents

EvoCUA: Evolving Computer Use Agents via Learning from Scalable Synthetic Experience

EvoCUA evolves computer-use agents via synthetic experience loops, achieving 56.7% success on OSWorld to set a new open-source state-of-the-art.

Taofeng Xue, Chong Peng, Mianqiu Huang, Linsen Guo and 11 more

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026 · ▲ 92 on Hugging Face · Code ★ 352

– ReadersNo votes yet

14/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 14 of 20 reviewers recommend it

lenient 4/5

medium 9/10

strict 1/5

80%Must read

?Must readVote to see the score

NeurIPS 2026Institute of automation, Chinese Li Auto Inc.LiAuto LLM reasoning & chain of thought

M2A: Synergizing Mathematical and Agentic Reasoning in Large Language Models

M2A synergizes mathematical and agentic reasoning via parameter-space model merging, improving SWE-Bench Verified solved rates from 44.0% to 51.2% without retraining.

JunJian Wang, Xin Zhou, Qiran Xu, Kun Zhan

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

12/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 12 of 20 reviewers recommend it

lenient 5/5

medium 7/10

strict 0/5

83%Must read

?Must readVote to see the score

NeurIPS 2026U Chinese Academy of Science Meituan Center for Research on Intellige Institute of automation, Chinese U Science and Technology of Chin RL for LLMs

TGRL: Temperature-Grouped Reinforcement Learning for Efficient Exploration in LLMs

TGRL turns temperature-induced rollout diversity into an explicit RLVR training signal via reward contrast and Jensen-Shannon divergence, accelerating convergence up to 36% while improving math, code, and agent benchmarks.

Zihan Lin, Xiaohan Wang, Jie Cao, Jiajun Chai and 3 more

Sydney Poster Session 6, Thu, Dec 10, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026 · ▲ 8 on Hugging Face · Code ★ 3

– ReadersNo votes yet

13/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

83%Must read

?Must readVote to see the score

NeurIPS 2026SUN YAT-SEN UNIVERSITY South China University of Techno Harbin Institute of Technology Institute of automation, Chinese Shanghai Jiaotong Vision-language-action models

RePO-VLA: Recovery-Driven Policy Optimization for Vision-Language-Action Models

RePO-VLA improves vision-language-action robustness by assigning roles to success, recovery, and failure trajectories, raising adversarial success from 20% to 75%.

Weijia Liufu, Xiaoyu Guo, Ruiyi Chen, Jingzhi Liu and 15 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

13/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 13 of 20 reviewers recommend it

lenient 5/5

medium 6/10

strict 2/5

91%Must read

?Must readVote to see the score

NeurIPS 2026Fudan Institute of automation, Chinese Shanghai Artificial Intelligence The Hong Kong Polytechnic UNSW, Sydney Jailbreaks & red teaming

SciHazard: A Benchmark for Measuring Scientific Safety Risks with Decomposed Harm Scoring

SciHazard benchmarks LLM scientific safety risks via decomposed harm scoring across 3,000 real-world grounded queries, finding deep research agents 32.3% more harmful than standard models.

Chunxiao Li, Yuan Xiong, Lijun Li, Tianyi Du and 3 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

17/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

74%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026Beijing Jiaotong Beijing University of Posts and Beijing jiaotong university Kuaishou- 快手科技 National University of Singapore Video-language models

CaC: Advancing Video Reward Models via Hierarchical Spatiotemporal Concentrating

CaC advances video reward models via hierarchical spatiotemporal concentrating, improving fine-grained anomaly accuracy by 25.7% and reducing generated-video anomalies by 11.7%.

Jiyuan Wang, Huan Ouyang, Chunyu Lin, Dewen Fan and 14 more

Sydney Poster Session 1, Tue, Dec 8, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

9/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 9 of 20 reviewers recommend it

lenient 4/5

medium 4/10

strict 1/5

71%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026Southeast nuaa Huawei Technologies Ltd.Nanyang Technological Institute of automation, Chinese LLM agents & planning

PhGPO: Pheromone-Guided Policy Optimization for Long-Horizon Tool Planning

PhGPO learns reusable tool-transition patterns from past trajectories via pheromone guidance to improve long-horizon tool planning.

Yu Li, Guangfeng Cai, shengtian yang, Han Luo and 4 more

Sydney Poster Session 6, Thu, Dec 10, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

6/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 6 of 20 reviewers recommend it

lenient 4/5

medium 2/10

strict 0/5

78%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026Institute of Automation, Chinese Institute of automation, Chinese Controllable generation

TrioPose: Native Triple-Stream Diffusion Transformers for Pose-Guided Text-to-Image Generation

TrioPose uses a triple-stream pose-aware DiT with relational bias masks and spatial loss weighting to generate accurate multi-person images, improving Human-Art AP by 30%.

dian gu, Zhengyi Yang

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

11/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 11 of 20 reviewers recommend it

lenient 3/5

medium 7/10

strict 1/5

80%Must read

?Must readVote to see the score

NeurIPS 2026Tongji The Chinese University of Hong K Institute of automation, Chinese Tianjin Inner Mongolia RL for LLMs

AffectGPT-RL: Revealing Roles of Reinforcement Learning in Open-Vocabulary Emotion Recognition

AffectGPT-RL applies reinforcement learning to open-vocabulary emotion recognition, improving performance by optimizing non-differentiable metrics and enabling reasoning.

Zheng Lian, Fan Zhang, Lan Chen, Yazhou Zhang and 7 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

12/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 12 of 20 reviewers recommend it

lenient 4/5

medium 8/10

strict 0/5

76%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026Institute of automation, Chinese GigaAI Peking ByteDance Inc.Tianjin Long-context modeling

LinearARD: Linear-Memory Attention Distillation for RoPE Restoration

LinearARD restores RoPE-scaled LLMs via linear-memory attention self-distillation, recovering 98.3% short-text performance with 4.25M tokens versus 256M.

Ning Yang, Hengyu Zhong, Wentao Wang, Baoliang Tian and 3 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

10/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

88%Must read

?Must readVote to see the score

NeurIPS 2026Huazhong University of Science a Institute of automation, Chinese Jilin Institute of Automation, Chinese Carnegie Mellon Video understanding

Rethinking the Readout: Unlocking Video Backbones for AI-Generated Video Detection

Standard video backbone readouts suppress patch-level temporal dynamics needed to detect AI-generated videos; a lightweight velocity-gated patch profiling readout reaches 95.28 AUC on frozen backbones.

Manni Cui, Ziheng Qin, ZiAn Wang, Ruiqi Liu and 7 more

Atlanta Poster Session 4, Thu, Dec 10, 4:30 PM–7:30 PM, Hall C1 · Published 2026

– ReadersNo votes yet

15/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 15 of 20 reviewers recommend it

lenient 5/5

medium 9/10

strict 1/5

72%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026Institute of Automation Chinese Tsinghua The University of Edinburgh Institute of automation, Chinese Institute of automation, Chinese Unified multimodal models

OBJECT-UNI: A Unified Model for Object-Centric Spatial Understanding and Controllable Generation

Object-Uni unifies object-centric spatial understanding and controllable generation via explicit pose variables, improving pose reasoning and geometrically consistent synthesis.

Mining Tan, yinuo Wang, Ziqi Zhou, Weize Quan and 5 more

Sydney Poster Session 1, Tue, Dec 8, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

8/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 8 of 20 reviewers recommend it

lenient 4/5

medium 4/10

strict 0/5

92%Must read

?Must readVote to see the score

NeurIPS 2026SpotlightTsinghua Xiaomi Corporation Peking Institute of Automation, Chinese The Hong Kong University of Scie Vision-language-action models

Revisiting Embodied Chain-of-Thought for Generalizable Robot Manipulation

Embodied chain-of-thought improves generalization when grounded in action guidance, and ERVLA uses reasoning-dropout supervision to avoid unstable autoregressive reasoning at inference, achieving state-of-the-art robot manipulation results.

Nan Sun, Yuan Zhang, Yongkun Yang, Wentao Zhao and 9 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

19/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 19 of 20 reviewers recommend it

lenient 5/5

medium 10/10

strict 4/5

72%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026U Science and Technology of Chin Unisound unisound Institute of automation, Chinese Vision-language models

BayesRAG: Probabilistic Mutual Evidence Corroboration for Multimodal Retrieval-Augmented Generation

BayesRAG applies Bayesian inference and Dempster-Shafer theory to fuse multimodal retrieval evidence, improving retrieval confidence via cross-modal mutual corroboration. It significantly outperforms state-of-the-art methods on challenging multimodal benchmarks.

Xuan Li, Yining Wang, Haocai Luo, Shengping Liu and 5 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026 · Code ★ 6

– ReadersNo votes yet

8/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 8 of 20 reviewers recommend it

lenient 4/5

medium 4/10

strict 0/5

78%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026Institute of Automation，Chinese U Science and Technology of Chin Beijing University of Aeronautic Institute of automation, Chinese Zhejiang Video generation

Physics-Informed Video Generation via Mixture-of-Experts Latent Alignment

PILA injects physics-structured latent guidance into frozen video generators via mixture-of-experts alignment, achieving state-of-the-art physical plausibility and visual quality.

Cong Wang, Hanxin Zhu, Jiayi Luo, Yonglin Tian and 5 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

11/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 11 of 20 reviewers recommend it

lenient 3/5

medium 8/10

strict 0/5

71%Highly rated

?Highly ratedVote to see the score

NeurIPS 2026Macau University of Science and Macau University of Science and The University of Queensland National University of Singapore Institute of automation, Chinese 3D vision & reconstruction

Marrying Optimal Transport and ODEs for Unified Continuous-Time 4D Reconstruction and Tracking

Uni4R unifies continuous-time 4D reconstruction and tracking via optimal-transport and ODE velocity fields, achieving state-of-the-art results on both tasks and continuous-time kinematics benchmarks.

Liying Yang, Hao Mo, Jialun Liu, Chen Liu and 6 more

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026

– ReadersNo votes yet

6/20 AI panelreviewers recommend it

Readers and the AI panel: vote on this paper to see what they said.

Only vote on papers you've read. Sign in with GitHub to vote.

AI panel: 6 of 20 reviewers recommend it

lenient 2/5

medium 4/10

strict 0/5