Draft:Tommi Jaakola
 | This article has multiple issues. Please help improve it or discuss these issues on the talk page. (Learn how and when to remove these template messages)
|
Tommi Jaakkola is a computer science professor at the Massachusetts Institute of Technology, serving as a Principal Investigator at MIT's Computer Science and Artificial Intelligence Laboratory. His work focuses on machine learning and natural language processing, specializing in topics such as inference, semi-supervised learning, information retrieval, and reinforcement learning. In 2017 he was elected a fellow of the Association for the Advancement of Artificial Intelligence in recognition of “significant contributions to the fields of machine learning, computational biology and natural language processing”.[1]
Career[edit]
Jaakkola graduated from Helsinki University of Technology in 1992 and received his Ph.D. from MIT in computational neuroscience in 1997. Following a postdoctoral position in computational molecular biology he joined the MIT EECS faculty in 1998.
Jaakkola's research aims to tackle inferential, algorithmic and estimation questions in machine learning, including large scale probabilistic distributed inference, deep learning, and causal inference. The applied side of his work has involved problems in natural language processing such as parsing, regulatory models in computational biology, computational chemistry, and computational functional genomics.
He has done extensive research on unsupervised systems that can be passively trained to learn language more like humans, so that they can work using just audio. One of his translation models[2] runs more efficiently than traditional “monolingual” translation models by viewing words as vectors that can be clustered with other words that have similar meanings.
Another major topic area is explainability and better understanding how machine learning systems produce outputs. With MIT professor Regina Barzilay he has demonstrated methods for better interpreting the rationale behind the decisions of language models.[3]
Jaakkola has also done work related to drug development, producing machine learning models that can jointly learn drug-target interaction and drug-drug synergy to uncover new drug combinations to better protect against coronaviruses.[4]
His papers have received more than 45,000 citations, with an h-index of 94.[5] In 2022 he received the AISTATS Test of Time Award for co-authoring the paper ”Approximate Inference in Additive Factorial HMMs with Application to Energy Disaggregation”.[6] He has also held editorial positions on journals such as the Journal of Machine Learning Research.[7]
References[edit]
- ^ "Elected AAAI Fellows". mit.edu.
- ^ Alvarez-Melis, David; Jaakkola, Tommi S. (2018). "Gromov-Wasserstein Alignment of Word Embedding Spaces". arXiv:1809.00013 [cs.CL].
- ^ Lei, Tao; Barzilay, Regina; Jaakkola, Tommi (2016). "Rationalizing Neural Predictions". arXiv:1606.04155 [cs.CL].
- ^ Jin, Wengong; Stokes, Jonathan M.; Eastman, Richard T.; Itkin, Zina; Zakharov, Alexey V.; Collins, James J.; Jaakkola, Tommi S.; Barzilay, Regina (2021-09-28). "Deep learning identifies synergistic drug combinations for treating COVID-19". Proceedings of the National Academy of Sciences. 118 (39). Bibcode:2021PNAS..11805070J. doi:10.1073/pnas.2105070118. ISSN 0027-8424. PMC 8488647. PMID 34526388.
- ^ "Tommi Jaakola". scholar.google.com.
- ^ Kolter, J. Zico; Jaakkola, Tommi (2012-03-21). "Approximate Inference in Additive Factorial HMMs with Application to Energy Disaggregation". Proceedings of the Fifteenth International Conference on Artificial Intelligence and Statistics. PMLR: 1472–1482.
- ^ "JMLR Editorial Board". jmlr.org.