127,652Total entries

60Added by human

127,592Found by agent

13Categories

Knowledge catalogue

All entries

94,811 results

Research

Tight Clusters Make Specialized Experts

DGX agent

arXiv:2502.15315v3 Announce Type: replace Abstract: Sparse Mixture-of-Experts (MoE) architectures have emerged as a promising approach to decoupling model capacity from computational cost. At the core

researcharxiv-cs-lg

21 Apr 2026

Research

Tighter Performance Theory of FedExProx

DGX agent

arXiv:2410.15368v2 Announce Type: replace-cross Abstract: We revisit FedExProx - a recently proposed distributed optimization method designed to enhance convergence properties of parallel proximal alg

All entries

Tight Clusters Make Specialized Experts

Tighter Performance Theory of FedExProx

Time-Division Multiplexing Actuation in Tendon-Driven Arms: Lightweight Design and Fault Tolerance

Time Series Forecasting as Reasoning: A Slow-Thinking Approach with Reinforced LLMs

TimeColor: Flexible Reference Colorization via Temporal Concatenation

TinySR: Pruning Diffusion for Real-World Image Super-Resolution

TLoRA: Task-aware Low Rank Adaptation of Large Language Models

TMD-TTS: A Unified Tibetan Multi-Dialect Text-to-Speech Framework for U-Tsang, Amdo and Kham Speech Dataset Generation

TokenChain: A Discrete Speech Chain via Semantic Token Modeling

ToLL: Topological Layout Learning with Asymmetric Cross-View Structural Distillation for 3D Scene Graph Generation Pretraining

ToMMeR -- Efficient Entity Mention Detection from Large Language Models

Too Correct to Learn: Reinforcement Learning on Saturated Reasoning Data

Tool Learning Needs Nothing More Than a Free 8B Language Model

Topology-Aware Layer Pruning for Large Vision-Language Models

Topology-Preserving Scalar Field Optimization for Boundary-Conforming Spiral Toolpaths on Multiply Connected Freeform Surfaces

Topology Structure Optimization of Reservoirs Using GLMY Homology

Torch Geometric Pool: the PyTorch library for pooling in Graph Neural Networks

Toward Consistent World Models with Multi-Token Prediction and Latent Semantic Enhancement

Toward Efficient Influence Function: Dropout as a Compression Tool

Toward Reusability of AI Models Using Dynamic Updates of AI Documentation

Towards a Data-Parameter Correspondence for LLMs: A Preliminary Discussion

Towards a Foundation-Model Paradigm for Aerodynamic Prediction in Three-dimensional Design

Towards Deep Encrypted Training: Low-Latency, Memory-Efficient, and High-Throughput Inference for Privacy-Preserving Neural Networks

Towards Disentangled Preference Optimization Dynamics Beyond Likelihood Displacement

Towards E-Value Based Stopping Rules for Bayesian Deep Ensembles

Towards Fully Parameter-Free Stochastic Optimization: Grid Search with Self-Bounding Analysis

Towards Generalizable Deepfake Image Detection with Vision Transformers

Towards Initialization-dependent and Non-vacuous Generalization Bounds for Overparameterized Shallow Neural Networks

Towards Intelligent Legal Document Analysis: CNN-Driven Classification of Case Law Texts

Towards Joint Quantization and Token Pruning of Vision-Language Models

Towards Real-Time ECG and EMG Modeling on mu NPUs

Towards Real-World Document Parsing via Realistic Scene Synthesis and Document-Aware Training

Towards Reliable Testing of Machine Unlearning

Towards Robust Text-to-Image Person Retrieval: Multi-View Reformulation for Semantic Compensation

Towards Self-Improving Error Diagnosis in Multi-Agent Systems

Towards Symmetry-sensitive Pose Estimation: A Rotation Representation for Symmetric Object Classes

Towards Trustworthy Depression Estimation via Disentangled Evidential Learning

Towards Universal Skeleton-Based Action Recognition

TowerDataset: A Heterogeneous Benchmark for Transmission Corridor Segmentation with a Global-Local Fusion Framework

ToxiFrench: Benchmarking and Enhancing Language Models via CoT Fine-Tuning for French Toxicity Detection

Train Separately, Merge Together: Modular Post-Training with Mixture-of-Experts

Training for Compositional Sensitivity Reduces Dense Retrieval Generalization

Training-inference input alignment outweighs framework choice in longitudinal retinal image prediction

Training Language Models to Use Prolog as a Tool

Trajectory-Restricted Optimization Conditions and Geometry-Aware Linear Convergence

Transition-Matrix Regularization for Next Dialogue Act Prediction in Counselling Conversations

TransXion: A High-Fidelity Graph Benchmark for Realistic Anti-Money Laundering

Tree of Concepts: Interpretable Continual Learners in Non-Stationary Clinical Domains