7 citations · 19 across the 17 of their papers we have counts for
17 papers
On the Principles Behind Neural Network Optimizers
Yushun Zhang
Reliable optimization is central to neural network (NN) training, yet Adam, the default optimizer for modern LLMs, rests on a fragile foundation. This thesis develops a principled…
PC Layer: Polynomial Weight Preconditioning for Improving LLM Pre-Training
Senmiao Wang, Tiantian Fang, Haoran Zhang +4
We propose a preconditioning (PC) layer, a weight parameterization via polynomial preconditioner that ensures stable weight conditioning throughout LLM training. The PC module resh…
DeepSeek-V4: Towards Highly Efficient Million-Token Context Intelligence
DeepSeek-AI, Anyi Xu, Bangcai Lin +315
We present a preview version of DeepSeek-V4 series, including two strong Mixture-of-Experts (MoE) language models -- DeepSeek-V4-Pro with 1.6T parameters (49B activated) and DeepSe…
Adam Converges Without Any Modification On Update Rules
Yushun Zhang, Bingran Li, Congliang Chen +2
Adam is the default algorithm for training neural networks, including large language models (LLMs). However, \citet{reddi2019convergence} provided an example that Adam diverges, ra…
Exact Causal Attention with 10% Fewer Operations
Dmitry Rybin, Yushun Zhang, Ding Tian +2
We present Exact Causal Attention (ECA), a Strassen-style algorithm that computes exact Causal Attention using 10\% fewer operations. ECA improves a special class of matrix multipl…
Exploring and Improving Initialization for Deep Graph Neural Networks: A Signal Propagation Perspective
Senmiao Wang, Yupeng Chen, Yushun Zhang +2
Graph Neural Networks (GNNs) often suffer from performance degradation as the network depth increases. This paper addresses this issue by introducing initialization methods that en…
Can Be Faster
Dmitry Rybin, Yushun Zhang, Zhi-Quan Luo
We present RXTX, a new algorithm for computing the product of matrix by its transpose for . RXTX uses fewer multiplications and fe…
Towards Quantifying the Hessian Structure of Neural Networks
Zhaorui Dong, Yushun Zhang, Jianfeng Yao +1
Empirical studies reported that the Hessian matrix of neural networks (NNs) exhibits a near-block-diagonal structure, yet its theoretical foundation remains unclear. In this work,…
Finite Horizon Optimization: Framework and Applications
Yushun Zhang, Dmitry Rybin, Zhi-Quan Luo
In modern engineering scenarios, there is often a strict upper bound on the number of algorithm iterations that can be performed within a given time limit. This raises the question…
MoFO: Momentum-Filtered Optimizer for Mitigating Forgetting in LLM Fine-Tuning
Yupeng Chen, Senmiao Wang, Yushun Zhang +5
Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of tasks. Typically, LLMs are first pre-trained on large corpora and subsequently fine-tu…
Adam-mini: Use Fewer Learning Rates To Gain More
Yushun Zhang, Congliang Chen, Ziniu Li +6
We propose Adam-mini, an optimizer that achieves on par or better performance than AdamW with 50% less memory footprint. Adam-mini reduces memory by cutting down the learning rate…
Why Transformers Need Adam: A Hessian Perspective
Yushun Zhang, Congliang Chen, Tian Ding +3
SGD performs worse than Adam by a significant margin on Transformers, but the reason remains unclear. In this work, we provide an explanation through the lens of Hessian: (i) Trans…
ReMax: A Simple, Effective, and Efficient Reinforcement Learning Method for Aligning Large Language Models
Ziniu Li, Tian Xu, Yushun Zhang +4
Reinforcement Learning from Human Feedback (RLHF) is key to aligning Large Language Models (LLMs), typically paired with the Proximal Policy Optimization (PPO) algorithm. While PPO…
Uncertainty and Explainable Analysis of Machine Learning Model for Reconstruction of Sonic Slowness Logs
Hua Wang, Yuqiong Wu, Yushun Zhang +4
Logs are valuable information for oil and gas fields as they help to determine the lithology of the formations surrounding the borehole and the location and reserves of subsurface…
When Expressivity Meets Trainability: Fewer than Neurons Can Work
Jiawei Zhang, Yushun Zhang, Mingyi Hong +2
Modern neural networks are often quite wide, causing large memory and computation costs. It is thus of great interest to train a narrower network. However, training narrow neural n…
Provable Adaptivity of Adam under Non-uniform Smoothness
Bohan Wang, Yushun Zhang, Huishuai Zhang +6
Adam is widely adopted in practical applications due to its fast convergence. However, its theoretical analysis is still far from satisfactory. Existing convergence analyses for Ad…
Adam Can Converge Without Any Modification On Update Rules
Yushun Zhang, Congliang Chen, Naichen Shi +2
Ever since Reddi et al. 2018 pointed out the divergence issue of Adam, many new variants have been designed to obtain convergence. However, vanilla Adam remains exceptionally popul…