Concentration

Follow this topic Ask about this topic

Momentum

6 papers in the last four weeks, against 2 the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 26

All topics

Does a Shared Temperature Imply a Shared Angular Scale in Probabilistic Contrastive Learning?
Sep 30, 2026Ningkang Peng, Qianfeng Yu, Jingyang Mao +4Contrastive Learning Cosine Similarity
Nanjing Normal University · Nanjing University of Chinese Medicine · Tohoku University
A Concentration Bound for Two-Timescale Actor-Critic Algorithm
Sep 24, 2026Prashansa Panda, Shalabh BhatnagarSoft Actor-Critic Finite Time Analysis
Department of Computer Science and Automation IISc Bangalore
Understanding Hyperspherical Geometry of ECAPA-TDNN Embedding and Its Impact on Zero-Shot Voice Conversion
Sep 21, 2026Mathilde Abrassart, Nicolas Obin, Axel RoebelVoice Conversion Automatic Speaker Verification
STMS Lab, IRCAM CNRS, Sorbonne Universit´e Paris, France
When Confidence Signals Disagree: Local and Global Confidence in Autoregressive Language Models
Sep 15, 2026Julio C. Amador Diaz LopezConfidence Calibration Autoregressive Language Models
SiftyML, LTD London, United Kingdom
From Learned-Mode AV-Traffic Pairing to Planner Decisions: A Marginal-Preserving Study on Argoverse 2
Sep 14, 2026Jingyu WangPlanning Governed
Independent Researcher.
A Dominant Supplier Slows Recursive Drift More Than It Steers It
Sep 10, 2026Yangze Liu, Zhongyi HanCollapse Ecosystems
Shandong University
ConCA: Concentration-Aware Channel Attention for Fine-Grained Visual Recognition
Aug 31, 2026Yu-Sheng Liu, Yu-Chen TungConvolutional Block Attention Module Fine-Grained Perception
Department of Physics, National Kaohsiung Normal University, Kaohsiung, 824, Taiwan
A concentration result for multilayer feedforward neural networks
Aug 15, 2026Vera KoponenNeural Network Multilayer Perceptrons
Sharp Root Anti-Concentration via Projective Incidence and Ordered Root Laws
Aug 3, 2026Zijun Wang, Yuchen Miao, Yifan Hu +1Concentration Bounds \Widetilde{\Mathcal{O}}(\Sqrt{T})$ Regret
Northeastern University, China · Beijing Jiaotong University, China · Xiangtan University, China
ReCo: Reweighting GRPO Against Distributional Concentration
Jul 29, 2026Junoh Park, Junseo Hwang, Wonguk Cho +1Reweighting Mathematical Reasoning Benchmarks
Graduate School of Data Science, Seoul National University
When Shippers Become Algorithms: Candidate Exposure, Information Design, and the Concentration of LLM-Mediated Freight Markets
Jul 22, 2026Takahiro Ezaki, Naoto Imura, Katsuhiro NishinariLogistics Agentic Commerce
Research Center for Advanced Science and Technology, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo, 153-8904, Japan · aResearch Center for Advanced Science and Technology, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo, 153-8904, Japan · bDepartment of Aeronautics and Astronautics, School of Engineering, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-8656, Japan
Countercurrent Multiplier Networks: A Renal-Inspired Iterative Operator with Provably Bounded Fixed-Point Dynamics
Jul 21, 2026Snigdha Chandan KhilarFixed-Point Iteration Chronic Kidney Disease
Independent Researcher
Concentration and Mean-Square Bounds for Contractive Stochastic Approximation: A Unified Elementary Approach
Jul 20, 2026Siddharth ChandakStochastic Approximation Concentration Bounds
Department of Electrical Engineering, Stanford University, Stanford, 94305, CA, USA
ConCent: Contact-Centric Real-to-Sim-to-Real Learning from One Demonstration
Jun 29, 2026Heecheol Kim, Namiko Saito, Katsushi Ikeuchi +1Sim-To-Real Reinforcement Learning Sim-To-Real Gap
Microsoft Research Asia - Tokyo · The University of Tokyo
Are Time-Series Foundation Models Ready for E-Nose Data? An Empirical Assessment of Their Embeddings
Jun 26, 2026Taeyeong Choi, Mohammed KamruzzamanTime Series Foundation Models Timestep Embeddings
Department of Information Technology Kennesaw State University Marietta, GA 30060, USA · Department of Agricultural and Biological Engineering University of Illinois Urbana-Champaign Urbana, IL 61801, USA
When Top-1 Fails: Calibrating LoRA Monitors for Masked Diffusion LMs
Jun 23, 2026Lucky Verma, Pratik YadavDiffusion Language Models Masked Diffusion Language Models
Independent Researcher · University of Maryland, Baltimore County
Who Owns the AI Recommendation? A Multi-Industry Empirical Map of Brand Category Ownership Across Large Language Models
Jun 22, 2026Dmitrij ŻatuchinBrand-As-Memory Recommendation Quality
Department of Information Technologies, Estonian Entrepreneurship University of Applied Sciences (EUAS), Tallinn, Estonia. · Rankfor.AI, Tallinn, Estonia.
Wall Shear Stress Reconstruction from Concentration: Differentiable Physics and Physics-Informed Neural Networks
Jun 4, 2026Mahmoud Elhadidy, Siva Viknesh, Roshan M. D'Souza +1Velocity Field Differentiable Physics
Department of Mechanical Engineering, University of Utah, Salt Lake City, UT, USA · Scientific Computing and Imaging Institute, University of Utah, Salt Lake City, UT, USA · Department of Mechanical Engineering, University of Wisconsin–Milwaukee, Milwaukee, WI, USA +1
From Mean-Field Limits to Semiclassical Concentration: Global Convergence of the Canonical Evolutionary Strategy
May 19, 2026Matías Neto, Nicolás Garay, Luis Martí +1Evolutionary Optimization Convergence
Inria Chile Research Center, Av Apoquindo 2827, Las Condes, RM, Chile · Departamento de Ingeniería Matemática, Universidad de Chile, Santiago, Chile
CaC: Advancing Video Reward Models via Hierarchical Spatiotemporal Concentrating
May 12, 2026Jiyuan Wang, Huan Ouyang, Jiuzhou Lin +15Training-Free Video Anomaly Detection Dense Temporal Annotation
BJTU · NTU · BUPT +1
Quantifying Concentration Phenomena of Mean-Field Transformers in the Low-Temperature Regime
May 11, 2026Albert Alcalde, Leon Bungert, Konstantin Riedl +1Transformer Encoder Transformer Architectures
Department of Mathematics Friedrich-Alexander-University Erlangen-Nürnberg · Institute of Mathematics CAIDAS University of Würzburg · Mathematical Institute University of Oxford Reuben College +1
Architecture, Not Scale: Circuit Localization in Large Language Models
May 9, 2026Sohan VenkateshMechanistic Interpretability Model Size
Manipal Institute of Technology Bengaluru.
Matrix-Decoupled Concentration for Autoregressive Sequences: Dimension-Free Guarantees for Sparse Long-Context Rewards
May 7, 2026Pei-Sen LiAutoregressive Language Models Autoregressive Model
School of Mathematics and Statistics, Beijing Institute of Technology, Beijing 100872, China
Artificial Jagged Intelligence as Uneven Optimization Energy Allocation Capability Concentration, Redistribution, and Optimization Governance
May 2, 2026Wesley Shu, Peng WeiArtificial Intelligence Systems Concentration
The Institute of Energetic Paradigm · College of Plant Protection Southwest University, China
Preserving Disagreement: Architectural Heterogeneity and Coherence Validation in Multi-Agent Policy Simulation
Apr 29, 2026Ariel SelaMulti-Agent Simulations Disagreement
Tel Aviv University
Multi-Branch Non-Homogeneous Image Dehazing via Concentration Partitioning and Image Fusion
Apr 27, 2026Yingming Zhang, Wuqi Su, Qing Xiao +1Single Image Dehazing Visible Image Fusion
School of Software and Internet of Things Engineering, Zhejiang Gongshang University · School of Computer Science and Technology, Tiangong University

Sep 30, 2026, cs.LGJ/K move · Enter open · S save

Does a Shared Temperature Imply a Shared Angular Scale in Probabilistic Contrastive Learning?

Ningkang Peng, Qianfeng Yu, Jingyang Mao, Xiaoqian Peng+3

Nanjing Normal University · Nanjing University of Chinese Medicine · Tohoku University

Contrastive Learning Cosine Similarity Gumbel-Softmax Relaxation Concentration Scaling

Open paper AskPDF arXiv

In probabilistic contrastive learning, a shared temperature is commonly interpreted as a shared similarity scale, but this interpretation does not hold for high-dimensional distributional class representations. We study the exact von Mises-Fisher (vMF) probabilistic score used by ProCo when representation dimension and class concentration grow jointly. We prove that the score retains a class-dependent leading angular gain

g_c=A_c/τ

, where

A_c

is the mean resultant length. This gain enters Softmax competition, pairwise decision boundaries, and feature gradients. On real CIFAR-LT, ImageNet-LT, and iNaturalist representations, the theory accurately predicts boundary movements and local gradient changes under the full vMF score. Classwise temperature adjustment also changes the cosine-zero intercept and finite-dimensional response. We construct intercept-preserving and Pure Angular controls to separate the leading gain from these accompanying changes. Complete gain equalization yields a shared-scale cosine prototype rule at leading order; a finite-dimensional margin condition guarantees agreement of the two classifiers. Across 16 frozen representation settings, prediction agreement is 98.43-99.99%, with disagreements concentrated at small cosine margins. In controlled contrastive-only training with the training-frequency prior, Pure Angular editing improves both learned representations at all tested CIFAR-10/100 imbalance factors and retains positive changes on ImageNet-LT. Thus vMF concentration not only describes class distributions, but also forms a decision and learning scale in high-dimensional probabilistic contrastive learning.

Momentum

Related topics

Latest papers 26