Publications (13)
Bilevel ZOFO: Efficient LLM Fine-Tuning and Meta-Training
Reza Shirkavand, Peiran Yu, Qi He +1
Fine-tuning pre-trained Large Language Models (LLMs) for downstream tasks using First-Order (FO) optimizers presents significant computational challenges. Parameter-Efficient Fine-…
Breaking the Correlation Plateau: On the Optimization and Capacity Limits of Attention-Based Regressors
Jingquan Yan, Yuwei Miao, Peiran Yu +1
Attention-based regression models are often trained by jointly optimizing Mean Squared Error (MSE) loss and Pearson correlation coefficient (PCC) loss, emphasizing the magnitude of…
Efficient Fine-Tuning and Concept Suppression for Pruned Diffusion Models
Reza Shirkavand, Peiran Yu, Shangqian Gao +3
Recent advances in diffusion generative models have yielded remarkable progress. While the quality of generated content continues to improve, these models have grown considerably i…
Iteratively reweighted algorithms with extrapolation
Peiran Yu, Ting Kei Pong
Iteratively reweighted algorithm is a popular algorithm for solving a large class of optimization problems whose objective is the sum of a Lipschitz differentiable loss fu…
Provably Mitigating Corruption, Overoptimization, and Verbosity Simultaneously in Offline and Online RLHF/DPO Alignment
Ziyi Chen, Junyi Li, Peiran Yu +1
Reinforcement learning from human feedback (RLHF) and direct preference optimization (DPO) are important techniques to align large language models (LLM) with human preference. Howe…
Convergence analysis for a variant of manifold proximal point algorithm based on Kurdyka-Åojasiewicz property
Peiran Yu, Liaoyuan Zeng, Ting Kei Pong
We incorporate an iteratively reweighted strategy in the manifold proximal point algorithm (ManPPA) in [12] to solve an enhanced sparsity inducing model for identifying sparse yet…
Analysis and algorithms for some compressed sensing models based on L1/L2 minimization
Liaoyuan Zeng, Peiran Yu, Ting Kei Pong
Recently, in a series of papers [32,38,39,41], the ratio of and norms was proposed as a sparsity inducing function for noiseless compressed sensing. In this paper…
Cost-Aware Contrastive Routing for LLMs
Reza Shirkavand, Shangqian Gao, Peiran Yu +1
We study cost-aware routing for large language models across diverse and dynamic pools of models. Existing approaches often overlook prompt-specific context, rely on expensive mode…
New Hybrid Fine-Tuning Paradigm for LLMs: Algorithm Design and Convergence Analysis Framework
Shaocong Ma, Peiran Yu, Heng Huang
Fine-tuning Large Language Models (LLMs) typically involves either full fine-tuning, which updates all model parameters, or Parameter-Efficient Fine-Tuning (PEFT), which adjusts a…
Zeroth-Order Methods for Stochastic Nonconvex Nonsmooth Composite Optimization
Ziyi Chen, Peiran Yu, Heng Huang
This work aims to solve a stochastic nonconvex nonsmooth composite optimization problem. Previous works on composite optimization problem requires the major part to satisfy Lipschi…
Convergence rate analysis of a sequential convex programming method with line search for a class of constrained difference-of-convex optimization problems
Peiran Yu, Ting Kei Pong, Zhaosong Lu
In this paper, we study the sequential convex programming method with monotone line search (SCP) in [46] for a class of difference-of-convex (DC) optimization problems with…
Kurdyka-Åojasiewicz exponent via inf-projection
Peiran Yu, Guoyin Li, Ting Kei Pong
Kurdyka-Lojasiewicz (KL) exponent plays an important role in estimating the convergence rate of many contemporary first-order methods. In particular, a KL exponent of for…
Revisiting Convergence: Shuffling Complexity Beyond Lipschitz Smoothness
Qi He, Peiran Yu, Ziyi Chen +1
Shuffling-type gradient methods are favored in practice for their simplicity and rapid empirical performance. Despite extensive development of convergence guarantees under various…