Research
Current Research Interests
- Nonlinear Optimization in Large-scale Machine Learning Problems: Stochastic Gradient Methods (SGDs), Coordinate Descent Methods (CDs), etc.
- Statistical Learning and Deep Learning: Neural Networks, Gaussian Process Modeling (GP), Deep GP.
- Parallel Computing: Mini-Batches in SGDs and CDs.
- Large-Scale Optimization in Power Systems: Optimization Methods in Alternating-Current Optimal Power Flows.
Publications
- Hybrid Methods in Solving Alternating-Current Optimal Power Flows. Jie Liu, Jakub Marecek and Martin Takac. arXiv:1510.02171
- Mini-Batch Semi-Stochastic Gradient Descent in the Proximal Setting. Jakub Konecny, Jie Liu, Peter Richtarik, Martin Takac. IEEE Journal of Selected Topics in Signal Processing, 10(2): 242-255.
- mS2GD: Mini-Batch Semi-Stochastic Gradient Descent in the Proximal Setting. Jakub Konecny, Jie Liu, Peter Richtarik, Martin Takac. NIPS Workshop on Optimization for Machine Learning, 2014.
Talks
Invited Talks
- Jun 30, 2016 Combining First and Second Order Methods in Solving Alternating- Current Power Flows, IBM Research, Dublin, Ireland.
- Jul 21, 2015 Modeling and Optimization: Theory and Applications (MOPTA 2015), Lehigh University, Bethlehem, PA, USA.
- Jul 13, 2015 22nd International Symposium on Mathematical Programming (ISMP 2015), Pittsburgh, PA, USA.
- Jul 6, 2015 Predictive Analyses for Time Series with Gaussian Processes, Siemens Corporate Research, Princeton, NJ, USA.
- Aug 2, 2012 Symposium Features Talks by LANS Summer Students of 2012, Argonne National Laboratory, Lemont, IL, USA.
Contributed Talks
- Mar 4, 2016 10th Annual Machine Learning Symposium, Poster Session, The New York Academy of Sciences, New York, NY, USA.
- Mar 13, 2015 9th Annual Machine Learning Symposium, Poster Session, The New York Academy of Sciences, New York, NY, USA.
- Dec 12, 2014 NIPS 2014 Workshop: Optimization for Machine Learning, Poster Session, Montréal, Québec, Canada.
- Mar 31, 2012 Applied Math Days, Rensselaer Polytechnic Institute, Troy, NY, USA.
