Good Papers

NeurIPS 2026 orals

Best rated first.

92%Must read
?Must readVote to see the score
NeurIPS 2026OralProgramma LabsNew YorkFacebookFAIR, Meta AIMeta FAIRAgent benchmarks & environments

Computer Use at the Edge of the Statistical Precipice

A 1MB replay script outperforms frontier agents on static benchmarks because of flawed environment design and evaluation; the paper proposes PRISM principles, DigiWorld, and hierarchical bootstrap aggregation to fix both.

Pierluca D Oro, Sneha Silwal, William R Wong, Yuxuan Sun and 5 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 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
91%Must read
?Must readVote to see the score
NeurIPS 2026OralNVIDIA | Università di Modena e NVIDIADepartment of Computer Science, ETH Zurich3D vision & reconstruction

Déjà View: Looping Transformers for Multi-View 3D Reconstruction

DéjàView loops a single transformer block for iterative multi-view 3D reconstruction, matching larger feed-forward models with far fewer parameters while treating refinement steps as an inference-time compute knob.

Alessandro Burzio, Tobias Fischer, Sven Elflein, Qunjie Zhou and 8 more

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

– ReadersNo votes yet
18/20 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: 18 of 20 reviewers recommend it
lenient 5/5
medium 10/10
strict 3/5
91%Must read
?Must readVote to see the score
NeurIPS 2026OralMITMassachusetts Institute of TechnU WyomingRostockFondazione Edmund MachSpatio-temporal forecasting

MOVEBENCH: A Benchmark for Global-Scale Wildlife Movement Forecasting

MoveBench introduces a 2.6M-location wildlife movement forecasting benchmark across 110 species and finds existing methods generalize poorly to unseen individuals and deep learning does not consistently beat simpler baselines.

Justin Kay, Shir Bar, Ellen O Aikens, Martin Becker and 27 more

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

100% Readers1 of 1 upvoted
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
91%Must read
?Must readVote to see the score
NeurIPS 2026OralKing's College London, UnivU Oxford & IDAIDepartment of Computer Science, King's College LondonU SussexJailbreaks & red teaming

The Attacker in the Mirror: Breaking Self-Consistency in Safety via Anchored Bipolicy Self-Play

Anchored Bipolicy Self-Play uses frozen-base LoRA adapters to separate attacker and defender roles, preventing self-consistency collapse and improving safety with 100x greater parameter efficiency.

Gabriele La Malfa, Emanuele La Malfa, Saar Cohen, Jie Zhang 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.

AI panel: 17 of 20 reviewers recommend it
lenient 5/5
medium 10/10
strict 2/5
91%Must read
?Must readVote to see the score
NeurIPS 2026OralU California, BerkeleyTsinghuaTsinghua University, TsinghuaPurdueU WashingtonAgent benchmarks & environments

MLS-Bench: A Holistic and Rigorous Assessment of AI Systems on Building Better AI

MLS-Bench evaluates AI agents on inventing scalable ML methods across 140 tasks, finding current systems fail to reliably surpass human-designed approaches due to insufficient scientific validation insight.

Bohan Lyu, Yucheng Yang, Siqiao Huang, Jiaru Zhang and 24 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 8 on Hugging Face · Code ★ 120

– 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.

AI panel: 17 of 20 reviewers recommend it
lenient 5/5
medium 8/10
strict 4/5
91%Must read
?Must readVote to see the score
NeurIPS 2026OralNanjing University of AeronauticZhejiangZhejiang LabJailbreaks & red teaming

LITMUS: Benchmarking Behavioral Jailbreaks of LLM Agents in Real OS Environments

LITMUS benchmarks LLM agent behavioral jailbreaks in real OS environments, revealing agents execute 40.64% of high-risk operations despite refusals and suffer pervasive execution hallucination.

Chiyu Zhang, Huiqin Yang, Bendong Jiang, Xiaolei Zhang and 7 more

Paris Poster Session 6, Fri, Dec 11, 2:30 PM–4:30 PM, Paris Poster Hall · 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.

AI panel: 17 of 20 reviewers recommend it
lenient 5/5
medium 8/10
strict 4/5
91%Must read
?Must readVote to see the score
NeurIPS 2026OralStanfordZhejiangThe University of QueenslandVision-language models

SALART-VQA: Diagnosing Whether VLMs Understand Salient Artifacts in Generated Images

SalArt-VQA evaluates VLM artifact understanding via fine-grained questions, revealing high detection recall but only 53% fully correct reasoning and a sensitivity-calibration tradeoff.

Xiaoxiao Sun, Ruotian Zhang, Junzhe Huang, James Burgess and 1 more

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026 · ▲ 1 on Hugging Face

– 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.

AI panel: 17 of 20 reviewers recommend it
lenient 5/5
medium 8/10
strict 4/5
91%Must read
?Must readVote to see the score
NeurIPS 2026OralMeta FAIRMetaFacebookLLM reasoning & chain of thought

Quantized Reasoning Models Think They Need to Think Longer, but They Do Not

Post-training quantization of reasoning models increases chain-of-thought length and overthinking errors without improving accuracy, yet penalizing overthinking markers reduces reasoning cost and fixes failures.

Sanae Lotfi, Polina Kirichenko, Steven Li, Zechun Liu

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

– 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.

AI panel: 17 of 20 reviewers recommend it
lenient 5/5
medium 9/10
strict 3/5
89%Must read
?Must readVote to see the score
NeurIPS 2026OralJohns HopkinsVideo understanding

Behavioral Geometric Supervision Aligns Video Foundation Models with Human Social Perception

Behavioral geometric supervision aligns video foundation models with human social perception by matching embedding geometry to human similarity judgments, nearly tripling V-JEPA 2.1 performance past language baselines and developing interpretable social attributes with zero-shot transfer.

Kathy Garcia, Leyla Isik

Sydney Poster Session 4, Wed, Dec 9, 5:00 PM–8:00 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 4/5
medium 9/10
strict 3/5
89%Must read
?Must readVote to see the score
NeurIPS 2026OralAnthropicMATSStanfordGoogle DeepMindAI oversight & deception

Neural Chameleons: Language Models Can Learn to Hide Their Thoughts from Unseen Activation Monitors

Fine-tuned LLMs learn to selectively hide internal representations from unseen activation monitors via low-dimensional subspace manipulation, evading even post-hoc safety probes with modest capability loss.

Max McGuinness, Alex Serrano Terre, Luke Bailey, Scott Emmons

Paris Poster Session 5, Fri, Dec 11, 11:30 AM–1:30 PM, Paris Poster Hall · 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 9/10
strict 2/5
88%Must read
?Must readVote to see the score
NeurIPS 2026OralU TorontoGenomics & single-cell

C3P: Contrastive promoter-protein pretraining yields representations capturing bacterial gene regulation

C3P uses contrastive promoter-protein pretraining to learn bacterial promoter representations that outperform genome language models on regulatory inference and zero-shot co-regulated gene retrieval.

Cameron Dufault, Scott Xu, Alan Moses

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

100% Readers1 of 1 upvoted
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 5/5
medium 8/10
strict 1/5
88%Must read
?Must readVote to see the score
NeurIPS 2026OralInstitute of Computing TechnologInstitute of Computing, Chinese Pruning & sparsity

Compact SO(3) Equivariant Atomistic Foundation Models via Structural Pruning

Structural pruning of SO(3) equivariant atomistic models reduces inference cost while outperforming small from-scratch models on accuracy and compute.

Chen Wang, Siyu Hu, Guangming Tan, Weile Jia

Sydney Poster Session 3, Wed, Dec 9, 10:00 AM–1:00 PM, Hall 1-4 · 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 8/10
strict 2/5
88%Must read
?Must readVote to see the score
NeurIPS 2026OralPrincetonMSLMathematical reasoning

Escaping the Cognitive Well: Efficient Competition Math with Off-the-Shelf Models

An inference pipeline using off-the-shelf models and conjecture extraction with context detachment achieves state-of-the-art IMO-style math performance at much lower cost by escaping the Cognitive Well.

Xingyu Dang, Rohit Agarwal, Rodrigo Porto, Anirudh Goyal and 2 more

Atlanta Poster Session 6, Fri, Dec 11, 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 4/5
medium 9/10
strict 2/5
88%Must read
?Must readVote to see the score
NeurIPS 2026OralKTH Royal Institute of TechnologMcGillKarolinska InstituteU CambridgeUnified multimodal models

MindAlign: Bridging EEG, Vision, and Language for Zero-Shot Visual Decoding

MindAlign aligns EEG, vision, and language via tri-modal contrastive learning to achieve 54.1% zero-shot visual decoding accuracy on Things-EEG2.

Zexuan Chen, Sichao Liu, Runhao Lu, Huichao Qi and 3 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · 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 4/5
medium 8/10
strict 3/5
88%Must read
?Must readVote to see the score
NeurIPS 2026OralU Wisconsin - MadisonU Wisconsin, MadisonStanfordLLM evaluation & benchmarks

Fisher-R1: Training LLM Agents for Reliable Hypothesis Testing

P-Bench reveals LLM agents make subtle inferential errors in hypothesis testing, and Fisher-R1 improves reliability via reinforcement learning to outperform GPT-5.4 and DeepSeek-V4-Pro.

Jiacheng Miao, Jin Mu, Guanhua Chen, James Zou

Atlanta Poster Session 6, Fri, Dec 11, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 1 on Hugging Face

– 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 8/10
strict 2/5
88%Must read
?Must readVote to see the score
NeurIPS 2026OralCalifornia Institute of TechnoloCaltechMulti-agent RL

Training Generalizable Collaborative Agents via Strategic Risk Aversion

Strategic risk aversion acts as an inductive bias for generalizable collaboration, yielding robust multi-agent policies with reduced free-riding and stronger equilibrium outcomes alongside unseen partners.

Chengrui Qu, Yizhou Zhang, Nicolas Lanzetti, Eric Mazumdar

Atlanta Poster Session 6, Fri, Dec 11, 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
86%Must read
?Must readVote to see the score
NeurIPS 2026OralHarvardMITHarvard / MSRLLM pretraining & scaling laws

Bridging Compute- and Data-Optimal Pretraining

A unified compute-data scaling framework introduces token effectiveness to bridge data-rich and fixed-corpus regimes, showing diminishing returns and three operational regimes that render classical compute-optimal allocations suboptimal.

Tian Qin, Kimia Hamidieh, David Alvarez-Melis

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

– 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 5/5
medium 8/10
strict 1/5
86%Must read
?Must readVote to see the score
NeurIPS 2026OralU WashingtonThe University of AlabamaNational University of SingaporeWestlakeDrexelNavigation & SLAM

Language-Conditioned World Modeling for Visual Navigation

LCVN introduces a language-conditioned navigation benchmark and compares diffusion-based latent imagination against unified autoregressive prediction for embodied agents. Latent imagination yields more temporally coherent rollouts, while unified prediction generalizes better to unseen environments.

Yifei Dong, Fengyi Wu, Yilong Dai, Lingdong Kong and 7 more

Atlanta Poster Session 5, Fri, Dec 11, 10:00 AM–1:00 PM, Hall C1 · Published 2026

– 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 7/10
strict 3/5
86%Must read
?Must readVote to see the score
NeurIPS 2026OralTsinghuaTsinghua University, ChinaLancasterVision-language-action models

PhaseLoRA: Control-Regime-Conditioned Low-Rank Adaptation for Continuous-Action Vision-Language-Action Policies

PhaseLoRA conditions LoRA updates on control-regime descriptors at each step, improving continuous-action VLA success by 12.2 points over high-rank baselines.

Yufei Guo, Yinan Wu, Haoran Duan, guiguang ding and 1 more

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

– 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 5/5
medium 8/10
strict 1/5
86%Must read
?Must readVote to see the score
NeurIPS 2026OralKAISTKorea Advanced Institute of ScieKorea Advanced Institute of ScieAmazonMulti-agent LLM systems

ExComm: Exploration-Stage Communication for Error-Resilient Agentic Test-Time Scaling

ExComm detects cross-agent factual conflicts during exploration to correct errors via soft belief updates and trajectory diversification, improving test-time scaling by up to 5.7%.

Woomin Song, Beomjun Kim, Daewon Choi, Sai Muralidhar Jayanthi and 3 more

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

– 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 10/10
strict 0/5
86%Must read
?Must readVote to see the score
NeurIPS 2026OralTexas TechExploration

Emergence of Physical Intelligence via Controllable Information Production

Controllable Information Production grounds intrinsic motivation in dynamics and control, unifying them to measure controllable information production and outperforming prior methods on robot tasks.

Tristan Shah, Stas Tiomkin

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

– 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
86%Must read
?Must readVote to see the score
NeurIPS 2026OralNational University of Defense TNanyang TechnologicalEmbodied agents

Resilience Matters for Embodied Agents System: New Metrics, Systematic Evaluation, and Optimization

Resilience metrics for embodied agents expose hidden recovery and stability differences beyond success rates, guiding deployment-specific optimization.

Yapeng Liu, Yuanzhao Zhai, Xudong Gong, Feng Dawei and 3 more

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

– 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 5/5
medium 9/10
strict 0/5
86%Must read
?Must readVote to see the score
NeurIPS 2026OralUC San DiegoU TorontoNational University of SingaporeSichuanU VirginiaSynthetic data

DD-Ranking: Rethinking the Evaluation of Dataset Distillation

DD-Ranking reveals dataset distillation gains come from extra evaluation techniques rather than image quality, proposing fair metrics to assess true synthetic dataset value.

Zekai Li, Xinhao Zhong, Samir Khaki, Zhiyuan Liang and 36 more

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · Published 2026

– 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 5/5
medium 8/10
strict 1/5
86%Must read
?Must readVote to see the score
NeurIPS 2026OralU Texas at AustinThe University of Texas at AustiDiffusion models

Query Lower Bounds for Diffusion Sampling

Diffusion sampling requires $\tilde\Omega(\sqrt{d})$ adaptive score queries for $d$-dimensional distributions with polynomial accuracy, proving multiscale schedules are necessary.

Zhiyang Xun, Eric Price

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

– 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 6/10
strict 4/5
83%Must read
?Must readVote to see the score
NeurIPS 2026OralThe Chinese University of Hong KSenseTime ResearchSensetimeUCLA | Tencent HunyuanUniversiti MalayaMultimodal reasoning

GeoSym127K: Scalable Symbolically-verifiable Synthesis for Multimodal Geometric Reasoning

GeoSym Engine automates symbolically-verifiable geometric reasoning data synthesis, and models trained on GeoSym127K achieve large gains on diagram-dependent geometry benchmarks.

Jinhao Jing, Zheng Ma, Jinwei Liang, Qiannian Zhao and 8 more

Sydney Poster Session 1, Tue, Dec 8, 10:00 AM–1:00 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 4/5
medium 9/10
strict 0/5
83%Must read
?Must readVote to see the score
NeurIPS 2026OralHebrew University of JerusalemDiffusion models

Colored Noise Diffusion Sampling

Colored Noise Sampling replaces uniform noise with dynamic frequency-dependent schedules to exploit diffusion spectral bias, substantially reducing FID across architectures as a training-free plug-in.

Hadar Davidson, Noam Issachar, Sagie Benaim

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026 · ▲ 19 on Hugging Face · Code ★ 45

– 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 3/5
medium 9/10
strict 1/5
83%Must read
?Must readVote to see the score
NeurIPS 2026OralMassachusetts Institute of TechnNew YorkOptimization

The Stability of Online Algorithms in Performative Prediction

No-regret online algorithms unconditionally converge to mixed performatively stable equilibria via randomized martingale arguments, avoiding distribution-response assumptions and PPAD hardness.

Gabriele Farina, Juan C Perdomo

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · 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 3/5
medium 6/10
strict 4/5
80%Must read
?Must readVote to see the score
NeurIPS 2026OralU GöttingenPrincetonMax Planck Institute for BiologiDeep RL

Unifying Goal-Conditioned RL and Unsupervised Skill Learning via Control-Maximization

Unifying GCRL and MISL as control maximization reveals formulation-specific bounds linking diverse pretraining skills to downstream goal sensitivity.

Alireza Modirshanechi, Benjamin Eysenbach, Peter Dayan, Eric Schulz

Paris Poster Session 2, Wed, Dec 9, 5:00 PM–7:00 PM, Paris Poster Hall · 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 2/5
medium 7/10
strict 3/5
80%Must read
?Must readVote to see the score
NeurIPS 2026OralDrexelMichigan StateU North Carolina at Chapel HillImage recognition

Non-Colliding Biometric Identities for Digital Entities: Geometry, Capacity, and Million-Scale Virtual Identity Provisioning

Biometric identity provisioning allocates virtual face identities as non-colliding gaps within the real identity manifold, scaling to ten million embeddings and one million photorealistic images via gap-aware generation.

Yuyang Ji, Yixuan Shen, Anil Jain, Xiaoming Liu and 1 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · 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 6/10
strict 2/5
80%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralU California, DavisBayesian & probabilistic methods

Exact Gaussian Moment Matching for Residual Networks: a Second-Order Method

Exact Gaussian moment matching propagates mean and covariance through residual networks with exact nonlinear layer formulas, cutting KL divergence errors by orders of magnitude versus approximate methods.

Simon Kuang, Xinfan Lin

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

100% Readers1 of 1 upvoted
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.

AI panel: 10 of 20 reviewers recommend it
lenient 2/5
medium 6/10
strict 2/5
78%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralU CambridgeSouthern University of Science aTsinghua University, TsinghuaIMATI-CNRNanjing3D vision & reconstruction

SAM 3D Animal: Promptable Animal 3D Reconstruction from Images in the Wild

SAM 3D Animal is a promptable framework using the SMAL+ model and Herd3D dataset to reconstruct multiple 3D animals from single images with keypoint and mask prompts, achieving state-of-the-art results.

Xuyi Hu, Jin Lyu, Jiuming Liu, Yebin Liu and 3 more

Sydney Poster Session 5, Thu, Dec 10, 10:00 AM–1:00 PM, Hall 1-4 · Published 2026 · ▲ 3 on Hugging Face · Code ★ 13

– 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 4/5
medium 6/10
strict 1/5
78%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralKTH Royal Institute of TechnologJohns HopkinsWhiting School of EngineeringGraph neural networks

Graph Cascades: Contagion-Based Mesoscopic Rewiring for Structure-Aware Graph Machine Learning

Graph Cascades uses contagion diffusion to build auxiliary edges in linear time, boosting GNN and graph transformer accuracy on heterophilic and high-degree graphs while failing on regular low-degree graphs.

Meher Chaitanya Pindiprolu, My Le, Luana Ruiz

Sydney Poster Session 4, Wed, Dec 9, 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 6/10
strict 2/5
78%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralShanghai Jiao TongHuazhong University of Science aTsinghua University, TsinghuaU California, San DiegoCarnegie MellonUnified multimodal models

LatentUM: Unleashing the Potential of Interleaved Cross-Modal Reasoning via a Latent-Space Unified Model

LatentUM unifies modalities in a shared latent space to enable efficient interleaved cross-modal reasoning and generation, achieving state-of-the-art visual planning and self-reflective generation results.

Jiachun Jin, Zetong Zhou, Xiao Yang, Hao Zhang and 3 more

Sydney Poster Session 2, Tue, Dec 8, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026 · ▲ 29 on Hugging Face · Code ★ 63

– 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 5/5
medium 5/10
strict 1/5
78%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralSorbonne University; ENS, ParisSorbonne Université - Faculté deSorbonne UniversitéBayesian & probabilistic methods

A multi-scale information geometry reveals the structure of mutual information in neural populations

A multi-scale Riemannian geometry extending Fisher information relates metric structure to mutual information and reveals visual cortex encoding features.

Simone Azeglio, Steeve Laquitaine, Ulisse Ferrari, Matthew Chalk

Sydney Poster Session 1, Tue, Dec 8, 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 4/5
medium 6/10
strict 1/5
78%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralTel AvivLightricksAdobe ResearchTel-AvivDeep RL

ParetoSlider: Diffusion Models Post-Training for Continuous Reward Control

ParetoSlider trains one diffusion model with continuous preference weights to approximate the full Pareto front, enabling inference-time navigation of trade-offs between conflicting generative goals without retraining.

Shelly Golan, Michael Finkelson, Ariel Bereslavsky, Yotam Nitzan and 1 more

Sydney Poster Session 6, Thu, Dec 10, 5:00 PM–8:00 PM, Hall 1-4 · Published 2026 · ▲ 15 on Hugging Face

– 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 5/5
medium 6/10
strict 0/5
78%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralQiuzhen College, TsinghuaTsinghuaCausal inference

Impossibility of Distribution-Free Predictive Inference for Individual Treatment Effects

Distribution-free predictive inference for individual treatment effects with continuous covariates requires trivial infinite-length prediction sets.

Chongguang Tao, Zheng Zhou, Yuhong Yang

Sydney Poster Session 5, Thu, Dec 10, 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 2/5
medium 5/10
strict 4/5
78%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralUniversity College LondonU California, RiversideGooglePrivacy

CLIOPATRA: Extracting Private Information from LLM Insights

CLIOPATRA attacks privacy-preserving LLM insight platforms with malicious chats to leak target medical histories with nearly 100% precision in 65% of cases, showing layered heuristic protections are insufficient.

Meenatchi Sundaram Muthu Selva Annamalai, Emiliano De Cristofaro, Peter Kairouz

Sydney Poster Session 4, Wed, Dec 9, 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 5/5
medium 5/10
strict 1/5
76%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralU Massachusetts at AmherstShanghaiTechUMassGoogle Deepmind / HarvardIBMSim-to-real

SyncWorld: Visual Calibration Enables World Models as Zero-Shot Simulators

SyncWorld learns action-visual mappings via visual calibration episodes to serve as zero-shot simulators across unseen robot settings without retraining.

Yuncong Yang, Zhengtao Han, Furkan Ozyurt, Zeyuan Yang and 5 more

Atlanta Poster Session 2, Wed, Dec 9, 4:30 PM–7:30 PM, Hall C1 · Published 2026 · ▲ 16 on Hugging Face · Code ★ 21

– 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.

AI panel: 10 of 20 reviewers recommend it
lenient 5/5
medium 5/10
strict 0/5
74%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralSouthern University of Science aShenzhensouthern university of science aDiffusion models

BrainWorld: A Structural-Prior-Conditioned Generative Model for Whole-Brain 4D fMRI Dynamics

BrainWorld uses anatomical MRI to condition whole-brain 4D fMRI generation, producing stable long-horizon trajectories and improving downstream performance via augmentation and transferable representations.

Junfeng Xia, Wenhao Ye, Junxiang Zhang, Xuanye Pan and 2 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 5/5
medium 4/10
strict 0/5
74%Highly rated
?Highly ratedVote to see the score
NeurIPS 2026OralThe Hong Kong University of ScieEindhoven University of TechnoloHong Kong University of Science 3D vision & reconstruction

TransmissiveGS: Residual-Guided Disentangled Gaussian Splatting for Transmissive Scene Reconstruction and Rendering

TransmissiveGS disentangles reflective and transmissive components via dual Gaussian representations and residual-guided multi-view separation for transmissive scene reconstruction.

Zhenyu Liang, Xiao Zhang, Tianchao Li, Jack C Cheng and 1 more

Atlanta Poster Session 1, Wed, Dec 9, 10:00 AM–1:00 PM, Hall C1 · 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 5/10
strict 0/5
Show 40 more papers