All Publications

  • 2019
  • 2018
  • "Combining Symbolic Expressions and Black-Box Function Evaluations In Neural Programs"
       Forough Arabshahi, Sameer Singh, Animashree Anandkumar
         Appeared at ICLR 2018, Vancouver, Canada. Also appeared at NIPS 2017, MLtrain workshop, Long Beach, CA.
    Download:    [Code]  [Poster]
  • 2017
  • "Efficient Exploration through Bayesian Deep Q-Networks"
       Kamyar Azizzadenesheli, Animashree Anandkumar
         Appeared at NIPS 2017, Deep RL Workshop, Long Beach, California
    Download:     [Code]
  • "Long-term forecasting using tensor-train RNNs"
       Rose Yu, Stephan Zheng, Anima Anandkumar, Yisong Yue
    Download:     [Code]
  • "Tensor Regression Networks"
       Jean Kossaifi, Zack Chase Lipton, Aran Khanna, Tommaso Furlanello, Anima Anandkumar
    Download:     [Code]
  • "Compact Tensor Pooling for Visual Question Answering"
       Yang Shi, Tommaso Furlanello, Anima Anandkumar
         Appeared at CVPR 2017 VQA Workshop, Honolulu, Hawaii
  • "Analyzing tensor power method dynamics in overcomplete regime"
       Anima Anandkumar, Rong Ge, Majid Janzamin
         Appeared at JMLR
  • "A clustering approach to learning sparsely used overcomplete dictionaries"
       Alekh Agarwal, Animashree Anandkumar, Praneeth Netrapalli
         Appeared at IEEE Transactions on Information Theory 2017
  • "Tensor Contraction Layers for Parsimonious Deep Nets"
       Jean Kossaifi, Aran Khanna, Zachary C. Lipton, Tommaso Furlanello, Anima Anandkumar
         Appeared at CVPR 2017, Workshop, Honolulu, Hawaii
    Download:     [Slides]
  • "Spectral Latent Dirichlet Allocation model on Spark"
       Furong Huang, Jencir Lee, and Anima Anandkumar
    Download:  White paper 2017   [Code]
  • "Reinforcement Learning in Rich-Observation MDPs using Spectral Methods"
       Kamyar Azizzadenesheli, Alessandro Lazaric, Anima Anandkumar
         Appeared at RLDM 2017, Ann Arbor, Michigan, USA
    Download:     [Poster]
  • "Homotopy Analysis for Tensor PCA"
       Anima Anandkumar, Yuan Deng, Rong Ge, Hossein Mobahi
         Appeared at COLT 2017, Amsterdam, Netherlands
    Download:     [Slides]
  • "Spectral Methods for Correlated Topic Models"
       Forough Arabshahi, Animashree Anandkumar
         Appeared at the 20th International Conference on Artificial Intelligence and Statistics (AISTATS), PMLR 54:1439-1447, 2017.
    Download:     [Code]   [Poster]
  • 2016
  • "Learning Sparsely Used Overcomplete Dictionaries via Alternating Minimization"
       Alekh Agarwal, Animashree Anandkumar, Prateek Jain, and Praneeth Netrapalli
         Appeared at SIAM J. Optim., 26(4), 2016.
  • "Experimental results: Reinforcement Learning of POMDPs using Spectral Methods"
       Kamyar Azizzadenesheli, Alessandro Lazaric, Anima Anandkumar
         Appeared at NIPS 2016, Barcelona, Spain, Workshop on Deep RL
    Download:     [Poster] ;[Animation]
  • "Online and Differentially-Private Tensor Decomposition"
       Yining Wang, Anima Anandkumar
         Appeared at NIPS 2016, Barcelona, Spain
  • "Tensor Contractions with Extended BLAS Kernels on CPU and GPU"
       Yang Shi, U.N. Niranjan, Anima Anandkumar, Cris Cecka
         Appeared at HiPC2016, Hyderabad, India, December 2016
    Download:     [Poster]   [Blog Post]   [Slides]
  • "PhD Thesis: Stochastic Optimization in High Dimension"
       Hanie Sedghi, 2016
  • "PhD Thesis: Discovery of Latent Factors in High-dimensional Data Using Tensor Methods"
       Furong Huang, 2016
  • "PhD Thesis: Non-convex Optimization in Machine Learning: Provable Guarantees Using Tensor Methods"
       Majid Janzamin, 2016
  • "Open Problem: Approximate Planning of POMDPs in the class of Memoryless Policies"
       Kamyar Azizzadenesheli, Alessandro Lazaric, Anima Anandkumar
         Appeared at COLT 2016, New York, USA
    Download:     [Talk]  [Slides]
  • "Unsupervised Learning of Word-Sequence Representations from Scratch via Convolutional Tensor Decomposition"
       Furong Huang, A. Anandkumar
         Appeared at ACL 2016
    Download:     [slides]
  • "Training Input-Output Recurrent Neural Networks through Spectral Methods"
       H. Sedghi, A. Anandkumar
  • "Reinforcement Learning of POMDPs using Spectral Methods"
       Kamyar Azizzadenesheli, Alessandro Lazaric, Anima Anandkumar
         Appeared at COLT 2016, New York, USA
    Download:     [Talk]  [Slides]
  • "Efficient approaches for escaping higher order saddle points in non-convex optimization"
       Anima Anandkumar, Rong Ge
         Appeared at COLT 2016, New York, USA
    Download:     [Blog Post]
  • "Provable Tensor Methods for Learning Mixtures of Generalized Linear Models"
       H. Sedghi, M. Janzamin, A. Anandkumar
         Appeared at AISTATS 2016
  • "Tensor vs Matrix Methods: Robust Tensor Decomposition under Block Sparse Perturbations"
       Anima Anandkumar, Prateek Jain, Yang Shi, U.N. Niranjan
         Appeared at AISTATS 2016
    Download:     [Poster]  [Slides]
  • 2015
  • "Feast at Play: Feature ExtrAction using Score function Tensors"
       Majid Janzamin, Hanie Sedghi, U.N. Niranjan, Anima Anandkumar
         Appeared at NIPS 2015
  • "Convolutional Dictionary Learning through Tensor Factorization"
       Furong Huang, Anima Anandkumar
         Appeared at NIPS 2015
    Download:    [Slides]  [Code]
  • "Scalable Latent Tree Model and its Application to Health Analytics"
       F. Huang, U. N. Niranjan, J. Perros, R. Chen, J. Sun, A. Anandkumar
         Appeared at NIPS 2015
    Download:    [Slides]
  • "Discovering Neuronal Cell Types and Their Gene Expression Profiles Using a Spatial Point Process Mixture Model"
       F. Huang, A. Anandkumar, C. Borgs, J. Chayes, E. Fraenkel, M. Hawrylycz, E. Lein, A. Ingrosso, S. Turaga
         Appeared at NIPS 2015
  • "Are You Going to the Party: Depends, Who Else is Coming?: [Learning Hidden Group Dynamics via Conditional Latent Tree Models]"
       Forough Arabshahi, Furong Huang, Anima Anandkumar, Carter T. Butts and Sean M. Fitzhugh
         Appeared at International Conference on Data Mining (ICDM), IEEE 2015
    Download:    [Code]
  • "Multi-Object Classification and Unsupervised Scene Understanding Using Deep Learning Features and Latent Tree Probabilistic Models"
       Tejaswi Nimmagadda, Anima Anandkumar
         Appeared at SUNw 2015
    Download:    [Project Page]
  • "Beating the Perils of Non-Convexity: Guaranteed Training of Neural Networks using Tensor Method"
       Majid Janzamin, Hanie Sedghi, Anima Anandkumar
    Download:    [Slides]  [Talk]
  • "Fast and Guaranteed Tensor Decomposition via Sketching"
       Yining Wang, Hsiao-Yu Tung, Alex Smola, Anima Anandkuma
         Appeared at NIPS 2015
    Download:    [Slides]
  • "A Scale Mixture Perspective of Multiplicative Noise in Neural Networks"
       Eric Nalisnick, Anima Anandkumar, Padhraic Smyth
  • "Learning Mixed Membership Community Models in Social Tagging Networks through Tensor Methods"
       A. Anandkumar, H. Sedghi
  • 2014
  • "Online Tensor Methods for Learning Latent Variable Models"
       F. Huang, U. N. Niranjan, M. U. Hakeem, A. Anandkumar
         Appeared at JMLR 2014
  • "Score Function Features for Discriminative Learning: Matrix and Tensor Framework"
       M. Janzamin, H. Sedghi, A. Anandkumar
         Appeared at JMLR 2014
  • "Provable Methods for Training Neural Networks with Sparse Connectivity"
       H. Sedghi, A. Anandkumar
         Appeared at NIPS Deep Learning Workshop. Dec. 2014
  • "PAnalyzing Tensor Power Method Dynamics in Overcomplete Regime"
       A. Anandkumar, R. Ge, M. Janzamin
  • "Non-convex Robust PCA"
       P. Netrapalli, U. N. Niranjan, S. Sanghavi, A. Anandkumar, P. Jain.Â
         An abridged version appears in NIPS 2014
    Download:    [Slides]
  • "Sample Complexity Analysis for Learning Overcomplete Latent Variable Models through Tensor Methods"
       A. Anandkumar, R. Ge, M. Janzamin
         An abridge version appears in COLT 2015
    Download:    [Abridged]  [Slides]
  • "Multi-Step Stochastic ADMM in High Dimensions: Applications in Sparse Optimization and Noisy Matrix Decomposition"
       H. Sedghi, A. Anandkumar, E. Jonckheere
         Appeared at NIPS 2014
    Download:    [Talk]
  • "Guaranteed Non-Orthogonal Tensor Decomposition via Alternating Rank-1 Updates"
       A. Anandkumar, R. Ge, M. Janzamin
         Appeared at NIPS 2014
    Download:    [Slides]  [Code]
  • "Nonparametric Estimation of Multi-View Latent Variable Models"
       L. Song, A. Anandkumar, B. Dai, B. Xie
         Appeared at ICML 2014
  • "Tensor Decompositions for Learning Latent Variable Models"
       A. Anandkumar, R. Ge, D. Hsu, S.M. Kakade and M. Telgarsky.Â
         Appeared at Journal of Machine Learning Research 2014
    Download:    [Blog Post]
  • 2013
  • Learning Loopy Graphical Models with Latent Variables: Efficient Methods and Guarantees, Ann. Statist., 41(2), 2013. W3Schools supplement slides code
  • Learning Sparsely Used Overcomplete Dictionaries via Alternating Minimization, 2013. W3Schools slides
  • Exact Recovery of Sparsely Used Overcomplete Dictionaries, 2013. W3Schools
  • When Are Overcomplete Topic Models Identifiable? Uniqueness of Tensor Tucker Decompositions with Structured Sparsity, 2013. W3Schools
  • A Tensor Spectral Approach to Learning Mixed Membership Community Models, COLT 2013. W3Schools slides
  • Learning Latent Bayesian Networks and Topic Models Under Expansion Constraints, ICML 2013. W3Schools
  • Fast, Concurrent and Distributed Load Balancing under Switching Costs and Imperfect Observations, IEEE INFOCOM 2013. W3Schools
  • 2012
  • High-Dimensional Structure Learning of Ising Models: Local Separation Criterion, Ann. Statist. 40(3), 2012. W3Schools supplement code slides
  • Learning Linear Bayesian Networks with Latent Variables, 2012. W3Schools
  • Feedback Message Passing for Inference in Gaussian Graphical Models, IEEE Trans. on Signal Processing, 60(8), 2012. W3Schools
  • High-Dimensional Covariance Decomposition into Sparse Markov and Independence Domains, ICML 2012. W3Schools W3Schools slides
  • A Method of Moments for Mixture Models and Hidden Markov Models, COLT 2012. W3Schools W3Schools
  • Learning High-Dimensional Mixtures of Graphical Models, NIPS 2012. W3Schools W3Schools
  • Two SVDs Suffice: Spectral decompositions for probabilistic topic modeling and latent Dirichlet allocation, NIPS 2012. W3Schools W3Schools
  • 2011
  • High-Dimensional Gaussian Graphical Model Selection: Walk-Summability and Local Separation Criterion, JMLR 13:229-337, 2012. NIPS 2011. W3Schools W3Schools talk slides
  • High-Dimensional Graphical Model Selection: Tractable Graph Families and Necessary Conditions, NIPS 2011. W3Schools
  • Spectral Methods for Learning Multivariate Latent Tree Structure, NIPS 2011. W3Schools
  • Robust Rate Maximization Game Under Bounded Channel Uncertainty, IEEE Trans. Vehicular Technology, 60:9, 2011. W3Schools
  • Summary Based Structures with Improved Sublinear Recovery for Compressed Sensing, IEEE ISIT 2011. W3Schools
  • Topology Discovery of Sparse Random Graphs With Few Participants, ACM SIGMETRICS 2011. W3Schools W3Schools slides
  • Learning Latent Tree Graphical Models, JMLR 2011. W3Schools project homepage Allerton version Allerton slides seminar slides
  • Learning High-Dimensional Markov Forest Distributions: Analysis of Error Rates, JMLR 2011. W3Schools Allerton version Allerton slides
  • Distributed Algorithms for Learning and Cognitive Medium Access with Logarithmic Regret, IEEE JSAC 2011. W3Schools
  • Energy-Latency Tradeoff for In-Network Function Computation in Random Networks, IEEE INFOCOM 2011. W3Schools slides
  • Index-Based Sampling Policies for Tracking Dynamic Networks under Sampling Constraints, IEEE INFOCOM 2011. W3Schools supplemental material
  • A Large-Deviation Analysis of the Maximum-Likelihood Learning of Markov Tree Structures, IEEE Trans. Information Theory,57:3, 2011. W3Schools
  • 2010
  • Scaling Laws for Random Spatial Graphical Models, IEEE ISIT 2010. W3Schools slides
  • Error Exponents for Composite Hypothesis Testing of Markov Forest Distributions, IEEE ISIT 2010. W3Schools slides proofs
  • Learning Gaussian Tree Models: Analysis of Error Exponents and Extremal Structures, IEEE Trans. Signal Processing, 58:5, 2010. W3Schools slides
  • Robust Rate Maximization Game Under Bounded Channel Uncertainty, IEEE ICASSP 2010. W3Schools
  • Opportunistic Spectrum Access with Multiple Users: Learning under Competition, IEEE INFOCOM 2010. W3Schools slides
  • Seeing Through Black Boxes : Tracking Transactions through Queues under Monitoring Resource Constraints, Elsevier Performance Evaluation 2010. W3Schools
  • 2009
  • Energy Scaling Laws for Distributed Inference in Random Fusion Networks, IEEE JSAC, 27:7, 2009. W3Schools
  • Selectively Retrofitting Monitoring in Distributed Systems, Workshop on MAMA 2009. W3Schools
  • Detection Error Exponent for Spatially Dependent Samples in Random Networks, IEEE ISIT 2009. W3Schools
  • Prize-Collecting Data Fusion for Cost-Performance Tradeoff in Distributed Inference, IEEE INFOCOM 2009. W3Schools tech report
  • Detection of Gauss-Markov random fields with nearest-neighbor dependency, IEEE Trans. Information Theory, 55:2, 2009. W3Schools
  • 2008 and Earlier
  • Optimal Node Density for Detection in Energy Constrained Random Networks, IEEE Trans. Signal Processing, 56:10, 2008. W3Schools
  • Distributed Estimation Via Random Access, IEEE Trans. Information Theory, 54:7, 2008. W3Schools
  • Tracking in a Spaghetti Bowl: Monitoring Transactions Using Footprints, ACM SIGMETRICS 2008. W3Schools
  • Minimum Cost Data Aggregation with Localized Processing for Statistical Inference, IEEE INFOCOM 2008. W3Schools
  • A Large Deviation Analysis of Detection over Multi-Access Channels with Random Number of Sensors, ICASSP 2006 (Best Paper Award). W3Schools