Two Speeds of Learning: A Representation-Readout Decomposition of Grokking and Double Descent
两种学习速度:Grokking 和双下降的表征-读出分解
Chi-Ning Chou, Oscar Uzdelewicz, Neng-Chun Chiu, Yao-Yuan Yang, SueYeon Chung
机构
*
Center for Computational Neuroscience(计算神经科学中心)
;
Flatiron Institute(Flatiron研究所)
;
Department of Physics(物理系)
;
Harvard University(哈佛大学)
;
Kempner Institute(Kempner研究所)
;
New York University(纽约大学)
Accelerating trajectory optimization with Sobolev-trained diffusion policies
基于Sobolev训练的扩散策略加速轨迹优化
Théotime Le Hellard, Franki Nguimatsia Tiofack, Quentin Le Lidec, Justin Carpentier
机构
*
Inria - Département d’Informatique de l’École normale supérieure, PSL Research University(法国国家科学研究中心-巴黎高等师范学院计算机系,PSL研究大学)
;
Courant Institute, New York University(纽约大学Courant研究所)
Prescribe-then-Select: Adaptive Policy Selection for Contextual Stochastic Optimization
先规定后选择:面向情境随机优化的自适应策略选择
Caio de Prospero Iglesias, Kimberly Villalobos Carballo, Dimitris Bertsimas
机构
*
Sloan School of Management(斯隆管理学院)
;
Massachusetts Institute of Technology(麻省理工学院)
;
Tandon School of Engineering(坦多工程学院)
;
New York University(纽约大学)
Reevaluating Policy Gradient Methods for Imperfect-Information Games
重新评估不完美信息博弈的策略梯度方法
Max Rudolph, Nathan Lichtle, Sobhan Mohammadpour, Alexandre Bayen, J. Zico Kolter, Amy Zhang, Gabriele Farina, Eugene Vinitsky, Samuel Sokota
机构
*
University of Texas at Austin(德克萨斯大学奥斯汀分校)
;
University of California, Berkeley(加州大学伯克利分校)
;
Massachusetts Institute of Technology(麻省理工学院)
;
Carnegie Mellon University(卡内基梅隆大学)
;
NYU Tandon School of Engineering(纽约大学坦顿工程学院)
Two-Parameter Flows for Learning Population Dynamics of Physical Systems
用于学习物理系统群体动力学的双参数流
Paul Schwerdtner, Tobias Blickhan, Benjamin Peherstorfer
机构
*
Courant Institute of Mathematical Sciences, New York University, 251 Mercer Street, New York, NY 10012, USA(数学科学学院,纽约大学,251 Mercer Street,纽约,NY 10012,美国)
To model human linguistic prediction, make LLMs less superhuman
为了模拟人类语言预测,让大语言模型不那么超人类
Byung-Doh Oh, Tal Linzen
机构
*
Division of Linguistics and Multilingual Studies, Nanyang Technological University, Singapore(南洋理工大学语言学与多语言研究系,新加坡)
;
Department of Linguistics, New York University, New York, USA(纽约大学语言学系,纽约,美国)
;
Center for Data Science, New York University, New York, USA(纽约大学数据科学中心,纽约,美国)
DiscoverPhysics: Benchmarking LLMs for Out-of-the-Box Scientific Thinking
DiscoverPhysics: 基准测试LLMs的即用型科学思维
Matt L. Wiemann, Lindsay M. Smith, Peter Melchior, Siddharth Mishra-Sharma, Andrew Gordon Wilson, Pavel Izmailov, Carolina Cuesta-Lázaro
机构
*
Princeton University(普林斯顿大学)
;
Boston University(波士顿大学)
;
New York University(纽约大学)
;
Flatiron Institute(Flatiron研究所)
;
Institute for Advanced Studies(高级研究研究所)
Optimal and Order-optimal Gated Priority-based Greedy Policies for Two-layer Multi-item Order Fulfillment
两层多物品订单履约的最优和阶最优门控优先级贪婪策略
Xi Chen, Yuze Chen, Ziyi Chen, Yuan Zhou
机构
*
Leonard N. Stern School of Business, New York University(纽约大学 Leonard N. Stern 商学院)
;
Qiuzhen College, Tsinghua University(清华大学邱泽学院)
;
Yau Mathematical Sciences Center & Department of Mathematical Sciences, Tsinghua University(清华大学姚数学科学中心及数学科学系)
Comments8 pages, 15 figures, 1 table, Accepted at the IEEE Robotics and Automation Letters (RA-L) and the IEEE International Conference on Robotics and Automation (ICRA), 2026
Journal refIEEE Robotics and Automation Letters, vol. 11, no. 3, pp. 3582-3589, March 2026
Hybrid least squares for learning functions from highly noisy data
混合最小二乘法:从高噪声数据中学习函数
Ben Adcock, Bernhard Hientzsch, Akil Narayan, Yiming Xu
机构
*
Department of Mathematics, Simon Fraser University(Simon Fraser大学数学系)
;
Courant Institute of Mathematical Sciences, New York University(纽约大学Courant数学科学研究所)
;
Scientific Computing and Imaging Institute, University of Utah(犹他大学科学计算与成像研究所)
;
Department of Mathematics, University of Kentucky(肯塔基大学数学系)