Francis Tom

Master's student, Language Technologies Institute
School of Computer Science, Carnegie Mellon University

I work with Prof. Jamie Callan. I am currently interested in information retrieval for LLM agents: how search engines should serve agents, how agents can treat memory as retrieval over their own experience, and how they learn to search through reinforcement learning.

Email: francist [at] andrew.cmu.edu · Google Scholar

Background

I have worked as a Data and Applied Scientist at Microsoft Bing on multimodal question answering, e-commerce search, and query autosuggestion. I took models from research to production at web scale, under tight latency budgets.

I have also worked on deep generative models for medical imaging and on audio classification, including work at IBM Research. I graduated from IIT Kharagpur in Electrical Engineering, with a minor in Computer Science.

Publications

• RespireNet: A Deep Neural Network for Accurately Detecting Abnormal Lung Sounds in Limited Data Setting
  S. Gairola, F. Tom, N. Kwatra, M. Jain. IEEE EMBC 2021.
• Visually Precise Query
  R. Dasgupta, F. Tom, S. Kumar, M. Das Gupta, Y. Kumar, B. N. Patro, V. P. Namboodiri. ACM Multimedia 2020.
• Learning a Deep Convolution Network with Turing Test Adversaries for Microscopy Image Super Resolution
  F. Tom, H. Sharma, D. Mundhra, T. R. Dastidar, D. Sheet. IEEE ISBI 2019.
• Ultracompression: Framework for High Density Compression of Ultrasound Volumes using Physics Modeling Deep Neural Networks
  D. China*, F. Tom*, S. Nandamuri*, A. Kar, M. Srinivasan, P. Mitra, D. Sheet. IEEE ISBI 2019.
• End-to-End Audio Replay Attack Detection using Deep Convolutional Networks with Attention
  F. Tom, M. Jain, P. Dey. Interspeech 2018.
• Simulating Patho-realistic Ultrasound Images using Deep Generative Networks with Adversarial Learning
  F. Tom, D. Sheet. IEEE ISBI 2018.

* Equal contribution.