Mononito Goswami

I am a Robotics Ph.D. Student at the Auton Lab in School of Computer Science at Carnegie Mellon University, advised by Prof. Artur Dubrawski.

I am interested in machine learning using imperfect data with an emphasis on healthcare and time-series modeling. Specifically, my research focusses on settings which challenge common assumptions on data, i.e., data is (1) labeled accurately, (2) centrally located, (3) sufficient to train complex models from scratch. I am also interested in evaluating trained machine learning models comprehensively in these practical settings, and foundation modeling for temporal data and clinical applications.

I was an Applied Scientist Intern at Amazon Web Services (AWS) AI Labs in the summer of 2022 and 2023! I’ll be interning at Google Cloud AI this summer!

Prior to pursuing my Ph.D., I earned my bachelor’s in computer engineering from Delhi Technological University (erstwhile Delhi College of Engineering) in India.

Feel free to drop me an email at mgoswami [at] andrew [dot] cmu [dot] edu if you are interested in my research and would like to collaborate!

I am a member of the Robotics Institute Climate Committee, so if you’re an RI student facing issues, let’s chat!

My CV can be found here.

Time-series Foundation Models

I worked on MOMENT, the first family of open-source large pre-trained time-series models! MOMENT serves as a building block for multiple time-series analysis tasks (forecasting, classification, anomaly detection, and imputation, etc.), is effective out-of-the-box, i.e. with no (or few) particular task-specific exemplars (enabling zero-shot forecasting, etc.), and (3) is tunable using in-distribution and task-specific data to improve performance! Check out our preprint and code on anomymous github.

We are currently working on open-sourcing MOMENT, which includes EVERYTHING from the code to compile the Time-series Pile, pre-train, and fine-tune MOMENT, to the pre-trained model weights! If you’re interested in our model, or have any questions or feature requests, please fill this Google form, and we’ll do our best to answer your questions, and update you as we open-source MOMENT and its subsequent improvements.

Awards

Center for Machine Learning and Health (CMLH) Fellowship 2021

Publications

JoLT: Jointly Learned Representations of Language and Time-Series.

Yifu Cai, Mononito Goswami, Arjun Choudhry, Arvind Srinivasan, Artur Dubrawski. Neural Information Processing Systems Workshop on Deep Generative Models for Health (DGM4H NeurIPS) (2023) (Poster).

PICSR: Prototype-Informed Cross-Silo Router for Federated Learning.

Eric Enouen, Sebastian Caldas, Mononito Goswami, Artur Dubrawski. *AAAI Conference on Artificial Intelligence (Student Abstract) (2024). 3-min presentation contest finalist**

JoLT: Jointly Learned Representations of Language and Time-Series for Clinical Time-series Interpretation.

Yifu Cai, Arvind Srinivasan, Mononito Goswami, Arjun Choudhry, Artur Dubrawski. AAAI Conference on Artificial Intelligence (Student Abstract) (2024). Best student abstract presentation award winner.

AQuA: A Benchmarking Tool for Label Quality Assessment

Mononito Goswami, Vedant Sanil, Arjun Choudhry, Arvind Srinivasan, Chalisa Udompanyawit, Artur Dubrawski
Neural Information Processing Systems (NeurIPS) 2023 Datasets & Benchmark Track (Poster). [ArXiv, code]

Unsupervised Model Selection for Time-series Anomaly Detection

Mononito Goswami, Cristian Challu, Laurent Callot, Lenon Minorics, Andrey Kan
International Conference of Learning Representations (ICLR) 2023 (Spotlight).
[ArXiv, code]

Encoding Expert Knowledge into Federated Learning Using Weak Supervision

Sebastian Caldas, Mononito Goswami, Artur Dubrawski
International Conference of Learning Representations 2023 Workshop on Machine Learning for Internet of Things (ML4IoT).

Classifying Unstructured Clinical Notes via Automatic Weak Supervision

Chufan Gao*, Mononito Goswami*, Jieshi Chen, Artur Dubrawski
Machine Learning for Healthcare Conference 2022.
[ArXiv, code]

Weakly Supervised Classification of Vital Sign Alerts as Real or Artifact

Arnab Dey, Mononito Goswami, Joo Heung Yoon, Gilles Clermont, Michael R. Pinsky, Marilyn Hravnak, Artur Dubrawski. AMIA 2022 Annual Symposium. American Medical Informatics Association.
[ArXiv, code]

Counterfactual Phenotyping with Censored Time-to-Events

Chirag Nagpal, Mononito Goswami, Keith Dufendach, Artur Dubrawski
ACM Conference on Knowledge Discovery and Data Mining 2022.
[ArXiv, code]

Weak Supervision for Affordable Modeling of Electrocardiogram Data

Mononito Goswami, Benedikt Boecking, Artur Dubrawski
AMIA 2021 Annual Symposium. American Medical Informatics Association.
[ArXiv]

Toward Learning at Scale in Developing Countries: Lessons from the Global Learning XPRIZE Field Study

Andrew A, McReynolds, Sheba P Naderzad, Mononito Goswami, Jack Mostow
ACM Conference on Learning@Scale 2020.
[ACM]

A binary PSO approach for improving the performance of wireless sensor networks

Ajay Kaushik*, Mononito Goswami*, Minkush Manuja*, S Indu, Daya Gupta
Wireless Personal Communications 113 (1), 263-297
[Springer]

A multi-task approach to open domain suggestion mining using language model for text over-sampling

Maitree Leekha*, Mononito Goswami*, Minni Jain
European Conference on Information Retrieval (ECIR) 2020
[Springer]

Discriminating Cognitive Disequilibrium and Flow in Problem Solving: A Semi-Supervised Approach Using Involuntary Dynamic Behavioral Signals

Mononito Goswami*, Lujie Chen*, Artur Dubrawski
AAAI Conference on Artificial Intelligence (Social Impact Track) 2020
[AAAI]

Modeling Involuntary Dynamic Behaviors to Support Intelligent Tutoring (Student Abstract)

Mononito Goswami, Lujie Chen, Chufan Gao, Artur Dubrawski
AAAI Conference on Artificial Intelligence (Student Abstract) 2020
[AAAI]

A multi-task approach to open domain suggestion mining (Student abstract)

M Jain*, Maitree Leekha*, Mononito Goswami*
AAAI Conference on Artificial Intelligence (Student Abstract) 2020
[AAAI]

Detecting intrusive transactions in databases using partially-ordered sequential rule mining and fractional-distance based anomaly detection

Indu Singh, Minkush Manuja*, Rishabh Mathur*, Mononito Goswami*
International Journal of Intelligent Engineering Informatics 8 (2), 138-171
[Inderscience]

What makes a better companion? Towards Social & Engaging Peer Learning

Rajni Jindal, Maitree Leekha*, Minkush Manuja*, Mononito Goswami*
European Conference on Artificial Intelligence (ECAI) 2020
[IOS Press]

Detecting Patterns of Physiological Response to Hemodynamic Stress via Unsupervised Deep Learning

Chufan Gao, Fabian Falck, Mononito Goswami, Anthony Wertz, Michael R. Pinsky, Artur Dubrawski
NeurIPS Machine Learning for Health (ML4H) Workshop 2020 (Short Paper)
[ArXiv]

Are you paying attention? detecting distracted driving in real-time

Maitree Leekha*, Mononito Goswami*, Rajiv Ratn Shah, Yifang Yin, Roger Zimmermann
IEEE Fifth International Conference on Multimedia Big Data (BigMM) 2019
[IEEE]

What’s Most Broken? A Tool to Assist Data-Driven Iterative Improvement of an Intelligent Tutoring System

Mononito Goswami*, Shiven Mian*, Jack Mostow
AAAI Conference on Artificial Intelligence (Student Abstract) 2019
[AAAI]

What’s Most Broken? Design and Evaluation of a Tool to Guide Improvement of an Intelligent Tutor

Shiven Mian*, Mononito Goswami*, Jack Mostow
International Conference on Artificial Intelligence in Education (AIED) 2019
[Springer]

* indicates equal contribution

Preprints

MOMENT: A Family Of Open Time-series Foundation Models

Mononito Goswami, Konrad Szafer, Arjun Choudhry, Yifu Cai, Shuo Li, Artur Dubrawski [ArXiv, code]

Encoding Expert Knowledge into Federated Learning Using Weak Supervision

Sebastian Caldas, Mononito Goswami, Artur Dubrawski

Towards Knowledge-driven Clinical Data Annotation Quality Assessment

Mononito Goswami, Benedikt Boecking, Patrick J. Coppler, Jonathan Elmer, Artur Dubrawski

Active Refinement of Weakly Supervised Models

Mononito Goswami*, Chufan Gao*, Benedikt Boecking, Saswati Ray, Artur Dubrawski

Towards Social & Engaging Peer Learning: Predicting Backchanneling and Disengagement in Children

Mononito Goswami*, Minkush Manuja*, Maitree Leekha*
[ArXiv]