About
Group
Research
Publications
News Archive

News Archive

  • May, 2024: 2 papers on performative prediction and contrastive learning accepted at ICML 2024:

 
 

  • April, 2024: I have visited Hanyang University and gave a talk on Graph Learning with Low Pass Graph Signal Processing at Sunwoo Kim's group. I have also attended IEEE ICASSP 2024 in Seoul.

  • April, 2024: 1 paper accepted at IEEE SAM 2024: On Detecting Low-pass Graph Signals under Partial Observations (with Sean)  —  it's an extension over our prior work for detecting if a set of graph signals is low pass or not without knowing the graph topology, now under partially observed nodes. I am also organizing a special session on ‘Recent Advances on Graph Signal Processing’.

 

  • December, 2023: Paper on ‘‘Learning Multiplex Graph with Inter-layer Coupling’’ (with Chenyue) accepted at ICASSP 2024. I have also given a presentation of this work at the LoG Meetup Shanghai in Oct - Slides.

 

  • November, 2023: With Mingyi Hong, we have organized a special session for Bilevel Optimization at Asilomar SSC. I have also visited UC Davis and gave a talk on ‘‘Graph Learning with Low Pass Graph Signal Processing’’ at the MADDD Seminar series. Thanks Naoki for hosting me.

  • July, 2023: Paper on ‘‘DoCoM: Compressed Decentralized Optimization with Near-Optimal Sample Complexity’’ (with Oscar) accepted by TMLR. Congrats to Oscar for his first journal paper. Accepted version is available at Openreview, also find the 10-minutes pitch on this work on Youtube.

 

  • July, 2023: Overview paper on ‘‘Stochastic Approximation Beyond Gradient for Signal Processing and Machine Learning’’ (with Aymeric, Gersende and Eric) accepted by IEEE TSP. In this paper, we provide a detailed overview on the recent state-of-the-art algorithms and theoretical analysis of non-gradient stochastic approximation algorithms. The unifying framework provides a convenient tool for studying algorithms such as stochastic EM, SGD with compression, TD learning, etc. Check out the preprint at here.

 

  • July, 2023: 2 Papers accepted by IEEE CDC 2023, see you in Singapore!

    • Fully Stochastic Distributed Convex Optimization on Time-Varying Graph with Compression (with Oscar)

    • Linear Speedup of Incremental Aggregated Gradient Methods on Streaming Data (with Xiaolu, Jin, Yuantao)

  • June, 2023: I have attended GSP Workshop 2023 and gave a plenary talk on Low Pass GSP. Slides PDF. Recordings can be found on Youtube.

 

  • April, 2023: 1 Paper accepted by ICML 2023 on ‘‘Network Effects on Performative Prediction Games’’ (with Xiaolu and Oscar). See the accompanying slides on this work. Preprint/camera ready can be found here.

MultiPP 

  • April, 2023: I have visited Universidad Rey Juan Carlos in Madrid and gave a seminar on ‘‘Stochastic Approximation Schemes with Decision Dependent Data’’. Slides PDF. Thanks Antonio for hosting me.

  • April, 2023: I have attended the Workshop on Games on networks in IMS, National University of Singapore. I presented about the work on ‘‘Network Effects on Performative Prediction Games’’ (with Xiaolu and Oscar). Slides PDF.

MultiPP 

  • February, 2023: Two papers accepted by ICASSP 2023, see you in Rhodes!

    • Product Graph Learning from Multi-attribute Graph Signals with Inter-layer Coupling (with Chenyue and Yiran, Congrats to Chenyue for 1st ICASSP!) preprint

ICASSP 

    • Central Nodes Detection from Partially Observed Graph Signals (with Yiran)

ICASSP 

  • Jan, 2023: Paper accepted by AISTATS 2023, see you in Valencia!

    • Incremental Aggregated Riemannian Gradient Method for Distributed PCA (with Xiaolu, Yuchen and Yuantao)

AISTATS 

  • Jan, 2023: Paper accepted by IEEE TSP IEEEXplore.

    • Online Inference for Mixture Model of Streaming Graph Signals with Sparse Excitation (with Yiran)

TSP 

  • September, 2022: Three papers accepted at NeurIPS 2022

    • Multi-agent Performative Prediction with Greedy Deployment and Consensus Seeking Agents (with Qiang and Oscar) preprint

NeurIPS 

    • Distributed Optimization for Overparameterized Problems: Achieving Optimal Dimension Independent Communication Complexity (with Bingqing, Ioannis, Oscar and Mingyi)

NeurIPS 

    • Inducing Equilibria via Incentives: Simultaneous Design-and-Play Ensures Global Convergence (with Boyi, Jiaying, Zhuoran, Mingyi, Yu and Zhaoran) preprint

  • August, 2022: Paper accepted at SIOPT

    • A two-timescale framework for bilevel optimization: Complexity analysis and application to actor-critic (with Mingyi, Zhaoran, Zhuoran) preprint

SIOPT 

  • July, 2022: Two papers accepted at IEEE CDC

    • On the Role of Data Homogeneity in Multi-Agent Non-convex Stochastic Optimization (with Qiang) preprint

CDC 

    • Stochastic Gradient Tracking Methods for Distributed Personalized Optimization over Networks (with Jimmy Xu et al.)

  • June, 2022: Paper accepted at IEEE TSP

    • Robust Distributed Optimization With Randomly Corrupted Gradients (with Berkay, Cesar and Mahnoosh) preprint

TSP 

  • May, 2022: I have given a seminar at the NUS's IORA Seminar on Stochastic Approximation Schemes with Decision Dependent Data. Thanks Wang-chi for hosting me. Slides

ICASSP 

  • April, 2022: Two papers accepted at IEEE TSP

    • Community Inference from Partially Observed Graph Signals: Algorithms and Analysis (with Yonina, Anna and Asu) paper

TSP 

  • Detecting Central Nodes from Low-rank Excited Graph Signals via Structured Factor Analysis (with Yiran - Congrats to Yiran for her first TSP paper!) paper

TSP 

  • January, 2022: Two papers accepted at ICASSP 2022

    • Joint Centrality Estimation and Graph Identification from Mixture of Low Pass Graph Signals (with Yiran)

    • On the Stability of Low Pass Graph Filter With a Large Number of Edge Rewires (with Sean and Yiran - Congrats to Sean for his first paper!) preprint

ICASSP 

  • January, 2022: One paper accepted at AISTATS 2022

    • State Dependent Performative Prediction with Stochastic Approximation (with Qiang - Congrats to Qiang for his first paper!) preprint

AISTATS 

  • January, 2022: One paper accepted at ICLR 2022

    • Decentralized Learning for Overparameterized Problems: A Multi-Agent Kernel Approximation Approach (with Prashant et al.) pre-proceeding

  • January, 2022: One paper accepted at ALT 2022

    • MISSO: Minimization by Incremental Stochastic Surrogate Optimization for Large Scale Nonconvex Problems (with Belhal et al.)

  • December, 2021: I have given a virtual seminar at the UMass Lowell's CS Colloquim on Low Pass Graph Signal Processing - Applications and Beyond. Thanks Yimin for hosting me.

  • September, 2021: Two papers accepted at NeurIPS 2021:

    • Tight High Probability Bounds for Linear Stochastic Approximation with Fixed Stepsize (with Alain et al.) pre-proceeding.

    • A near-optimal algorithm for stochastic bilevel optimization via double-momentum (with Prashant et al.) pre-proceeding.

  • August, 2021: I have given a tutorial at the ZJU-CSE Summer School 2021 on Decentralized learning in the nonconvex world: Recent results. Thanks Jinming for the invitation.

  • July, 2021: One paper accepted at APSIPA 2021:

    • An Empirical Study on Compressed Decentralized Stochastic Gradient Algorithms with Overparameterized Models (with Arjun - Congrats to Arjun for his first paper!) preprint.

  • May, 2021: The paper “On the Stability of Random Matrix Product with Markovian Noise: Application to Linear Stochastic Approximation and TD Learning” (with Alain, Eric, Alexey and Sergey) is accepted at COLT 2021, see the preprint.

  • May, 2021: The paper “On Robustness of the Normalized Subgradient Method with Randomly Corrupted Subgradients” (with Berkay, Cesar and Mahnoosh) is accepted and presented at ACC 2021, see the preprint.

ACC 

  • February, 2021: Three papers accepted at ICASSP 2021:

    • Idenftifying First-Order Lowpass Graph Signals Using Perron Frobenius Theorem (with Yiran) preprint

    • Provably Fast Asynchronous and Distributed Algorithms for PageRank Centrality Computation (with Yiran)

    • GEOM-SPIDER-EM: Faster Variance Reduced Stochastic Expectation Maximization for Nonconvex Finite-sum Optimization (with Gersende and Eric) preprint

ICASSP 

  • January, 2021: With Prof. Usman Khan (Tufts), we are organizing a special track on Signal Processing for Self-Aware and Social Autonomous Systems at the inaugural IEEE-ICAS in Montreal this August! The deadline of submission is March 3, 2021.

  • December, 2020: New paper Federated Block Coordinate Descent Scheme for Learning Global and Personalized Models (with Ryan, Anna, Nurullah, Kari and Ken) accepted at AAAI 2021 preprint

AAAI 

  • September, 2020: Two papers accepted at NeurIPS 2020:

    • Provably Efficient Neural GTD for Off-Policy Learning (with Zhuoran, Zhaoran, Mingyi) pre-proceeding.

    • A Stochastic Path Integral Differential EstimatoR Expectation Maximization Algorithm (with Gersende, Eric) pre-proceeding.

NeurIPS 

  • September, 2020: New paper Resilient Primal-Dual Optimization Algorithms for Distributed Resource Allocation accepted by IEEE Transactions on Control of Networked Systems (with Burkay, Cesar, Mahnoosh). This paper studies how to robustify distributed primal-dual optimization using a median filtering for robust mean estimation. Checkout the preprint here

TCNS 

  • August, 2020: New paper A User Guide to Low-Pass Graph Signal Processing and its Applications (with Raksha and Anna) accepted by IEEE SP Magazine for a special issue on Graph Signal Processing. Checkout the preprint here. In this paper, we overview a number of network dynamics models which lead to low-pass graph signals, and demonstrate how to leverage the low-pass properties for making inference and data analytics.

SPM 

  • August, 2020: New paper Exact Blind Community Detection from Signals on Multiple Graphs (with Mitch, Michael and Santiago) accepted by IEEE TSP. Checkout the preprint here. In this paper, we show that communities can be detected exactly from graph signals generated from multiple SBM-PPM graphs, regardless of the spectral resolution limit.

TSP 

  • May, 2020: Our paper, Finite Time Analysis of Linear Two-timescale Stochastic Approximation with Markovian Noise (with Maxim, Eric, Alex and Vladislav), has been accepted by COLT 2020. Checkout the preprint here.

  • Mar., 2020: Our paper, Block-randomized stochastic proximal gradient for low-rank tensor factorization (with Xiao, Shahana, Cheng and Kejun), has been accepted by IEEE TSP. Checkout the paper here.

  • Feb., 2020: Our paper, Accelerating Incremental Gradient Optimization with Curvature Information (with Wilbur, Cesar, Angelia and Anna), has been accepted by Computational Optimization and Applications. The preprint here will be updated soon.

  • Feb., 2020: With Prof. Usman Khan (Tufts), we are organizing a special track on *Signal Processing for Self-Aware and Social Autonomous Systems* at the inaugural [https://2020.ieee-icas.org/ *IEEE-ICAS*] in Montreal this August! Please consider sending your work there The IEEE-ICAS conference has been postponed to 2021 due to the coronavirus outbreak.

  • Feb., 2020: Our paper, Hybrid Inexact BCD for Coupled Structured Matrix Factorization in Hyperspectral Super-Resolution (with Ryan and Ken), has been accepted by IEEE TSP. See the preprint here.

  • Jan., 2020: Our paper, Estimating Centrality Blindly from Low-pass Filtered Graph Signals (with Yiran), has been accepted by IEEE ICASSP 2020. This is Yiran's first paper. Congrats! See the preprint here.

  • Jan., 2020: Our paper, Distributed Learning in the Non-Convex World: From Batch to Streaming Data, and Beyond (with Tsung-hui, Mingyi, Xinwei, Songtao), has been accepted by IEEE SP Magazine as a special issue paper on Distributed, Streaming Machine Learning. This paper provides a selective review on non-convex distributed learning, accompanied with extensive numerical experiments to illustrates some practical concerns in applying distributed algorithms on machine learning. See the preprint here.

  • Dec., 2019: Our paper, Blind community detection from low-rank excitations of a graph filter (with Santiago, Asu, Anna, Ali), has been accepted by IEEE TSP. This paper studies how to bypass graph topology inference to detect the underlying community/clusters of the graph. IEEE Explore.

  • Nov., 2019: I have given a talk at ESD, SUTD on Non-asymptotic Analysis of Biased Stochastic Approximation Scheme.

  • Nov., 2019: I have given a talk on Malicious Agent Detection in Social Network at the CUHK Conference on FinTech link.

  • Oct., 2019: I have given a talk at EECS, OSU on Non-asymptotic Analysis of Biased Stochastic Approximation Scheme.

  • Sept., 2019: Two papers accepted at NeurIPS 2019. This year the acceptance rate was 21.1%, and our papers are titled On the Global Convergence of (Fast) Incremental Expectation Maximization Methods (with Belhal, Eric and Marc), and Variance Reduced Policy Evaluation with Smooth Function Approximation (with Zhuoran, Zhaoran, Mingyi and Kexin). Preprints will be available soon.

  • July, 2019: One paper accepted at IEEE CDC 2019: Resilient Distributed Optimization Algorithms for Resource Allocation (with Cesar and Mahnoosh). See the preprint Here.

  • June, 2019: Our paper, Joint Network Topology and Dynamics Recovery from Perturbed Stationary Points (with Anna, Baruch, Amir), has been accepted by IEEE TSP. This paper studies the identifiability of network structure from stationary points of certain non-linear dynamics. Preprint will be available soon.

  • June, 2019: I have given a talk at LIDS, MIT on Non-asymptotic Analysis of Biased Stochastic Approximation Scheme.

  • Mar, 2019: Our paper, Non-asymptotic Analysis on Biased Stochastic Approximation Scheme (with Belhal, Blazej, Eric), has been accepted by COLT 2019. See the preprint Here.