We’re interested in machine learning, optimization algorithms and text understanding, as well as several application domains.
Stochastic Zeroth-Order Optimisation Algorithms with Variance ReductionEPFL, 2019-06-21.
A comparison of model-parallel training methods for deep learning2019-06-18.
Decentralized Stochastic Optimization and Gossip Algorithms with Compressed Communication2019-06-09. ICML 2019 – International Conference on Machine Learning, Long Beach, California, USA, 9-15 June 2019.
Local SGD Converges Fast and Communicates Little2019-05-06. ICLR 2019 – International Conference on Learning Representations, New Orleans, USA, May 6-9, 2019.
Error Feedback Fixes SignSGD and other Gradient Compression Schemes2019. 36th International Conference on Machine Learning (ICML) 2019, Long Beach, USA, June 9-15, 2019. p. 3252-3261.
Overcoming Multi-model Forgetting2019. ICML 2019 – 36th International Conference on Machine Learning, Long Beach, California, USA, June 09-15, 2019. p. 594-603.
Open-Vocabulary Keyword Spotting with Audio and Text Embeddings2019. INTERSPEECH 2019 – IEEE International Conference on Acoustics, Speech, and Signal Processing, Graz, Austria,
Better Word Embeddings by Disentangling Contextual n-Gram Information2019. NAACL 2019 – Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Minneapolis, Minnesota, USA, June 2-7. 2019.
Training DNNs with Hybrid Block Floating Point2018-12-04. NeurIPS 2018 – Neural Information Processing Systems, Montréal Canada, December 2-8, 2018.
Accelerated Stochastic Matrix Inversion: General Theory and Speeding up BFGS Rules for Faster Second-Order Optimization2018-12-02. 32nd Annual Conference on Neural Information Processing Systems (NIPS), Montréal, Canada, December 2-8, 2018.
Sparsified SGD with Memory2018-12-02. NeurIPS 2018 – 32nd Annual Conference on Neural Information Processing Systems, Montréal, Canada, December 2-8, 2018.
Unsupervised Learning of Representations for Lexical Entailment Detection2018-09-04.
Convex Optimization using Sparsified Stochastic Gradient Descent with Memory2018-06-27.
Unsupervised learning of sentence embeddings using compositional n-gram features2018-05-01. NAACL 2018 – Conference of the North American Chapter of the Association for Computational Linguistics.
Simple Unsupervised Keyphrase Extraction using Sentence Embeddings2018-05-01. CoNLL 2018 – SIGNLL Conference on Computational Natural Language Learning.
On Matching Pursuit and Coordinate Descent2018-05-01. ICML 2018 – 35th International Conference on Machine Learning.
Adaptive balancing of gradient and update computation times using global geometry and approximate subproblems2018-05-01.
Learning Word Vectors for 157 Languages2018-02-19. Language Resources and Evaluation Conference, Miyazaki, Japan, May 7-12, 2018.
A Distributed Second-Order Algorithm You Can Trust2018. 35th International Conference on Machine Learning (ICML 2018), Stockholm, Sweden, 10-15 July 2018. p. 1358-1366.
Training DNNs with Hybrid Block Floating Point2018-01-01. NeurIPS 2018 – 32nd Conference on Neural Information Processing Systems, Montreal, CANADA, Dec 02-08, 2018.
CoCoA: A General Framework for Communication-Efficient Distributed OptimizationJournal Of Machine Learning Research. 2018-01-01. Vol. 18.
Some results on a class of mixed van der Waerden numbersRocky Mountain Journal of Mathematics. 2018. Vol. 48, num. 3, p. 885-904. DOI : 10.1216/RMJ-2018-48-3-885.
Gene locations may contribute to predicting gene regulatory relationshipsJOURNAL OF ZHEJIANG UNIVERSITY-SCIENCE B. 2018. Vol. 19, num. 1, p. 25-37. DOI : 10.1631/jzus.B1700303.
Prediction of patient-reported physical activity scores from wearable accelerometer data: a feasibility study2018. ICNR2018 – International Conference on NeuroRehabilitation, Pisa, Italy, October 16 – 20, 2018.
Asynchronous updates for stochastic gradient descent2017-06-09.
Fully Quantized Distributed Gradient Descent2017.
Safe Adaptive Importance Sampling2017. Neural Information Processing Systems (NIPS), Long Beach, USA, December 4-9, 2017.
Approximate Steepest Coordinate Descent2017. ICML 2017 – International Conference on Machine Learning, Sydney, Australia, Aug 6-11, 2017.
Unsupervised Learning of Sentence Embeddings using Compositional n-Gram Features
A Unified Optimization View on Generalized Matching Pursuit and Frank-Wolfe2017. 20th International Conference on Artificial Intelligence and Statistics (AISTATS) 2017, Fort Lauderdale, California, United States, April 20-22, 2017.
Faster Coordinate Descent via Adaptive Importance Sampling2017. 20th International Conference on Artificial Intelligence and Statistics (AISTATS) 2017, Fort Lauderdale, Florida, USA, April 20-22, 2017.
CoCoA: A General Framework for Communication-Efficient Distributed Optimization2016.
Screening Rules for Convex Problems
Pursuits in Structured Non-Convex Matrix Factorizations
Primal-Dual Rates and Certificates2016. ICML 2016 – International Conference on Machine Learning, USA, NY, New York City, June 19-24, 2016. p. 783-792.
Audio Based Bird Species Identification using Deep Learning Techniques2016. Conference and Labs of the Evaluation Forum (CLEF) 2016, Évora, Portugal, 5-8 September, 2016. p. 547-559.
L1-Regularized Distributed Optimization: A Communication-Efficient Primal-Dual Framework
On the Global Linear Convergence of Frank-Wolfe Optimization Variants2015. Neural Information Processing Systems (NIPS) 2015, Montreal, Quebec, Canada, December 7-12, 2015. p. 496-504.
Adding vs. Averaging in Distributed Primal-Dual Optimization2015. ICML 2015 – International Conference on Machine Learning, Lille, France, 6-11 July 2015. p. 1973-1982.
An Equivalence between the Lasso and Support Vector MachinesRegularization, Optimization, Kernels, and Support Vector Machines; Chapman and Hall/CRC, 2014. p. 1-26.
Communication-Efficient Distributed Dual Coordinate Ascent.2014. Neural Information Processing Systems (NIPS) 2014, Montreal, Quebec, Canada, December 8-13 2014. p. 3068-3076.
Revisiting Frank-Wolfe: Projection-Free Sparse Convex Optimization2013. ICML 2013 – International Conference on Machine Learning, Atlanta, GA, USA, 16-21 June 2013. p. 427-435.
Block-Coordinate Frank-Wolfe Optimization for Structural SVMs2013. ICML 2013 – International Conference on Machine Learning, Atlanta, GA, USA, 16-21 June 2013. p. 53-61.
Regularization Paths with Guarantees for Convex Semidefinite Optimization.2012. International Conference on Artificial Intelligence and Statistics (AISTATS) 2012, La Palma, Canary Islands, April 21-23, 2012. p. 432-439.
Optimizing over the Growing Spectrahedron2012. European Symposia on Algorithms (ESA) 2012, Ljubljana, Slovenia, September 10-12, 2012. p. 503-514. DOI : 10.1007/978-3-642-33090-2_44.
Approximating parameterized convex optimization problemsACM Trans. Algorithms. 2012. Vol. 9, num. 1, p. 10:1-10:17. DOI : 10.1145/2390176.2390186.
Sparse convex optimization methods for machine learningETH Zürich, 2011.
Convex Optimization without Projection Steps
A Simple Algorithm for Nuclear Norm Regularized Problems.2010. International Conference on Machine Learning (ICML) 2010, Haifa, Israel, June 21-24, 2010. p. 471-478.
Approximating Parameterized Convex Optimization Problems.2010. European Symposia on Algorithms (ESA) 2010, Liverpool, UK, September 6-8, 2010. p. 524-535. DOI : 10.1007/978-3-642-15775-2_45.
Approximating Parameterized Convex Optimization ProblemsEuropean Symposia on Algorithms (ESA) 2010; Springer Berlin Heidelberg, 2010. p. 524-535.
A Combinatorial Algorithm to Compute Regularization Paths