[AI] Stream of Belief: Eliciting Collective Intelligence through Latent Belief Dynamics
Y Yao, J Dong, Y Du [MIT & Harvard & UCLA] (2026)
https://t.co/EysTdhipzP
[AI] DNeo: Your Parallel Draft Model Needs Better Target Feature Encoding
J Sandler, S Bergsma, M Elhoushi, B C Zhang… [University of Virginia & Cerebras Systems] (2026)
https://t.co/GP6D7ExOmi
[AI] The Resolution Hypothesis: Agent Evolution's Key Is Identifying Improvements
Y Zeng, S Lan, A Lu, K Yan… [ShanghaiTech University & ByteDance] (2026)
https://t.co/GHINzYmyyT
[AI] Composable Soft-Prompt Distillation of RL Experts: Toward Continual Post-Training
Z Zhang, C Yang, X He, C Liu, Y Yin... [UC Santa Barbara & CMU & University of California, Santa Cruz] (2026)
https://t.co/E8W4woDe9A
[AI] Weight-Space Programming Doesn't Need Transformers
H Ok, J Lee [Pohang University of Science and Technology (POSTECH)] (2026)
https://t.co/Gp8JXgugRd
[CV] π⁴: A Foundation Model for 4D Scene Geometry and Motion
Y Wang, W Zhao, N Yang, H Guo… [Shanghai Artificial Intelligence Laboratory] (2026)
https://t.co/OD7gukt3Q2
[RO] ScaleVLA: Scaling Vision-Language-Action Models to 552 Billion Parameters
R Yu, D Liu, X Chen, Z Qi… [Tsinghua University & Peking University] (2026)
https://t.co/LgFGZTZdy2
[AI] The Content–Geometry Tradeoff in World Models: When JEPA Learns Latent Maps
M Ren, Y Wang, C Hoang, Y Kuang, Y LeCun [NYU & AMI Labs] (2026)
https://t.co/pUQrwBuyCg
[AI] BuggyMLE: Scaling Machine Learning Engineering Difficulty with Bug Injection
J Wu, R Qiang, Y Zhuang, G Li… [Google Research & Google DeepMind] (2026)
https://t.co/2JRE5lUX9v
[CL] AltRA-Test: An Alternative Annotator Test for LLM Judges Without Shared Annotators
N Omar, R Dror, A Gueta, A Gal… [Technion & University of Haifa & Google Research] (2026)
https://t.co/gnN6JR4AVk
[LG] Revealing Pathway Interactions in Network Computations with Complete Path Contribution
S Liu, Y Faragalla, S Ganguli, S A. Baccus [Stanford University] (2026)
https://t.co/hoaTY3tCfc
[CL] Frontier models need only one text to unmask its human or AI author
A Desai, D Böloni-Turgut, T Brassington, M Choi… [University of Maryland & Simon Fraser University] (2026)
https://t.co/w9TlK0gVRd
[LG] WeightCode: Installing New Capabilities into Neural Networks via Weight-Space Programming
B Li, T Cui, B Zhang, M Li… [University of Macau & The Chinese University of Hong Kong & Sun Yat-sen University] (2026)
https://t.co/2ULyGGeACS
[LG] Toaster: Nonlinear Memory from Cascaded Linear Memories and Target Propagation
Z Li, A Behrouz, Y Deng, P Kacham… [Google Research] (2026)
https://t.co/bLBjobC94l
[AI] In Situ Training of Optical Neural Networks by Coding Agents
I Oguz, C Christian, C Moser, D Psaltis… [Stanford University & Ecole polytechnique fédérale de Lausanne] (2026)
https://t.co/krjLuv1e1t
[LG] Counterfactual World Models: JEPAs That Reason About "What If"
R Balestriero, Y LeCun, K Varaku [Brown University & New York University] (2026)
https://t.co/bwlpQBbCKX