AI研究论文每日精选

Is Preference Alignment Always the Best Option to Enhance LLM-Based Translation? An Empirical Analysis

HarmoniCa: Harmonizing Training and Inference for Better Feature Cache in Diffusion Transformer Acceleration

FactAlign: Long-form Factuality Alignment of Large Language Models

Quantifying Generalization Complexity for Large Language Models

Selective Aggregation for Low-Rank Adaptation in Federated Learning

General Preference Modeling with Preference Representations for Aligning Language Models

RATIONALYST: Pre-training Process-Supervision for Improving Reasoning

Closed-loop Long-horizon Robotic Planning via Equilibrium Sequence Modeling

LEOPARD : A Vision Language Model For Text-Rich Multi-Image Tasks

HelpSteer2-Preference: Complementing Ratings with Preferences

E.T. Bench: Towards Open-Ended Event-Level Video-Language Understanding

MOSEL: 950,000 Hours of Speech Data for Open-Source Speech Foundation Model Training on EU Languages

SonicSim: A customizable simulation platform for speech processing in moving sound source scenarios

From Code to Correctness: Closing the Last Mile of Code Generation with Hierarchical Debugging

InfiniPot: Infinite Context Processing on Memory-Constrained LLMs

EmoKnob: Enhance Voice Cloning with Fine-Grained Emotion Control

Not All LLM Reasoners Are Created Equal

VLMGuard: Defending VLMs against Malicious Prompts via Unlabeled Data

Old Optimizer, New Norm: An Anthology

PHI-S: Distribution Balancing for Label-Free Multi-Teacher Distillation

3DGS-DET: Empower 3D Gaussian Splatting with Boundary Guidance and Box-Focused Sampling for 3D Object Detection

ComfyGen: Prompt-Adaptive Workflows for Text-to-Image Generation

BordIRlines: A Dataset for Evaluating Cross-lingual Retrieval-Augmented Generation

EVER: Exact Volumetric Ellipsoid Rendering for Real-time View Synthesis