no code implementations • ICML 2020 • Sijia Liu, Songtao Lu, Xiangyi Chen, Yao Feng, Kaidi Xu, Abdullah Al-Dujaili, Mingyi Hong, Una-May O'Reilly
In this paper, we study the problem of constrained min-max optimization in a black-box setting, where the desired optimizer cannot access the gradients of the objective function but may query its values.
no code implementations • ICML 2020 • Haoran Sun, Songtao Lu, Mingyi Hong
Similarly, for online problems, the proposed method achieves an $\mathcal{O}(m \epsilon^{-3/2})$ sample complexity and an $\mathcal{O}(\epsilon^{-1})$ communication complexity, while the best existing bounds are $\mathcal{O}(m\epsilon^{-2})$ and $\mathcal{O}(\epsilon^{-2})$.
no code implementations • 31 Jan 2025 • Han Zhong, Yutong Yin, Shenao Zhang, Xiaojun Xu, Yuanxin Liu, Yifei Zuo, Zhihan Liu, Boyi Liu, Sirui Zheng, Hongyi Guo, LiWei Wang, Mingyi Hong, Zhaoran Wang
Large Language Models (LLMs) have demonstrated remarkable capabilities in complex reasoning tasks, yet generating reliable reasoning processes remains a significant challenge.
1 code implementation • 5 Dec 2024 • Jiangweizhi Peng, Zhiwei Tang, Gaowen Liu, Charles Fleming, Mingyi Hong
Our method introduces a novel optimization framework that leverages both the continuous prompt embedding and the injected noise trajectory in the sampling process to generate safe images.
no code implementations • 25 Nov 2024 • Rajesh Shrestha, Mingjie Shao, Mingyi Hong, Wing-Kin Ma, Xiao Fu
In frequency division duplex (FDD) massive MIMO systems, a major challenge lies in acquiring the downlink channel state information}\ (CSI) at the base station (BS) from limited feedback sent by the user equipment (UE).
no code implementations • 12 Nov 2024 • Bingqing Song, Boran Han, Shuai Zhang, Jie Ding, Mingyi Hong
While the Transformer architecture has achieved remarkable success across various domains, a thorough theoretical foundation explaining its optimization dynamics is yet to be fully developed.
no code implementations • 29 Oct 2024 • Zhiqi Bu, Xiaomeng Jin, Bhanukiran Vinzamuri, Anil Ramakrishna, Kai-Wei Chang, Volkan Cevher, Mingyi Hong
Machine unlearning has been used to remove unwanted knowledge acquired by large language models (LLMs).
no code implementations • 4 Oct 2024 • Xinwei Zhang, Zhiqi Bu, Borja Balle, Mingyi Hong, Meisam Razaviyayn, Vahab Mirrokni
This approach led to the development of DP optimizers that have comparable performance with their non-private counterparts in fine-tuning tasks or in tasks with a small number of training parameters.
no code implementations • 12 Sep 2024 • Xun Xian, Ganghua Wang, Xuan Bi, Jayanth Srinivasa, Ashish Kundu, Charles Fleming, Mingyi Hong, Jie Ding
To understand this vulnerability, we discovered that the deviation from the query's embedding to that of the poisoned document tends to follow a pattern in which the high similarity between the poisoned document and the query is retained, thereby enabling precise retrieval.
no code implementations • 24 Aug 2024 • Xinwei Zhang, Zhiqi Bu, Mingyi Hong, Meisam Razaviyayn
More specifically, by defining the ``frequency domain'' for both the gradient and differential privacy (DP) noise, we have developed a new component, called DOPPLER.
no code implementations • 11 Jun 2024 • Chenliang Li, Siliang Zeng, Zeyi Liao, Jiaxiang Li, Dongyeop Kang, Alfredo Garcia, Mingyi Hong
Aligning human preference and value is an important requirement for building contemporary foundation models and embodied AI.
1 code implementation • 4 Jun 2024 • Andi Han, Jiaxiang Li, Wei Huang, Mingyi Hong, Akiko Takeda, Pratik Jawanpuria, Bamdev Mishra
In this work, we propose to parameterize the weights as a sum of low-rank and sparse matrices for pretraining, which we call SLTrain.
no code implementations • 29 May 2024 • Zhiwei Tang, Jiangweizhi Peng, Jiasheng Tang, Mingyi Hong, Fan Wang, Tsung-Hui Chang
In this work, we focus on the alignment problem of diffusion models with a continuous reward function, which represents specific objectives for downstream tasks, such as increasing darkness or improving the aesthetics of images.
1 code implementation • 28 May 2024 • Jiaxiang Li, Siliang Zeng, Hoi-To Wai, Chenliang Li, Alfredo Garcia, Mingyi Hong
State-of-the-art techniques such as Reinforcement Learning from Human Feedback (RLHF) often consist of two stages: 1) supervised fine-tuning (SFT), where the model is fine-tuned by learning from human demonstration data; 2) Preference learning, where preference data is used to learn a reward model, which is in turn used by a reinforcement learning (RL) step to fine-tune the model.
1 code implementation • 24 May 2024 • Yimeng Zhang, Xin Chen, Jinghan Jia, Yihua Zhang, Chongyu Fan, Jiancheng Liu, Mingyi Hong, Ke Ding, Sijia Liu
The techniques of machine unlearning, also known as concept erasing, have been developed to address these risks.
1 code implementation • 16 Apr 2024 • Chung-Yiu Yau, Hoi-To Wai, Parameswaran Raman, Soumajyoti Sarkar, Mingyi Hong
We follow the global contrastive learning loss as introduced in SogCLR, and propose EMC$^2$ which utilizes an adaptive Metropolis-Hastings subroutine to generate hardness-aware negative samples in an online fashion during the optimization.
2 code implementations • 28 Feb 2024 • Zhiqi Bu, Xinwei Zhang, Mingyi Hong, Sheng Zha, George Karypis
The superior performance of large foundation models relies on the use of massive amounts of high-quality data, which often contain sensitive, private and copyrighted material that requires formal protection.
1 code implementation • 18 Feb 2024 • Yihua Zhang, Pingzhi Li, Junyuan Hong, Jiaxiang Li, Yimeng Zhang, Wenqing Zheng, Pin-Yu Chen, Jason D. Lee, Wotao Yin, Mingyi Hong, Zhangyang Wang, Sijia Liu, Tianlong Chen
In the evolving landscape of natural language processing (NLP), fine-tuning pre-trained Large Language Models (LLMs) with first-order (FO) optimizers like SGD and Adam has become standard.
no code implementations • 23 Jan 2024 • Xun Xian, Ganghua Wang, Xuan Bi, Jayanth Srinivasa, Ashish Kundu, Mingyi Hong, Jie Ding
Subsequently, we employ a classifier that is jointly trained with the watermark to detect the presence of the watermark.
no code implementations • 17 Jan 2024 • Kaan Ozkara, Can Karakus, Parameswaran Raman, Mingyi Hong, Shoham Sabach, Branislav Kveton, Volkan Cevher
Following the introduction of Adam, several novel adaptive optimizers for deep learning have been proposed.
no code implementations • 5 Jan 2024 • Ruichen Jiang, Parameswaran Raman, Shoham Sabach, Aryan Mokhtari, Mingyi Hong, Volkan Cevher
In this paper, we introduce a novel subspace cubic regularized Newton method that achieves a dimension-independent global convergence rate of ${O}\left(\frac{1}{mk}+\frac{1}{k^2}\right)$ for solving convex optimization problems.
1 code implementation • 24 Nov 2023 • Xinwei Zhang, Zhiqi Bu, Zhiwei Steven Wu, Mingyi Hong
In our work, we propose a new error-feedback (EF) DP algorithm as an alternative to DPSGD-GC, which not only offers a diminishing utility bound without inducing a constant clipping bias, but more importantly, it allows for an arbitrary choice of clipping threshold that is independent of the problem.
no code implementations • 16 Oct 2023 • Ganghua Wang, Xun Xian, Jayanth Srinivasa, Ashish Kundu, Xuan Bi, Mingyi Hong, Jie Ding
The growing dependence on machine learning in real-world applications emphasizes the importance of understanding and ensuring its safety.
2 code implementations • 15 Sep 2023 • Ran Wei, Siliang Zeng, Chenliang Li, Alfredo Garcia, Anthony McDonald, Mingyi Hong
We consider a Bayesian approach to offline model-based inverse reinforcement learning (IRL).
no code implementations • 1 Aug 2023 • Yihua Zhang, Prashant Khanduri, Ioannis Tsaknakis, Yuguang Yao, Mingyi Hong, Sijia Liu
Overall, we hope that this article can serve to accelerate the adoption of BLO as a generic tool to model, analyze, and innovate on a wide array of emerging SP and ML applications.
no code implementations • 16 Mar 2023 • Xinwei Zhang, Mingyi Hong, Jie Chen
In this paper, we propose a model splitting method that splits a backbone GNN across the clients and the server and a communication-efficient algorithm, GLASU, to train such a model.
no code implementations • 4 Mar 2023 • Yihua Zhang, Pranay Sharma, Parikshit Ram, Mingyi Hong, Kush Varshney, Sijia Liu
We propose a new IRM variant to address this limitation based on a novel viewpoint of ensemble IRM games as consensus-constrained bi-level optimization.
1 code implementation • NeurIPS 2023 • Siliang Zeng, Chenliang Li, Alfredo Garcia, Mingyi Hong
Offline inverse reinforcement learning (Offline IRL) aims to recover the structure of rewards and environment dynamics that underlie observed actions in a fixed, finite set of demonstrations from an expert agent.
no code implementations • 9 Nov 2022 • Jinghan Jia, Mingyi Hong, Yimeng Zhang, Mehmet Akçakaya, Sijia Liu
We find a new instability source of MRI image reconstruction, i. e., the lack of reconstruction robustness against spatial transformations of an input, e. g., rotation and cutout.
no code implementations • NeurIPS 2021 • Jiawei Zhang, Yushun Zhang, Mingyi Hong, Ruoyu Sun, Zhi-Quan Luo
Third, we consider a constrained optimization formulation where the feasible region is the nice local region, and prove that every KKT point is a nearly global minimizer.
1 code implementation • 8 Oct 2022 • Yihua Zhang, Yuguang Yao, Parikshit Ram, Pu Zhao, Tianlong Chen, Mingyi Hong, Yanzhi Wang, Sijia Liu
To reduce the computation overhead, various efficient 'one-shot' pruning methods have been developed, but these schemes are usually unable to find winning tickets as good as IMP.
no code implementations • 4 Oct 2022 • Siliang Zeng, Mingyi Hong, Alfredo Garcia
Other approaches in the inverse reinforcement learning (IRL) literature emphasize policy estimation at the expense of reduced reward estimation accuracy.
no code implementations • 4 Oct 2022 • Siliang Zeng, Chenliang Li, Alfredo Garcia, Mingyi Hong
To reduce the computational burden of a nested loop, novel methods such as SQIL [1] and IQ-Learn [2] emphasize policy estimation at the expense of reward estimation accuracy.
no code implementations • 23 Jun 2022 • Xun Xian, Mingyi Hong, Jie Ding
The privacy of machine learning models has become a significant concern in many emerging Machine-Learning-as-a-Service applications, where prediction services based on well-trained models are offered to users via pay-per-query.
2 code implementations • 13 Jun 2022 • Gaoyuan Zhang, Songtao Lu, Yihua Zhang, Xiangyi Chen, Pin-Yu Chen, Quanfu Fan, Lee Martie, Lior Horesh, Mingyi Hong, Sijia Liu
Spurred by that, we propose distributed adversarial training (DAT), a large-batch adversarial training framework implemented over multiple machines.
no code implementations • 11 Jun 2022 • Sagar Shrestha, Xiao Fu, Mingyi Hong
This work revisits the joint beamforming (BF) and antenna selection (AS) problem, as well as its robust beamforming (RBF) version under imperfect channel state information (CSI).
no code implementations • 4 Jun 2022 • Ioannis Tsaknakis, Bhavya Kailkhura, Sijia Liu, Donald Loveland, James Diffenderfer, Anna Maria Hiszpanski, Mingyi Hong
Existing knowledge integration approaches are limited to using differentiable knowledge source to be compatible with first-order DL training paradigm.
no code implementations • 27 Apr 2022 • Xinwei Zhang, Mingyi Hong, Nicola Elia
Distributed algorithms have been playing an increasingly important role in many applications such as machine learning, signal processing, and control.
1 code implementation • ICLR 2022 • Yimeng Zhang, Yuguang Yao, Jinghan Jia, JinFeng Yi, Mingyi Hong, Shiyu Chang, Sijia Liu
To tackle this problem, we next propose to prepend an autoencoder (AE) to a given (black-box) model so that DS can be trained using variance-reduced ZO optimization.
no code implementations • 28 Dec 2021 • Bingqing Song, Haoran Sun, Wenqiang Pu, Sijia Liu, Mingyi Hong
We then provide a series of theoretical results to further understand the properties of the two approaches.
2 code implementations • 23 Dec 2021 • Yihua Zhang, Guanhua Zhang, Prashant Khanduri, Mingyi Hong, Shiyu Chang, Sijia Liu
We first show that the commonly-used Fast-AT is equivalent to using a stochastic gradient algorithm to solve a linearized BLO problem involving a sign operation.
no code implementations • 30 Oct 2021 • Jian Du, Song Li, Xiangyi Chen, Siheng Chen, Mingyi Hong
The equivalent privacy costs controlled by maintaining the same gradient clipping thresholds and noise powers in each step result in unstable updates and a lower model accuracy when compared to the non-DP counterpart.
no code implementations • 11 Oct 2021 • Siliang Zeng, Tianyi Chen, Alfredo Garcia, Mingyi Hong
The flexibility in our design allows the proposed MARL-CAC algorithm to be used in a {\it fully decentralized} setting, where the agents can only communicate with their neighbors, as well as a {\it federated} setting, where the agents occasionally communicate with a server while optimizing their (partially personalized) local models.
no code implementations • 4 Oct 2021 • Boyi Liu, Jiayang Li, Zhuoran Yang, Hoi-To Wai, Mingyi Hong, Yu Marco Nie, Zhaoran Wang
To regulate a social system comprised of self-interested agents, economic incentives are often required to induce a desirable outcome.
no code implementations • ICLR 2022 • Prashant Khanduri, Haibo Yang, Mingyi Hong, Jia Liu, Hoi To Wai, Sijia Liu
We analyze the optimization and the generalization performance of the proposed framework for the $\ell_2$ loss.
no code implementations • 25 Jun 2021 • Xinwei Zhang, Xiangyi Chen, Mingyi Hong, Zhiwei Steven Wu, JinFeng Yi
Recently, there has been a line of work on incorporating the formal privacy notion of differential privacy with FL.
no code implementations • NeurIPS 2021 • Prashant Khanduri, Pranay Sharma, Haibo Yang, Mingyi Hong, Jia Liu, Ketan Rajawat, Pramod K. Varshney
Despite extensive research, for a generic non-convex FL problem, it is not clear, how to choose the WNs' and the server's update directions, the minibatch sizes, and the local update frequency, so that the WNs use the minimum number of samples and communication rounds to achieve the desired solution.
1 code implementation • 3 May 2021 • Haoran Sun, Wenqiang Pu, Xiao Fu, Tsung-Hui Chang, Mingyi Hong
However, it is often challenging for these approaches to learn in a dynamic environment.
1 code implementation • 1 May 2021 • Sagar Shrestha, Xiao Fu, Mingyi Hong
However, such deep learning (DL)-based SC approaches encounter serious challenges in both off-line model learning (training) and completion (generalization), possibly because the latent state space for generating the radio maps is prohibitively large.
no code implementations • 29 Apr 2021 • Wenqiang Pu, Shahana Ibrahim, Xiao Fu, Mingyi Hong
This work offers a unified stochastic algorithmic framework for large-scale CPD decomposition under a variety of non-Euclidean loss functions.
no code implementations • 25 Feb 2021 • Chi Zhang, Jinghan Jia, Burhaneddin Yaman, Steen Moeller, Sijia Liu, Mingyi Hong, Mehmet Akçakaya
Although deep learning (DL) has received much attention in accelerated MRI, recent studies suggest small perturbations may lead to instabilities in DL-based reconstructions, leading to concern for their clinical application.
no code implementations • NeurIPS 2021 • Prashant Khanduri, Siliang Zeng, Mingyi Hong, Hoi-To Wai, Zhaoran Wang, Zhuoran Yang
We focus on bilevel problems where the lower level subproblem is strongly-convex and the upper level objective function is smooth.
1 code implementation • 14 Feb 2021 • Shixiang Chen, Alfredo Garcia, Mingyi Hong, Shahin Shahrampour
The global function is represented as a finite sum of smooth local functions, where each local function is associated with one agent and agents communicate with each other over an undirected connected graph.
no code implementations • 22 Jan 2021 • Shixiang Chen, Alfredo Garcia, Mingyi Hong, Shahin Shahrampour
We study the convergence properties of Riemannian gradient method for solving the consensus problem (for an undirected connected graph) over the Stiefel manifold.
no code implementations • ICLR 2021 • Naichen Shi, Dawei Li, Mingyi Hong, Ruoyu Sun
Removing this assumption allows us to establish a phase transition from divergence to non-divergence for RMSProp.
no code implementations • 31 Dec 2020 • Han Shen, Kaiqing Zhang, Mingyi Hong, Tianyi Chen
Asynchronous and parallel implementation of standard reinforcement learning (RL) algorithms is a key enabler of the tremendous success of modern RL.
1 code implementation • 22 Dec 2020 • Xinwei Zhang, Wotao Yin, Mingyi Hong, Tianyi Chen
To the best of our knowledge, this is the first formulation and algorithm developed for the hybrid FL.
no code implementations • NeurIPS 2020 • Hoi-To Wai, Zhuoran Yang, Zhaoran Wang, Mingyi Hong
This paper studies a gradient temporal difference (GTD) algorithm using neural network (NN) function approximators to minimize the mean squared Bellman error (MSBE).
no code implementations • NeurIPS 2020 • Songtao Lu, Meisam Razaviyayn, Bo Yang, Kejun Huang, Mingyi Hong
To the best of our knowledge, this is the first time that first-order algorithms with polynomial per-iteration complexity and global sublinear rate are designed to find SOSPs of the important class of non-convex problems with linear constraints (almost surely).
4 code implementations • 16 Nov 2020 • Haoran Sun, Wenqiang Pu, Minghe Zhu, Xiao Fu, Tsung-Hui Chang, Mingyi Hong
We propose to build the notion of continual learning (CL) into the modeling process of learning wireless systems, so that the learning model can incrementally adapt to the new episodes, {\it without forgetting} knowledge learned from the previous episodes.
no code implementations • 8 Nov 2020 • Minghe Zhu, Tsung-Hui Chang, Mingyi Hong
It is well-known that the problem of finding the optimal beamformers in massive multiple-input multiple-output (MIMO) networks is challenging because of its non-convexity, and conventional optimization based algorithms suffer from high computational costs.
no code implementations • 19 Aug 2020 • Lingyun Zhou, Xihan Chen, Mingyi Hong, Shi Jin, Qingjiang Shi
Unmanned aerial vehicle (UAV) swarm has emerged as a promising novel paradigm to achieve better coverage and higher capacity for future wireless network by exploiting the more favorable line-of-sight (LoS) propagation.
no code implementations • 10 Jul 2020 • Mingyi Hong, Hoi-To Wai, Zhaoran Wang, Zhuoran Yang
Bilevel optimization is a class of problems which exhibit a two-level structure, and its goal is to minimize an outer objective function with variables which are constrained to be the optimal solution to an (inner) optimization problem.
no code implementations • NeurIPS 2020 • Xiangyi Chen, Zhiwei Steven Wu, Mingyi Hong
Deep learning models are increasingly popular in many machine learning applications where the training data may contain sensitive information.
no code implementations • 24 Jun 2020 • Yingxue Zhou, Xiangyi Chen, Mingyi Hong, Zhiwei Steven Wu, Arindam Banerjee
We obtain this rate by providing the first analyses on a collection of private gradient-based methods, including adaptive algorithms DP RMSProp and DP Adam.
no code implementations • 20 Jun 2020 • Mingyi Hong, Siliang Zeng, Junyu Zhang, Haoran Sun
However, by constructing some counter-examples, we show that when certain local Lipschitz conditions (LLC) on the local function gradient $\nabla f_i$'s are not satisfied, most of the existing decentralized algorithms diverge, even if the global Lipschitz condition (GLC) is satisfied, where the sum function $f$ has Lipschitz gradient.
no code implementations • 15 Jun 2020 • Meisam Razaviyayn, Tianjian Huang, Songtao Lu, Maher Nouiehed, Maziar Sanjabi, Mingyi Hong
The min-max optimization problem, also known as the saddle point problem, is a classical optimization problem which is also studied in the context of zero-sum games.
1 code implementation • 22 May 2020 • Xinwei Zhang, Mingyi Hong, Sairaj Dhople, Wotao Yin, Yang Liu
Aiming at designing FL algorithms that are provably fast and require as few assumptions as possible, we propose a new algorithm design strategy from the primal-dual optimization perspective.
no code implementations • 14 Jan 2020 • Tsung-Hui Chang, Mingyi Hong, Hoi-To Wai, Xinwei Zhang, Songtao Lu
In particular, we {provide a selective review} about the recent techniques developed for optimizing non-convex models (i. e., problem classes), processing batch and streaming data (i. e., data types), over the networks in a distributed manner (i. e., communication and computation paradigm).
no code implementations • 24 Dec 2019 • Yang Liu, Yan Kang, Xinwei Zhang, Liping Li, Yong Cheng, Tianjian Chen, Mingyi Hong, Qiang Yang
We introduce a collaborative learning framework allowing multiple parties having different sets of attributes about the same user to jointly build models without exposing their raw data or model parameters.
no code implementations • 16 Dec 2019 • Seyed Amir Hossein Hosseini, Burhaneddin Yaman, Steen Moeller, Mingyi Hong, Mehmet Akçakaya
These methods unroll iterative optimization algorithms to solve the inverse problem objective function, by alternating between domain-specific data consistency and data-driven regularization via neural networks.
no code implementations • NeurIPS 2019 • Hoi-To Wai, Mingyi Hong, Zhuoran Yang, Zhaoran Wang, Kexin Tang
Policy evaluation with smooth and nonlinear function approximation has shown great potential for reinforcement learning.
no code implementations • NeurIPS 2019 • Zhuoran Yang, Yongxin Chen, Mingyi Hong, Zhaoran Wang
Despite the empirical success of the actor-critic algorithm, its theoretical understanding lags behind.
no code implementations • 28 Nov 2019 • Guoyong Zhang, Xiao Fu, Jun Wang, Xi-Le Zhao, Mingyi Hong
Spectrum cartography aims at estimating power propagation patterns over a geographical region across multiple frequency bands (i. e., a radio map)---from limited samples taken sparsely over the region.
1 code implementation • NeurIPS 2019 • Xiangyi Chen, Sijia Liu, Kaidi Xu, Xingguo Li, Xue Lin, Mingyi Hong, David Cox
In this paper, we propose a zeroth-order AdaMM (ZO-AdaMM) algorithm, that generalizes AdaMM to the gradient-free regime.
no code implementations • 13 Oct 2019 • Haoran Sun, Songtao Lu, Mingyi Hong
Similarly, for online problems, the proposed method achieves an $\mathcal{O}(m \epsilon^{-3/2})$ sample complexity and an $\mathcal{O}(\epsilon^{-1})$ communication complexity, while the best existing bounds are $\mathcal{O}(m\epsilon^{-2})$ and $\mathcal{O}(\epsilon^{-2})$, respectively.
no code implementations • 14 Jul 2019 • Zhuoran Yang, Yongxin Chen, Mingyi Hong, Zhaoran Wang
Despite the empirical success of the actor-critic algorithm, its theoretical understanding lags behind.
no code implementations • 9 Jul 2019 • Songtao Lu, Meisam Razaviyayn, Bo Yang, Kejun Huang, Mingyi Hong
This paper proposes low-complexity algorithms for finding approximate second-order stationary points (SOSPs) of problems with smooth non-convex objective and linear constraints.
1 code implementation • 10 Jun 2019 • Yi Wei, Ming-Min Zhao, Mingyi Hong, Min-Jian Zhao, Ming Lei
Furthermore, in order to reduce the memory costs, a novel quantized LcgNet is proposed, where a low-resolution nonuniform quantizer is integrated into the LcgNet to smartly quantize the aforementioned step-sizes.
1 code implementation • 10 Jun 2019 • Kaidi Xu, Hongge Chen, Sijia Liu, Pin-Yu Chen, Tsui-Wei Weng, Mingyi Hong, Xue Lin
Graph neural networks (GNNs) which apply the deep neural networks to graph data have achieved significant performance for the task of semi-supervised node classification.
no code implementations • NeurIPS 2020 • Xiangyi Chen, Tiancong Chen, Haoran Sun, Zhiwei Steven Wu, Mingyi Hong
We show that these algorithms are non-convergent whenever there is some disparity between the expected median and mean over the local gradients.
no code implementations • ICLR 2019 • Songtao Lu, Rahul Singh, Xiangyi Chen, Yongxin Chen, Mingyi Hong
By developing new primal-dual optimization tools, we show that, with a proper stepsize choice, the widely used first-order iterative algorithm in training GANs would in fact converge to a stationary solution with a sublinear rate.
no code implementations • ICLR 2019 • Sijia Liu, Pin-Yu Chen, Xiangyi Chen, Mingyi Hong
Our study shows that ZO signSGD requires $\sqrt{d}$ times more iterations than signSGD, leading to a convergence rate of $O(\sqrt{d}/\sqrt{T})$ under mild conditions, where $d$ is the number of optimization variables, and $T$ is the number of iterations.
no code implementations • 21 Feb 2019 • Songtao Lu, Ioannis Tsaknakis, Mingyi Hong, Yongxin Chen
In this work, we consider a block-wise one-sided non-convex min-max problem, in which the minimization problem consists of multiple blocks and is non-convex, while the maximization problem is (strongly) concave.
no code implementations • 11 Jan 2019 • Qi Cai, Mingyi Hong, Yongxin Chen, Zhaoran Wang
We study the global convergence of generative adversarial imitation learning for linear quadratic regulators, which is posed as minimax optimization.
no code implementations • ICLR 2019 • Xiangyi Chen, Sijia Liu, Ruoyu Sun, Mingyi Hong
We prove that under our derived conditions, these methods can achieve the convergence rate of order $O(\log{T}/\sqrt{T})$ for nonconvex stochastic optimization.
no code implementations • ICML 2018 • Mingyi Hong, Meisam Razaviyayn, Jason Lee
In this work, we study two first-order primal-dual based algorithms, the Gradient Primal-Dual Algorithm (GPDA) and the Gradient Alternating Direction Method of Multipliers (GADMM), for solving a class of linearly constrained non-convex optimization problems.
no code implementations • NeurIPS 2018 • Hoi-To Wai, Zhuoran Yang, Zhaoran Wang, Mingyi Hong
Despite the success of single-agent reinforcement learning, multi-agent reinforcement learning (MARL) remains challenging due to complex interactions between agents.
Multi-agent Reinforcement Learning
reinforcement-learning
+2
no code implementations • 24 Apr 2018 • Charilaos I. Kanatsoulis, Xiao Fu, Nicholas D. Sidiropoulos, Mingyi Hong
In this work, we propose a new computational framework for large-scale SUMCOR GCCA that can easily incorporate a suite of structural regularizers which are frequently used in data analytics.
no code implementations • 28 Feb 2018 • Songtao Lu, Mingyi Hong, Zhengdao Wang
The alternating gradient descent (AGD) is a simple but popular algorithm which has been applied to problems in optimization, machine learning, data ming, and signal processing, etc.
no code implementations • 27 Oct 2017 • Davood Hajinezhad, Mingyi Hong, Alfredo Garcia
In this paper, we consider distributed optimization problems over a multi-agent network, where each agent can only partially evaluate the objective function, and it is allowed to exchange messages with its immediate neighbors.
no code implementations • ICML 2017 • Mingyi Hong, Davood Hajinezhad, Ming-Min Zhao
In this paper we consider nonconvex optimization and learning over a network of distributed nodes.
no code implementations • 24 Mar 2017 • Songtao Lu, Mingyi Hong, Zhengdao Wang
The proposed algorithm is guaranteed to converge to the set of Karush-Kuhn-Tucker (KKT) points of the nonconvex SymNMF problem.
11 code implementations • ICML 2017 • Bo Yang, Xiao Fu, Nicholas D. Sidiropoulos, Mingyi Hong
To recover the `clustering-friendly' latent representations and to better cluster the data, we propose a joint DR and K-means clustering approach in which DR is accomplished via learning a deep neural network (DNN).
no code implementations • 10 Jul 2016 • Xingguo Li, Tuo Zhao, Raman Arora, Han Liu, Mingyi Hong
In particular, we first show that for a family of quadratic minimization problems, the iteration complexity $\mathcal{O}(\log^2(p)\cdot\log(1/\epsilon))$ of the CBCD-type methods matches that of the GD methods in term of dependency on $p$, up to a $\log^2 p$ factor.
no code implementations • 31 May 2016 • Xiao Fu, Kejun Huang, Mingyi Hong, Nicholas D. Sidiropoulos, Anthony Man-Cho So
Generalized canonical correlation analysis (GCCA) aims at finding latent low-dimensional common structure from multiple views (feature vectors in different domains) of the same entities.
no code implementations • NeurIPS 2016 • Davood Hajinezhad, Mingyi Hong, Tuo Zhao, Zhaoran Wang
We study a stochastic and distributed algorithm for nonconvex problems whose objective consists of a sum of $N$ nonconvex $L_i/N$-smooth functions, plus a nonsmooth regularizer.
no code implementations • 25 May 2016 • Xingguo Li, Haoming Jiang, Jarvis Haupt, Raman Arora, Han Liu, Mingyi Hong, Tuo Zhao
Many machine learning techniques sacrifice convenient computational structures to gain estimation robustness and modeling flexibility.
no code implementations • 28 Nov 2015 • Mingyi Hong, Tsung-Hui Chang
We consider solving a convex, possibly stochastic optimization problem over a randomly time-varying multi-agent network.
Optimization and Control Information Theory Information Theory
no code implementations • 9 Sep 2015 • Tsung-Hui Chang, Mingyi Hong, Wei-Cheng Liao, Xiangfeng Wang
By formulating the learning problem as a consensus problem, the ADMM can be used to solve the consensus problem in a fully parallel fashion over a computer network with a star topology.
no code implementations • 9 Sep 2015 • Tsung-Hui Chang, Wei-Cheng Liao, Mingyi Hong, Xiangfeng Wang
Unfortunately, a direct synchronous implementation of such algorithm does not scale well with the problem size, as the algorithm speed is limited by the slowest computing nodes.
1 code implementation • NeurIPS 2014 • Meisam Razaviyayn, Mingyi Hong, Zhi-Quan Luo, Jong-Shi Pang
In this work, we propose an inexact parallel BCD approach where at each iteration, a subset of the variables is updated in parallel by minimizing convex approximations of the original objective function.
Optimization and Control
no code implementations • 21 Jan 2014 • Brendan P. W. Ames, Mingyi Hong
To accomplish this task, we propose a heuristic, called sparse zero-variance discriminant analysis (SZVD), for simultaneously performing linear discriminant analysis and feature selection on high dimensional data.
no code implementations • 11 Sep 2012 • Meisam Razaviyayn, Mingyi Hong, Zhi-Quan Luo
The block coordinate descent (BCD) method is widely used for minimizing a continuous function f of several block variables.
Optimization and Control