Ph.D. Candidate Shashanka Ubaru Takes Home the ICMLA’s Best Paper Award
Ph.D. candidate Shashanka Ubaru received the best paper award at the 16th IEEE International Conference on Machine Learning Applications (ICMLA) for his paper, “UoI-NMF_cluster: A Robust Nonnegative Matrix Factorization Algorithm for Improved Parts-Based Decomposition and Reconstruction of Noisy Data.”
Ubaru was joined by co-authors Keshen Wu and Krostofer E. Bouchard, both from Lawrence Berkeley National Labs, in receiving the award. Their paper presents a novel algorithm to confront an ongoing problem with noisy data. As data collection becomes more prevalent and the volumes of data increase on an exponential scale, important insights are potentially being overlooked because there is too much information to sift through, even with the latest machine learning tools.
Ubaru’s team works with a family of algorithms called nonnegative matrix factorization (NMF) that have been helpful in sifting through all this data, but only up to a certain point. Where prior NMF algorithms have shown to have drawbacks such as making certain assumptions on the type of input data or failing to give interpretable results altogether, Ubaru and his team have created an NMF method of analyzing data that pulls interpretable and predictive structures from complex and noisy data.
Advised by Professor Yousef Saad, Ubaru’s ongoing research interest are in machine learning, numerical linear algebra, and coding theory applications. In particular, he uses computational linear algebra tools and error correcting coding theory to solve problems related to machine learning, data analysis and signal processing. He also works on the use of machine learning for scientific data applications, particularly in neuroscience and material science.
ICMLA was held in Cancun, Mexico and brings together researcher and practitioners to present their latest achievements and innovations in the area of machine learning.
Please join CS&E in sending our congratulations to Shashanka Ubaru.