Cover Image

Angles Don't Lie: Unlocking Training-Efficient RL Through the Model's Own Signals

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Wang, Qinsi, Ke, Jinghan, Ye, Hancheng, Lin, Yueqian, Fu, Yuzhe, Zhang, Jianyi, Keutzer, Kurt, Xu, Chenfeng, Chen, Yiran
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Dobi-SVD: Differentiable SVD for LLM Compression and Some New Perspectives
by: Wang, Qinsi, et al.
Published: (2025)

FlashFPS: Efficient Farthest Point Sampling for Large-Scale Point Clouds via Pruning and Caching
by: Fu, Yuzhe, et al.
Published: (2026)

KVCOMM: Online Cross-context KV-cache Communication for Efficient LLM-based Multi-agent Systems
by: Ye, Hancheng, et al.
Published: (2025)

HippoMM: Hippocampal-inspired Multimodal Memory for Long Audiovisual Event Understanding
by: Lin, Yueqian, et al.
Published: (2025)

CoreMatching: A Co-adaptive Sparse Inference Framework with Token and Neuron Pruning for Comprehensive Acceleration of Vision-Language Models
by: Wang, Qinsi, et al.
Published: (2025)

CoreInfer: Accelerating Large Language Model Inference with Semantics-Inspired Adaptive Sparse Activation
by: Wang, Qinsi, et al.
Published: (2024)

Keyframe-oriented Vision Token Pruning: Enhancing Efficiency of Large Vision Language Models on Long-Form Video Processing
by: Liu, Yudong, et al.
Published: (2025)

SpeechPrune: Context-aware Token Pruning for Speech Information Retrieval
by: Lin, Yueqian, et al.
Published: (2024)

T2S-Bench & Structure-of-Thought: Benchmarking and Prompting Comprehensive Text-to-Structure Reasoning
by: Wang, Qinsi, et al.
Published: (2026)

Don't flatten, tokenize! Unlocking the key to SoftMoE's efficacy in deep RL
by: Sokar, Ghada, et al.
Published: (2024)

Immiscible Diffusion: Accelerating Diffusion Training with Noise Assignment
by: Li, Yiheng, et al.
Published: (2024)

Improved Immiscible Diffusion: Accelerate Diffusion Training by Reducing Its Miscibility
by: Li, Yiheng, et al.
Published: (2025)

Benchmarking is Broken -- Don't Let AI be its Own Judge
by: Cheng, Zerui, et al.
Published: (2025)

Don't Forget the Nonlinearity: Unlocking Activation Functions in Efficient Fine-Tuning
by: Yin, Bo, et al.
Published: (2025)

s3: You Don't Need That Much Data to Train a Search Agent via RL
by: Jiang, Pengcheng, et al.
Published: (2025)

FlashSVD: Memory-Efficient Inference with Streaming for Low-Rank Models
by: Shao, Zishan, et al.
Published: (2025)

AsyncVoice Agent: Real-Time Explanation for LLM Planning and Reasoning
by: Lin, Yueqian, et al.
Published: (2025)

Lips Don't Lie: A Case Report
by: Mallika Reddy, et al.
Published: (2025)

GenAI at the Edge: Comprehensive Survey on Empowering Edge Devices
by: Navardi, Mozhgan, et al.
Published: (2025)

A Lesson in Splats: Teacher-Guided Diffusion for 3D Gaussian Splats Generation with 2D Supervision
by: Peng, Chensheng, et al.
Published: (2024)

LLMs Don't Know Their Own Decision Boundaries: The Unreliability of Self-Generated Counterfactual Explanations
by: Mayne, Harry, et al.
Published: (2025)

FractalCloud: A Fractal-Inspired Architecture for Efficient Large-Scale Point Cloud Processing
by: Fu, Yuzhe, et al.
Published: (2025)

Sparse Refinement for Efficient High-Resolution Semantic Segmentation
by: Liu, Zhijian, et al.
Published: (2024)

Looking Backward: Streaming Video-to-Video Translation with Feature Banks
by: Liang, Feng, et al.
Published: (2024)

Don't Trade Off Safety: Diffusion Regularization for Constrained Offline RL
by: Guo, Junyu, et al.
Published: (2025)

IoT-MCP: Bridging LLMs and IoT Systems Through Model Context Protocol
by: Yang, Ningyuan, et al.
Published: (2025)

Vision Transformers Don't Need Trained Registers
by: Jiang, Nick, et al.
Published: (2025)

MARS: Efficient, Adaptive Co-Scheduling for Heterogeneous Agentic Systems
by: Wang, Yifei, et al.
Published: (2026)

Vision-Zero: Scalable VLM Self-Improvement via Strategic Gamified Self-Play
by: Wang, Qinsi, et al.
Published: (2025)

Voice Evaluation of Reasoning Ability: Diagnosing the Modality-Induced Performance Gap
by: Lin, Yueqian, et al.
Published: (2025)

MITRA: A Large-Scale Parallel Corpus and Multilingual Pretrained Language Model for Machine Translation and Semantic Retrieval for Pāli, Sanskrit, Buddhist Chinese, and Tibetan
by: Nehrdich, Sebastian, et al.
Published: (2026)

Don't Waste Mistakes: Leveraging Negative RL-Groups via Confidence Reweighting
by: Feng, Yunzhen, et al.
Published: (2025)

LLaViDA: A Large Language Vision Driving Assistant for Explicit Reasoning and Enhanced Trajectory Planning
by: Liu, Yudong, et al.
Published: (2025)

Don't be salesmen
Published: (1997)

PrivAct: Internalizing Contextual Privacy Preservation via Multi-Agent Preference Training
by: Cheng, Yuhan, et al.
Published: (2026)

Don't double it: Efficient Agent Prediction in Occlusions
by: Rothenhäusler, Anna, et al.
Published: (2026)

Jet-RL: Enabling On-Policy FP8 Reinforcement Learning with Unified Training and Rollout Precision Flow
by: Xi, Haocheng, et al.
Published: (2026)

SD-NAE: Generating Natural Adversarial Examples with Stable Diffusion
by: Lin, Yueqian, et al.
Published: (2023)

Segment Any Motion in Videos
by: Huang, Nan, et al.
Published: (2025)

HallE-Control: Controlling Object Hallucination in Large Multimodal Models
by: Zhai, Bohan, et al.
Published: (2023)