 |
Somayeh Sojoudi Electrical Engineering and Computer Sciences University of California, Berkeley Email: sojoudi (at) berkeley.edu Office: 5114 Etcheverry Hall |
Brief Biography
I am an Assistant Professor in Electrical Engineering and Computer Sciences and Mechanical Engineeing departments at UC Berkeley. I am also on the faculty of the Tsinghua-Berkeley Shenzhen Institute (TBSI). I work on various problems in optimization theory, machine learning, algorithms, data science, and control theory.
Open postdoctoral positions
Several postdoctoral positions are available in my group in the broad areas of optimization theory and machine learning. To apply, please send me a CV and three sample publications.
News
September 2019: New paper on learning and control: “Efficient Learning of Distributed Linear-Quadratic Controllers”.
September 2019: New paper on homotopy and optimization: "Homotopy Method for Finding the Global Solution of Post-contingency Optimal Power Flow”.
August 2019: Our paper “Global Optimality Guarantees for Nonconvex Unsupervised Video Segmentation” to appear in 57th Annual Allerton Conference on Communication, Control, and Computing, 2019.
July 2019: Our paper “Motion Segmentation and Synthesis for Latency Mitigation in a Cloud Robotic Tele-Operation System” to appear in The International Symposium on Robotics Research (ISRR), 2019.
July 2019: Our paper “Learning Sparse Dynamical Systems from a Single Sample Trajectory” to appear in IEEE Conference on Decision and Control, 2019.
July 2019: New paper on learning: “Global Optimality Guarantees for Nonconvex Unsupervised Video Segmentation”.
June 2019: Our paper “Sharp Restricted Isometry Bounds for the Inexistence of Spurious Local Minima in Nonconvex Matrix Recovery” to appear in the Journal of Machine Learning Research.
May 2019: New paper on robust learning: “Power up! Robust Graph Convolutional Network against Evasion Attacks based on Graph Powering”.
May 2019: New paper on online optimization: “Absence of spurious local trajectories in time-varying optimization”.
April 2019: Cedric Josz will join Columbia University as an Assistant Professor in Fall 2019.
March 2019: New paper on Fog Robotics: “Mitigating Network Latency for Fog Robotics using Motion Segmentation and Synthesis”.
March 2019: New paper on sparse system identification: “Learning Sparse Dynamical Systems from a Single Sample Trajectory”.
January 2019: Our paper “Mathematical Models of Physiological Responses to Exercise” to appear in the 2019 American Control Conference.
January 2019: Our paper “A Fog Robotic System for Dynamic Visual Servoing” to appear in the 2019 International Conference on Robotics and Automation (ICRA).
December 2018: New paper on non-smooth optimization for learning: “Exact Guarantees on the Absence of Spurious Local Minima for Non-negative Robust Principal Component Analysis”.
December 2018: Our paper “Transient Stability Analysis of Power Systems via Occupation Measures” to appear in IEEE PES Innovative Smart Grid Technologies Conference, 2019.
December 2018: Our paper “Graphical Lasso and Thresholding: Equivalence and Closed-form Solutions” to appear in the Journal of Machine Learning Research.
December 2018: Our paper “Linear-Time Algorithm for Learning Large-Scale Sparse Graphical Models” to appear in IEEE Access.
December 2018: I received a grant from Air Force Office of Scientific Research (AFOSR).
November 2018: Our paper “Conic Optimization for Control, Energy Systems, and Machine Learning: Applications and Algorithms” to appear in Annual Reviews in Control.
November 2018: Our paper “Graphical Lasso and Thresholding: Equivalence and Closed-form Solutions” won the INFORMS 2018 Data Mining Best Paper Award.
October 2018: The Berkeley team led by me has been selected to receive funding to participate in the ARPA-E $4M Cash Prize Grid Optimization (GO) Competition (link 1, link 2).
September 2018: Video of my talk on conic optimization and graphical models: “Learning Large-Scale Sparse Graphical Models: Theory, Algorithm, and Applications”
September 2018: I received a seed grant from Siebel Energy Institute.
September 2018: I received a grant from NSF.
September 2018: New paper: “A Fog Robotic System for Dynamic Visual Servoing”
September 2018: Two papers to appear in 32nd Annual Conference on Neural Information Processing Systems (NIPS): “How Much Restricted Isometry is Needed In Nonconvex Matrix Recovery?” and “A Theory on the Absence of Spurious Optimality”.
August 2018: Elizabeth Glista has joined my group as a PhD student.
August 2018: Brendon Anderson has joined my group as a PhD student.
August 2018: New paper on statistical analysis of system identification: Sample Complexity of Sparse System Identification Problem.
August 2018: New tutorial paper: Conic Optimization With Applications to Machine Learning and Energy Systems.
August 2018: Our paper Data-Driven Sparse System Identification to appear in 56th Annual Allerton Conference on Communication, Control, and Computing, 2018.
July 2018: I have been appointed as an Associate Editor of IEEE Access.
July 2018: Our paper Non-Asymptotic Analysis of Block-Regularized Regression Problem to appear in IEEE Conference on Decision and Control, 2018.
June 2018: I received a grant from the Office of Naval Research.
May 2018: New paper on spurious local minima for nonsmooth optimization: “A Theory on the Absence of Spurious Optimality”
May 2018: New paper on spurious local minima for nonconvex optimization: “How Much Restricted Isometry is Needed In Nonconvex Matrix Recovery?”
May 2018: Our paper A Cloud-Based Robust Semaphore Mirroring System for Social Robots has been accepted for the 2018 IEEE International Conference on Automation Science and Engineering.
May 2018: Our paper Large-Scale Sparse Inverse Covariance Estimation via Thresholding and Max-Det Matrix Completion has been accepted for the 2018 International Conference on Machine Learning (ICML).
May 2018: I will teach a 16-hour short course on Optimization and Machine Learning in Shenzhen, China, May 8-13.
April 2018: Salar Fattahi won the Katta G. Murty Best Paper Prize in Optimization from Berkeley IEOR for the paper Graphical Lasso and Thresholding: Equivalence and Closed-form Solutions.
April 2018: My group has organized a session on sparse semidefinite programming for 23rd International Symposium on Mathematical Programming.
March 2018: New paper on sparse system identification: Data-Driven Sparse System Identification .
March 2018: New paper on system identification for physiology: Mathematical Models of Physiological Responses to Exercise .
March 2018: New paper on statistical learning: Non-Asymptotic Analysis of Block-Regularized Regression Problem .
March 2018: New paper on stability of power systems: Transient Stability Analysis of Power Systems via Occupation Measures .
March 2018: New paper on numerical algorithm for learning graphical models: Large-Scale Sparse Inverse Covariance Estimation via Thresholding and Max-Det Matrix Completion.
February 2018: Our paper Sparse Inverse Covariance Estimation for Chordal Structures has been accepted for the 2018 European Control Conference.
January 2018: Our paper “Closed-form Solution and Sparsity Path for Inverse Covariance Estimation Problem” has been selected as one of the five best student award finalists for the American Control Conference 2018.
January 2018: Our papers “Closed-form Solution and Sparsity Path for Inverse Covariance Estimation Problem” and “Conic Optimization Theory: Convexification Techniques and Numerical Algorithms” have been accepted for the 2018 American Control Conference.
January 2018: Our tutorial session “Optimization and Machine Learning: Theory, Algorithms, and Applications” has been accepted for the 2018 American Control Conference (contributors: Sam Coogan, Murat Arcak, Javad Lavaei, Richard Zhang, Cedric Josz, and Somayeh Sojoudi).
December 2017: Our paper Convexification of Generalized Network Flow Problem to appear in Mathematical Programming.
December 2017: Video of my talk at GraphXD seminar series: Data-Driven Methods for Learning Sparse Graphical Models.
December 2017: I have given talks at McGill University, University of Melbourne and UC Berkeley's GraphXD seminar series on learning sparse graphical models.
November 2017: New paper on structure learning: Sparse Inverse Covariance Estimation for Chordal Structures.
November 2017: I have become an affiliate faculty member of the Tsinghua-Berkeley-Shenzhen-Institute.
October 2017: New tutorial paper on conic optimization: Conic Optimization Theory: Convexification Techniques and Numerical Algorithms
September 2017: I have been appointed as an Associate Editor of Systems & Control Letters.
September 2017: New paper on graphical models: Graphical Lasso and Thresholding: Equivalence and Closed-form Solutions.
August 2017: Dr. Cedric Josz has joined my group as a postdoctoral scholar.
August 2017: I will give a talk on machine learning at INFORMS 2017.
August 2017: I gave a seminar at Hong Kong University of Science and Technology.
July 2017: I have joined the Departments of Electrical Engineering and Computer Sciences and Mechanical Engineering as an Assistant Professor in Residence.
June 2017: I have been appointed as a Conference Editorial Board (CEB) Associate Editor of the IEEE Control Systems Society.
May 2017: I have given talks at University of Michigan (EECS), UC Berkeley (IEOR), Georgia Institute of Technology (CSE), UCLA (EE), Rice (EE and CAAM), Ohio State University (ISE), UIUC (ECE), Georgia Institute of Technology (ISyE and ECE) and UC Berkeley (EECS and ME) in the Spring semester.
December 2016: The paper Finding Low-rank Solutions of Sparse Linear Matrix Inequalities using Convex Optimization to appear in SIAM Journal on Optimization.
October 2016: I will give talks at UC San Diego, UT Austin, UC Berkeley, UCLA, Caltech and Stanford University in the Fall semester.
September 2016: Our paper Convex Relaxation for Optimal Power Flow Problem: Mesh Networks has received the 2016 INFORMS ENRE Energy Best Publication Award.
August 2016: I have been appointed as an Associate Editor of IEEE Transactions on Smart Grid.
August 2016: I will give the tutorial talk “Data-Driven Computational Methods for Sparse Network Estimation” at IEEE Multi-Conference on Systems and Control 2016.
July 2016: The paper Graphical Lasso and Thresholding: Conditions for Equivalence to appear in 55th IEEE Conference on Decision and Control 2016.
July 2016: I will give a talk at INFORMS Annual Meeting 2016 on machine learning.
June 2016: I will give a talk at International Conference on Continuous Optimization on machine learning.
June 2016: Our article Semidefinite Relaxations for Nonlinear Optimization Problems has appeared in the INFORMS Optimization Society Newsletter 2016.
May 2016: The paper “Equivalence of Graphical Lasso and Thresholding for Sparse Graphs” to appear in Journal of Machine Learning Research.
March 2016: I will give a talk at INFORMS International Conference on data analytics and graphical lasso.
February 2016: New paper on machine learning : Graphical Lasso and Thresholding: Conditions for Equivalence
October 2015: I am a co-recipient of the INFORMS Optimization Society Prize for Young Researchers.
August 2015: Slides of my talks at Princeton, Johns Hopkins University and UC Berkeley: Computational Methods for Complex Systems: From Brain to Power Networks
June 2015: Two new papers on brain connectivity using ECoG: Causal Cortical Interactions in Speech Processing: Event Related Analysis of Human ECoG and Large-Scale Cortical Interactions During Auditory, Visual and Audiovisual Speech Processing: Event Related Analysis of Human ECoG
May 2015: The paper Corpus Callosum Area and Brain Volume in Autism Spectrum Disorder: Quantitative Analysis of Structural MRI from the ABIDE Database to appear in Journal of Autism and Developmental Disorders.
January 2015: Video of my talk at CMU: Computational Methods for the Analysis and Modeling of Brain Networks
December 2014: I am a co-author of a best student paper award finalist for the 53rd IEEE Conference on Decision and Control.
December 2014: Tutorial paper for the 53rd IEEE Conference on Decision and Control: Graph-Theoretic Algorithms for Polynomial Optimization Problems
October 2014: I chaired the session “Control Applications and Nonlinear Control” for IEEE 52nd Annual Allerton Conference on Communication, Control and Computing, 2014.
October 2014: I co-chaired the session ‘‘Stochastic Systems and Control" for IEEE 52nd Annual Allerton Conference on Communication, Control and Computing, 2014.
October 2014: Two papers accepted in Allerton: Study of the Brain Functional Network Using Synthetic Data and Equivalence of Graphical Lasso and Thresholding for Sparse Graphs
October 2014: New paper on low-rank optimization: “Finding Low-rank Solutions of Sparse Linear Matrix Inequalities using Convex Optimization”
July 2014: The paper Exactness of Semidefinite Relaxations for Nonlinear Optimization Problems with Underlying Graph Structure to appear in SIAM Journal on Optimization
July 2014: The paper “Low-Rank Solutions of Matrix Inequalities With Applications to Polynomial Optimization and Matrix Completion Problems” to appear in IEEE Conference on Decision and Control 2014.