Research

Top Latest Rising Hidden gems

Today This week This month

Open-source gems →Radar →

Hugging Face Daily Papers3d ago

Towards Looped Models Done Right, Part I: Topology, Input Injection, Recurrent-State Design

One Communicator, Three Vendors: Extending the MoA Predictive Theory to a Real NVIDIA+AMD+CPU MPI Ensemble

Apple Machine Learning Research6d ago

Mix, Don't Tune: Bilingual Pre-Training Outperforms Hyperparameter Search in Data-Constrained Settings

Hugging Face Daily Papers4d ago

Explaining the Roles of Experts in Mixture-of-Experts LLMs for Personalization

Hugging Face Daily Papers5d ago

Cortex: Consolidating Context into Parametric Memory

Local proxies for memory-efficient training

Hugging Face Daily Papers5d ago

LatentLoop: Scaling Looped MoE Models with Asymmetric Compute Allocation

Minimax-optimal Scaling Laws in Ridge Regression via Transfer Learning

Data-Efficient Classifier Pretraining Through Diffusion Models

Architecting for Hardware Diversity: Meta's End-to-End Co-Design for High-Performance AI Training on AMD GPUs

What Accuracy Hides In Learned Transformer Programs

How Finite-Rank Bottlenecks Shape the Low-Rank Adaptation Landscape

BMuon: Budgeting Muon's Spectral Geometry for Structured Compressibility

Hugging Face Daily Papers5d ago

Your Transformer Is Secretly Elastic

OrderFlowDiT: Transport-Regularized Diffusion Transformer for Probabilistic Limit-Order Supply Forecasting

Hugging Face Daily Papers5d ago

RepForcing: Representation-Space Initialization for Autoregressive Video Distillation

Hugging Face Daily Papers5d ago

CoTranslator: Reverse to Forward Chain-of-Thought Translator for Reasoning Distillation

Towards Hyperparameter Transfer for Differentially Private Optimization

ToPA: Block-wise Toeplitz Adaptation for Expressive and Efficient Fine-Tuning

Hugging Face Daily Papers5d ago

DAGFormer: Global Topology Prediction for Transformer Cross-Layer Connections

Rethinking Routing Exploration in Modern Mixture-of-Experts

Hugging Face Daily Papers5d ago

VASPO: Self-Distillation with Privileged Value Estimation

Hugging Face Daily Papers5d ago

Glance Forcing: Making Bidirectional Diffusion Autoregressive at a Glance

Priming Base Models for Agentic Post-Training: An Open Midtraining Recipe

Hugging Face Daily Papers5d ago

The COUNT: Counting N-grams Unlocks Scaling on Host Memory and Training-Free Adaptation

Hugging Face Daily Papers5d ago

MorphAct: Robot Policies Morphing with the Worlds They Act In

Ori: Drift-Aware Streaming $k$-PCA for Very Low-Rank LLM Training

Hugging Face Daily Papers6d ago

Tev: Scaling Long-Horizon Agentic Environments with Dense Progress Rewards

Task Structure and Scalability in Multi-Robot Authoring

Vertical Sliced Wasserstein Distances for Meta-Measures

Topology-Inspired Fine-Tuning for Wind-Field Super-Resolution

Compiling Multi-Threshold Operators for Quantized Non-Monotonic and Fine-Granular Activations in Custom-Tailored DNN Accelerators

The Ephemeral Cost of Switching Optimizers Across LLM Training Stages

FRACTAL: A Self-Similar Multi-Agent Orchestration Architecture with Formal Communication, Convergence, and Depth Bounds

When Is Muon Optimal? Radial Fragility and Minimax Geometry

Don't Retrain, Remix: Adaptation of Web Agents using Mixture Hypernetworks

The importance of tensor handling and implementation of physics-informed DeepONets

Personalized Projected Ascent for Federated Stackelberg Optimization

HTN-Orchestrate: A Hierarchical Task Network Framework for Multi-Agent Orchestration

FocusOPD: Structure-Guided On-Policy Distillation for Diffusion Models

Interfaces and Modules: A Scalable Representation for Weight-Space Learning across Architectures

Flat Loss, Evolving Features: Grokking in Modular Addition

Hugging Face Daily Papers4d ago

TIMERSI: Towards Self-Improved Agentic Deployment for Time-Series Foundation Models

Federated Art Chase: Privacy Leakage and Data Poisoning in Federated Fine-Tuning

Pairwise AUC Optimization Needs Corrective Power: A Unified View

Ranking and Partitioning of Attention Heads for Sparse and Continual Reinforcement Learning Fine-tuning

The Wealth of Agents: Scaling Agentic AI Across Developers

A Mixture-of-Experts Framework for Multi-Stream Quickest Change Detection