| 1 |
Trainable Vector Quantization for Large Language Models |
提出可训练向量量化方法以优化大语言模型压缩 |
distillation large language model |
|
|
| 2 |
Reliability-Weighted Direct Preference Optimization: Penalized Weights for Robust Preference Learning |
提出RW-DPO以解决偏好学习中的噪声数据问题 |
preference learning DPO direct preference optimization |
|
|
| 3 |
Hyperbolic Inverse Reinforcement Learning |
提出超曲率逆强化学习以提升奖励函数的解释性与预测准确性 |
reinforcement learning inverse reinforcement learning |
|
|
| 4 |
SPT-Flow: One Integration, Many Futures in Continuous-Time World Models |
提出SPT-Flow以解决连续时间世界模型中的随机条件利用问题 |
world model world models |
|
|
| 5 |
MUON-RT Reusing the Teacher's Trajectory for Momentum-Free Muon for LLM Self-Distillation |
提出Muon-RT以解决大语言模型的动量缓冲问题 |
teacher-student distillation large language model |
|
|
| 6 |
Towards Understanding the Impact of Plasticity Loss on Reinforcement Learning in Stochastic Environments |
研究塑性损失对随机环境中强化学习的影响 |
reinforcement learning deep reinforcement learning SAC |
|
|
| 7 |
TalentRL: Strategy-Guided Reinforcement Learning for Hierarchical Skill Routing |
提出TalentRL以解决技能选择策略不足的问题 |
reinforcement learning large language model |
|
|
| 8 |
Statistical Foundations of Reinforcement Learning: A Non-Asymptotic Perspective |
提出非渐近统计方法以提升强化学习的数据效率 |
reinforcement learning offline RL |
|
|
| 9 |
rl_rob: A transparent, from-scratch Python library for teaching reinforcement learning from tabular methods to deep policy gradients |
提出rl_rob库以透明化强化学习教学过程 |
reinforcement learning deep reinforcement learning |
|
|
| 10 |
RAMA: Retrieval-Augmented Multi-Agent Routing with Episodic Memory for LEO Satellite Networks |
提出RAMA以解决LEO卫星网络中的动态路由问题 |
reinforcement learning deep reinforcement learning DRL |
|
|
| 11 |
CausalXRL: Explainable Reinforcement Learning through Causal Graph Reasoning |
提出CausalXRL以解决强化学习可解释性问题 |
reinforcement learning |
|
|
| 12 |
When Is a Game a Markov Decision Process? Correspondences Between Game Theory and Reinforcement Learning |
提出游戏理论与强化学习之间的定理对应关系 |
reinforcement learning |
|
|
| 13 |
Continual Reinforcement Learning Is an Integration Problem |
提出十项设计标准以解决持续强化学习的整合问题 |
reinforcement learning |
|
|
| 14 |
From Representation to Readout: Mechanistic Analysis of Membership Signals in Language Models |
提出DUG指标以分析语言模型中的成员信号 |
reinforcement learning large language model |
|
|
| 15 |
SparseGDN: Chunk Sparse in Linear Attention |
提出SparseGDN以解决长上下文序列建模的计算瓶颈问题 |
linear attention |
|
|
| 16 |
Coarse-Graining Optimal Transport for Rollout-Free Post-Training of Flow Matching |
提出ReCouple框架以优化流匹配的后训练过程 |
flow matching |
|
|