Searcharxiv⌕ Search

arXiv subjects

Yichuan Deng

Publications and source records attributed to Yichuan Deng.

At least 19 recordsLinked to original sources
arXiv2025

OpenThoughts: Data Recipes for Reasoning Models

Reasoning models have made rapid progress on many benchmarks involving math, code, and science. Yet, there are still many open questions about the best training recipes for reasoning since state-of-the-art models often rely on proprietary datasets with little to no public information available. To address this, the goal of the OpenThoughts project is to create open-source datasets for training reasoning models. After initial explorations, our OpenThoughts2-1M dataset led to OpenThinker2-32B, the first model trained on public reasoning data to match DeepSeek-R1-Distill-32B on standard reasoning benchmarks such as AIME and LiveCodeBench. We then improve our dataset further by systematically investigating each step of our data generation pipeline with 1,000+ controlled experiments, which led to OpenThoughts3. Scaling the pipeline to 1.2M examples and using QwQ-32B as teacher yields our OpenThoughts3-7B model, which achieves state-of-the-art results: 53% on AIME 2025, 51% on LiveCodeBench 06/24-01/25, and 54% on GPQA Diamond - improvements of 15.3, 17.2, and 20.5 percentage points compared to the DeepSeek-R1-Distill-Qwen-7B. All of our datasets and models are available on https://openthoughts.ai.

cs.LG↗
arXiv2024

How Sparse Attention Approximates Exact Attention? Your Attention is Naturally $n^C$-Sparse

Sparse Attention is a technique that approximates standard attention computation with sub-quadratic complexity. This is achieved by selectively ignoring smaller entries in the attention matrix during the softmax function computation. Variations of this technique, such as pruning KV cache, sparsity-based fast attention, and Sparse Transformer, have been extensively utilized for efficient Large Language Models (LLMs) deployment. Despite its widespread use, a theoretical understanding of the conditions under which sparse attention performs on par with traditional attention remains elusive. This work aims to $\textbf{bridge this gap by examining the inherent sparsity of standard attention processes}$. Our theoretical framework reveals several brand-new key insights: $\bullet$ Attention is $n^{C}$-sparse, implying that considering only the largest $\Omega(n^{C})$ entries out of all $n$ entries is sufficient for sparse attention to approximate the exact attention matrix with decreasing loss. Here, $n$ represents the input length and $C \in (0, 1)$ is a constant. $\bullet$ Stable $o(\log(n))$-sparse attention, which approximates attention computation with $\log(n)$ or fewer entries, may not be feasible since the error will persist at a minimum of $O(1)$. $\bullet$ An adaptive strategy ($\alpha \cdot n^C, \alpha \in \mathbb{R}$) for the window size of efficient attention methods rather than a fixed one is guaranteed to perform more accurately and efficiently in a task for inference on flexible context lengths.

cs.LG↗
arXiv2024

Training Overparametrized Neural Networks in Sublinear Time

The success of deep learning comes at a tremendous computational and energy cost, and the scalability of training massively overparametrized neural networks is becoming a real barrier to the progress of artificial intelligence (AI). Despite the popularity and low cost-per-iteration of traditional backpropagation via gradient decent, stochastic gradient descent (SGD) has prohibitive convergence rate in non-convex settings, both in theory and practice. To mitigate this cost, recent works have proposed to employ alternative (Newton-type) training methods with much faster convergence rate, albeit with higher cost-per-iteration. For a typical neural network with $m=\mathrm{poly}(n)$ parameters and input batch of $n$ datapoints in $\mathbb{R}^d$, the previous work of [Brand, Peng, Song, and Weinstein, ITCS'2021] requires $\sim mnd + n^3$ time per iteration. In this paper, we present a novel training method that requires only $m^{1-α} n d + n^3$ amortized time in the same overparametrized regime, where $α\in (0.01,1)$ is some fixed constant. This method relies on a new and alternative view of neural networks, as a set of binary search trees, where each iteration corresponds to modifying a small subset of the nodes in the tree. We believe this view would have further applications in the design and analysis of deep neural networks (DNNs).

cs.LG↗
arXiv2024

Enhancing Stochastic Gradient Descent: A Unified Framework and Novel Acceleration Methods for Faster Convergence

Based on SGD, previous works have proposed many algorithms that have improved convergence speed and generalization in stochastic optimization, such as SGDm, AdaGrad, Adam, etc. However, their convergence analysis under non-convex conditions is challenging. In this work, we propose a unified framework to address this issue. For any first-order methods, we interpret the updated direction $g_t$ as the sum of the stochastic subgradient $\nabla f_t(x_t)$ and an additional acceleration term $\frac{2|\langle v_t, \nabla f_t(x_t) \rangle|}{\|v_t\|_2^2} v_t$, thus we can discuss the convergence by analyzing $\langle v_t, \nabla f_t(x_t) \rangle$. Through our framework, we have discovered two plug-and-play acceleration methods: \textbf{Reject Accelerating} and \textbf{Random Vector Accelerating}, we theoretically demonstrate that these two methods can directly lead to an improvement in convergence rate.

cs.LG↗
arXiv2023

Unmasking Transformers: A Theoretical Approach to Data Recovery via Attention Weights

In the realm of deep learning, transformers have emerged as a dominant architecture, particularly in natural language processing tasks. However, with their widespread adoption, concerns regarding the security and privacy of the data processed by these models have arisen. In this paper, we address a pivotal question: Can the data fed into transformers be recovered using their attention weights and outputs? We introduce a theoretical framework to tackle this problem. Specifically, we present an algorithm that aims to recover the input data $X \in \mathbb{R}^{d \times n}$ from given attention weights $W = QK^\top \in \mathbb{R}^{d \times d}$ and output $B \in \mathbb{R}^{n \times n}$ by minimizing the loss function $L(X)$. This loss function captures the discrepancy between the expected output and the actual output of the transformer. Our findings have significant implications for the Localized Layer-wise Mechanism (LLM), suggesting potential vulnerabilities in the model's design from a security and privacy perspective. This work underscores the importance of understanding and safeguarding the internal workings of transformers to ensure the confidentiality of processed data.

cs.LG↗
arXiv2023

Why Softmax Attention Outperforms Linear Attention

Large transformer models have achieved state-of-the-art results in numerous natural language processing tasks. Among the pivotal components of the transformer architecture, the attention mechanism plays a crucial role in capturing token interactions within sequences through the utilization of softmax function. Conversely, linear attention presents a more computationally efficient alternative by approximating the softmax operation with linear complexity. However, it exhibits substantial performance degradation when compared to the traditional softmax attention mechanism. In this paper, we bridge the gap in our theoretical understanding of the reasons behind the practical performance gap between softmax and linear attention. By conducting a comprehensive comparative analysis of these two attention mechanisms, we shed light on the underlying reasons for why softmax attention outperforms linear attention in most scenarios.

cs.CL↗
arXiv2023

Clustered Linear Contextual Bandits with Knapsacks

In this work, we study clustered contextual bandits where rewards and resource consumption are the outcomes of cluster-specific linear models. The arms are divided in clusters, with the cluster memberships being unknown to an algorithm. Pulling an arm in a time period results in a reward and in consumption for each one of multiple resources, and with the total consumption of any resource exceeding a constraint implying the termination of the algorithm. Thus, maximizing the total reward requires learning not only models about the reward and the resource consumption, but also cluster memberships. We provide an algorithm that achieves regret sublinear in the number of time periods, without requiring access to all of the arms. In particular, we show that it suffices to perform clustering only once to a randomly selected subset of the arms. To achieve this result, we provide a sophisticated combination of techniques from the literature of econometrics and of bandits with constraints.

cs.LG↗
arXiv2023

Convergence of Two-Layer Regression with Nonlinear Units

Large language models (LLMs), such as ChatGPT and GPT4, have shown outstanding performance in many human life task. Attention computation plays an important role in training LLMs. Softmax unit and ReLU unit are the key structure in attention computation. Inspired by them, we put forward a softmax ReLU regression problem. Generally speaking, our goal is to find an optimal solution to the regression problem involving the ReLU unit. In this work, we calculate a close form representation for the Hessian of the loss function. Under certain assumptions, we prove the Lipschitz continuous and the PSDness of the Hessian. Then, we introduce an greedy algorithm based on approximate Newton method, which converges in the sense of the distance to optimal solution. Last, We relax the Lipschitz condition and prove the convergence in the sense of loss value.

cs.LG↗
arXiv2023

Zero-th Order Algorithm for Softmax Attention Optimization

Large language models (LLMs) have brought about significant transformations in human society. Among the crucial computations in LLMs, the softmax unit holds great importance. Its helps the model generating a probability distribution on potential subsequent words or phrases, considering a series of input words. By utilizing this distribution, the model selects the most probable next word or phrase, based on the assigned probabilities. The softmax unit assumes a vital function in LLM training as it facilitates learning from data through the adjustment of neural network weights and biases. With the development of the size of LLMs, computing the gradient becomes expensive. However, Zero-th Order method can approximately compute the gradient with only forward passes. In this paper, we present a Zero-th Order algorithm specifically tailored for Softmax optimization. We demonstrate the convergence of our algorithm, highlighting its effectiveness in efficiently computing gradients for large-scale LLMs. By leveraging the Zeroth-Order method, our work contributes to the advancement of optimization techniques in the context of complex language models.

cs.LG↗
arXiv2023

Efficient Algorithm for Solving Hyperbolic Programs

Hyperbolic polynomials is a class of real-roots polynomials that has wide range of applications in theoretical computer science. Each hyperbolic polynomial also induces a hyperbolic cone that is of particular interest in optimization due to its generality, as by choosing the polynomial properly, one can easily recover the classic optimization problems such as linear programming and semidefinite programming. In this work, we develop efficient algorithms for hyperbolic programming, the problem in each one wants to minimize a linear objective, under a system of linear constraints and the solution must be in the hyperbolic cone induced by the hyperbolic polynomial. Our algorithm is an instance of interior point method (IPM) that, instead of following the central path, it follows the central Swath, which is a generalization of central path. To implement the IPM efficiently, we utilize a relaxation of the hyperbolic program to a quadratic program, coupled with the first four moments of the hyperbolic eigenvalues that are crucial to update the optimization direction. We further show that, given an evaluation oracle of the polynomial, our algorithm only requires $O(n^2d^{2.5})$ oracle calls, where $n$ is the number of variables and $d$ is the degree of the polynomial, with extra $O((n+m)^3 d^{0.5})$ arithmetic operations, where $m$ is the number of constraints.

math.OC↗
arXiv2023

Faster Robust Tensor Power Method for Arbitrary Order

Tensor decomposition is a fundamental method used in various areas to deal with high-dimensional data. \emph{Tensor power method} (TPM) is one of the widely-used techniques in the decomposition of tensors. This paper presents a novel tensor power method for decomposing arbitrary order tensors, which overcomes limitations of existing approaches that are often restricted to lower-order (less than $3$) tensors or require strong assumptions about the underlying data structure. We apply sketching method, and we are able to achieve the running time of $\widetilde{O}(n^{p-1})$, on the power $p$ and dimension $n$ tensor. We provide a detailed analysis for any $p$-th order tensor, which is never given in previous works.

cs.LG↗
arXiv2023

Attention Scheme Inspired Softmax Regression

Large language models (LLMs) have made transformed changes for human society. One of the key computation in LLMs is the softmax unit. This operation is important in LLMs because it allows the model to generate a distribution over possible next words or phrases, given a sequence of input words. This distribution is then used to select the most likely next word or phrase, based on the probabilities assigned by the model. The softmax unit plays a crucial role in training LLMs, as it allows the model to learn from the data by adjusting the weights and biases of the neural network. In the area of convex optimization such as using central path method to solve linear programming. The softmax function has been used a crucial tool for controlling the progress and stability of potential function [Cohen, Lee and Song STOC 2019, Brand SODA 2020]. In this work, inspired the softmax unit, we define a softmax regression problem. Formally speaking, given a matrix $A \in \mathbb{R}^{n \times d}$ and a vector $b \in \mathbb{R}^n$, the goal is to use greedy type algorithm to solve \begin{align*} \min_{x} \| \langle \exp(Ax), {\bf 1}_n \rangle^{-1} \exp(Ax) - b \|_2^2. \end{align*} In certain sense, our provable convergence result provides theoretical support for why we can use greedy algorithm to train softmax function in practice.

cs.LG↗
arXiv2023

Solving Tensor Low Cycle Rank Approximation

Large language models have become ubiquitous in modern life, finding applications in various domains such as natural language processing, language translation, and speech recognition. Recently, a breakthrough work [Zhao, Panigrahi, Ge, and Arora Arxiv 2023] explains the attention model from probabilistic context-free grammar (PCFG). One of the central computation task for computing probability in PCFG is formulating a particular tensor low rank approximation problem, we can call it tensor cycle rank. Given an $n \times n \times n$ third order tensor $A$, we say that $A$ has cycle rank-$k$ if there exists three $n \times k^2$ size matrices $U , V$, and $W$ such that for each entry in each \begin{align*} A_{a,b,c} = \sum_{i=1}^k \sum_{j=1}^k \sum_{l=1}^k U_{a,i+k(j-1)} \otimes V_{b, j + k(l-1)} \otimes W_{c, l + k(i-1) } \end{align*} for all $a \in [n], b \in [n], c \in [n]$. For the tensor classical rank, tucker rank and train rank, it has been well studied in [Song, Woodruff, Zhong SODA 2019]. In this paper, we generalize the previous ``rotation and sketch'' technique in page 186 of [Song, Woodruff, Zhong SODA 2019] and show an input sparsity time algorithm for cycle rank.

cs.DS↗
arXiv2023

Randomized and Deterministic Attention Sparsification Algorithms for Over-parameterized Feature Dimension

Large language models (LLMs) have shown their power in different areas. Attention computation, as an important subroutine of LLMs, has also attracted interests in theory. Recently the static computation and dynamic maintenance of attention matrix has been studied by [Alman and Song 2023] and [Brand, Song and Zhou 2023] from both algorithmic perspective and hardness perspective. In this work, we consider the sparsification of the attention problem. We make one simplification which is the logit matrix is symmetric. Let $n$ denote the length of sentence, let $d$ denote the embedding dimension. Given a matrix $X \in \mathbb{R}^{n \times d}$, suppose $d \gg n$ and $\| X X^\top \|_{\infty} < r$ with $r \in (0,0.1)$, then we aim for finding $Y \in \mathbb{R}^{n \times m}$ (where $m\ll d$) such that \begin{align*} \| D(Y)^{-1} \exp( Y Y^\top ) - D(X)^{-1} \exp( X X^\top) \|_{\infty} \leq O(r) \end{align*} We provide two results for this problem. $\bullet$ Our first result is a randomized algorithm. It runs in $\widetilde{O}(\mathrm{nnz}(X) + n^ω ) $ time, has $1-δ$ succeed probability, and chooses $m = O(n \log(n/δ))$. Here $\mathrm{nnz}(X)$ denotes the number of non-zero entries in $X$. We use $ω$ to denote the exponent of matrix multiplication. Currently $ω\approx 2.373$. $\bullet$ Our second result is a deterministic algorithm. It runs in $\widetilde{O}(\min\{\sum_{i\in[d]}\mathrm{nnz}(X_i)^2, dn^{ω-1}\} + n^{ω+1})$ time and chooses $m = O(n)$. Here $X_i$ denote the $i$-th column of matrix $X$. Our main findings have the following implication for applied LLMs task: for any super large feature dimension, we can reduce it down to the size nearly linear in length of sentence.

cs.DS↗
arXiv2023

An Improved Sample Complexity for Rank-1 Matrix Sensing

Matrix sensing is a problem in signal processing and machine learning that involves recovering a low-rank matrix from a set of linear measurements. The goal is to reconstruct the original matrix as accurately as possible, given only a set of linear measurements obtained by sensing the matrix [Jain, Netrapalli and Shanghavi, 2013]. In this work, we focus on a particular direction of matrix sensing, which is called rank-$1$ matrix sensing [Zhong, Jain and Dhillon, 2015]. We present an improvement over the original algorithm in [Zhong, Jain and Dhillon, 2015]. It is based on a novel analysis and sketching technique that enables faster convergence rates and better accuracy in recovering low-rank matrices. The algorithm focuses on developing a theoretical understanding of the matrix sensing problem and establishing its advantages over previous methods. The proposed sketching technique allows for efficiently extracting relevant information from the linear measurements, making the algorithm computationally efficient and scalable. Our novel matrix sensing algorithm improves former result [Zhong, Jain and Dhillon, 2015] on in two senses: $\bullet$ We improve the sample complexity from $\widetilde{O}(ε^{-2} dk^2)$ to $\widetilde{O}(ε^{-2} (d+k^2))$. $\bullet$ We improve the running time from $\widetilde{O}(md^2 k^2)$ to $\widetilde{O}(m d^2 k)$. The proposed algorithm has theoretical guarantees and is analyzed to provide insights into the underlying structure of low-rank matrices and the nature of the linear measurements used in the recovery process. It advances the theoretical understanding of matrix sensing and provides a new approach for solving this important problem.

cs.IT↗
arXiv2023

Streaming Kernel PCA Algorithm With Small Space

Principal Component Analysis (PCA) is a widely used technique in machine learning, data analysis and signal processing. With the increase in the size and complexity of datasets, it has become important to develop low-space usage algorithms for PCA. Streaming PCA has gained significant attention in recent years, as it can handle large datasets efficiently. The kernel method, which is commonly used in learning algorithms such as Support Vector Machines (SVMs), has also been applied in PCA algorithms. We propose a streaming algorithm for Kernel PCA problems based on the traditional scheme by Oja. Our algorithm addresses the challenge of reducing the memory usage of PCA while maintaining its accuracy. We analyze the performance of our algorithm by studying the conditions under which it succeeds. Specifically, we show that, when the spectral ratio $R := λ_1/λ_2$ of the target covariance matrix is lower bounded by $C \cdot \log n\cdot \log d$, the streaming PCA can be solved with $O(d)$ space cost. Our proposed algorithm has several advantages over existing methods. First, it is a streaming algorithm that can handle large datasets efficiently. Second, it employs the kernel method, which allows it to capture complex nonlinear relationships among data points. Third, it has a low-space usage, making it suitable for applications where memory is limited.

cs.LG↗
arXiv2022

Dynamic Kernel Graph Sparsifiers

A geometric graph associated with a set of points $P= \{x_1, x_2, \cdots, x_n \} \subset \mathbb{R}^d$ and a fixed kernel function $\mathsf{K}:\mathbb{R}^d\times \mathbb{R}^d\to\mathbb{R}_{\geq 0}$ is a complete graph on $P$ such that the weight of edge $(x_i, x_j)$ is $\mathsf{K}(x_i, x_j)$. We present a fully-dynamic data structure that maintains a spectral sparsifier of a geometric graph under updates that change the locations of points in $P$ one at a time. The update time of our data structure is $n^{o(1)}$ with high probability, and the initialization time is $n^{1+o(1)}$. Under certain assumption, our data structure can be made robust against adaptive adversaries, which makes our sparsifier applicable in iterative optimization algorithms. We further show that the Laplacian matrices corresponding to geometric graphs admit a randomized sketch for maintaining matrix-vector multiplication and projection in $n^{o(1)}$ time, under sparse updates to the query vectors, or under modification of points in $P$.

cs.DS↗
arXiv2022

Discrepancy Minimization in Input-Sparsity Time

A recent work by [Larsen, SODA 2023] introduced a faster combinatorial alternative to Bansal's SDP algorithm for finding a coloring $x \in \{-1, 1\}^n$ that approximately minimizes the discrepancy $\mathrm{disc}(A, x) := | A x |_{\infty}$ of a real-valued $m \times n$ matrix $A$. Larsen's algorithm runs in $\widetilde{O}(mn^2)$ time compared to Bansal's $\widetilde{O}(mn^{4.5})$-time algorithm, with a slightly weaker logarithmic approximation ratio in terms of the hereditary discrepancy of $A$ [Bansal, FOCS 2010]. We present a combinatorial $\widetilde{O}(\mathrm{nnz}(A) + n^3)$-time algorithm with the same approximation guarantee as Larsen's, optimal for tall matrices where $m = \mathrm{poly}(n)$. Using a more intricate analysis and fast matrix multiplication, we further achieve a runtime of $\widetilde{O}(\mathrm{nnz}(A) + n^{2.53})$, breaking the cubic barrier for square matrices and surpassing the limitations of linear-programming approaches [Eldan and Singh, RS&A 2018]. Our algorithm relies on two key ideas: (i) a new sketching technique for finding a projection matrix with a short $\ell_2$-basis using implicit leverage-score sampling, and (ii) a data structure for efficiently implementing the iterative Edge-Walk partial-coloring algorithm [Lovett and Meka, SICOMP 2015], and using an alternative analysis to enable ''lazy'' batch updates with low-rank corrections. Our results nearly close the computational gap between real-valued and binary matrices, for which input-sparsity time coloring was recently obtained by [Jain, Sah and Sawhney, SODA 2023].

cs.DS↗