6 papers
Preference-Based Self-Distillation: Beyond KL Matching via Reward Regularization
Xin Yu, Liuchen Liao, Yiwen Zhang +3
On-policy distillation is an efficient alternative to reinforcement learning, offering dense token-level training signals. However, its reliance on a stronger external teacher has…
Uno-Orchestra: Parsimonious Agent Routing via Selective Delegation
Zhiqing Cui, Haotong Xie, Jiahao Yuan +11
Large language model (LLM) multi-agent systems typically rely on rigid orchestration, committing either to flat per-query routing or to hand-engineered task decomposition, so decom…
A New Inexact Manifold Proximal Linear Algorithm with Adaptive Stopping Criteria
Zhong Zheng, Xin Yu, Shiqian Ma +1
This paper proposes a new inexact manifold proximal linear (IManPL) algorithm for solving nonsmooth, nonconvex composite optimization problems over an embedded submanifold. At each…
Train Large, Deploy Compact: Structured Compression for Compact Low-Rank Adaptation
Xin Yu, Cong Xie, Ziyu Zhao +4
Low-rank adaptation (LoRA) has become a widely used paradigm for parameter-efficient fine-tuning of large language models, yet its representational capacity often lags behind full…
AltLoRA: Towards Better Gradient Approximation in Low-Rank Adaptation with Alternating Projections
Xin Yu, Yujia Wang, Jinghui Chen +1
Low-Rank Adaptation (LoRA) has emerged as an effective technique for reducing memory overhead in fine-tuning large language models. However, it often suffers from sub-optimal perfo…
Understanding the Statistical Accuracy-Communication Trade-off in Personalized Federated Learning with Minimax Guarantees
Xin Yu, Zelin He, Ying Sun +2
Personalized federated learning (PFL) offers a flexible framework for aggregating information across distributed clients with heterogeneous data. This work considers a personalized…