Jwala Dhamala

Applied Scientist at Amazon Alexa AI

I am a scientist at Amazon Alexa AI (NLU), CA. I am interested in Aritifical Intelligence research, particularly in the field of Natural Language Processing. Presently I work on improving fairness, robustness and interpretability of NLP models.

Previously, I received my PhD degree in Computing and Information Sciences from the Rochester Institute of Technology (RIT) where I was working with Dr. Linwei Wang in the Computational Biomedicine Lab. During my PhD, I worked on personalization and uncertainty qualtification in multi-scale 3D simulation models of cardiac electrophysiology which gave me an opportunity to work at an intersection of machine learning (Bayesian models and optimization, generative models, graph convolutional models) and healthcare problem (cardiac model personalization).



News

  • [May, 2022] NEW! Three papers accepted to ACL 2022 [paper 1] [paper 2] [paper 3]
  • [May, 2022] NEW! We are organizing a workshop TrustNLP at NAACL 2022. Please consider attending and contributing.
  • [Oct, 2021] I will be presenting at WeCNLP Summit. Lets connect if you are also attending the summit.
  • [July, 2021] We are organizing a workshop REsponsible AI at KDD 2021. Please consider attending and contributing.
  • [May, 2021] We are organizing a workshop TrustNLP at NAACL 2021. Please consider attending and contributing.
  • [Jan, 2021] Paper on dataset and metrics for measuring biases in open-ended language generation accepted at ACM FAccT. Dataset: https://github.com/amazon-research/bold
  • [Dec, 2020] I participated in a panel discussion on fairness in AI and ML (organized alongside NeurIPS 2020).
  • [Oct, 2020] Paper lead by our intern Ansel on evaluating the effectiveness of ENAS accepted at EMNLP workshop.
  • [Feb, 2020] Paper on model personalization via HD Bayesian optimization and generative model is accepted at Medical Image Analysis.
  • [Feb, 2020] Successfully defended my PhD thesis.
  • [Dec, 2019] I joined Amazon Alexa NU-AI in Cambridge, MA as a Research Scientist.
  • [Jun, 2019] Paper early accepted and finalist for Young Scientist Award at MICCAI 2019
  • [Oct, 2018] Paper (collaboration of Philips Research and MIT CSAIL) accepted at IEEE Sensors Letters and NeurIPS ML4H; received NeurIPS ML4H workshop travel grant
  • [Oct, 2018] Abstract accepted in WiML Workshop 2018 and received WiML 2018 travel grant
  • [Sep, 2018] MICCAI 2018 paper is finalist for young scientist award; received MICCAI 2018 travel grant

Publications

Journal Publications

  • Embedding High-dimensional Bayesian Optimization via Generative Modeling: Parameter Personalization of Cardiac Electrophysiological Models
    Jwala Dhamala, Hermenegild J. Arevalo, John L. Sapp, Milan Horacek, Katherine C. Wu, Natalia A. Trayanova, and Linwei Wang
    Medical Image Analysis (MedIA), in submission
  • Multivariate Time-series Similarity Assessment via Unsupervised Representation Learning and Stratified Locality Sensitive Hashing: Application to Early Acute Hypotensive Episode Detection
    Jwala Dhamala, Emmanuel Azuh, Abdullah Al-Dujaili, Jonathan Rubin, and Una-May O'Reilly.
    IEEE Sensors Letters, 2018; NeurIPS Machine Learning for Health Workshop, 2018
  • Quantifying the Uncertainty in Model Parameters using Gaussian Process-Based Markov Chain Monte Carlo in Cardiac Electrophysiology
    Jwala Dhamala, Hermenegild J. Arevalo, John L. Sapp, Milan Horacek, Katherine C. Wu, Natalia A. Trayanova, and Linwei Wang
    Medical Image Analysis (MedIA), 2018
  • Spatially-Adaptive Multi-Scale Optimization for Local Parameter Estimation in Cardiac Electrophysiology
    Jwala Dhamala, Hermenegild J. Arevalo, John L. Sapp, Milan Horacek, Katherine C. Wu, Natalia A. Trayanova, and Linwei Wang
    IEEE Transactions on Medical Imaging (IEEE TMI), 2017

Conference Publications

  • BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generation
    Jwala Dhamala, Tony Sun, Varun Kumar, Satyapriya Krishna, Yada Pruksachatkun, Kai-Wei Chang, Rahul Gupta
    ACM Conference on Fairness, Accountability, and Transparency (ACM FAccT), 2021
  • Bayesian Optimization on Large Graphs via a Graph Convolutional Generative Model: Application in Cardiac Model Personalization
    Jwala Dhamala, John L. Sapp, Milan Horacek, and Linwei Wang
    Medical Image Computing and Computer Assisted Intervention (MICCAI), 2019
  • High-dimensional Bayesian Optimization of Personalized Cardiac Model Parameters via an Embedded Generative Model
    Jwala Dhamala, John L. Sapp, Milan Horacek, and Linwei Wang
    Medical Image Computing and Computer Assisted Intervention (MICCAI), 2018
  • Quantifying the Uncertainty in Model Parameters using Gaussian Process-Based Markov Chain Monte Carlo: An Application to Cardiac Electrophysiological Models
    Jwala Dhamala, John L. Sapp, Milan Horacek, and Linwei Wang
    Information Processing in Medical Imaging (IPMI), 2017
  • Spatially-Adaptive Multi-scale Optimization for Local Parameter Estimation: Application in Cardiac Electrophysiological Models
    Jwala Dhamala, John L. Sapp, Milan Horacek, and Linwei Wang
    Medical Image Computing and Computer Assisted Intervention (MICCAI), 2016

Projects

Graph Convolutional Generative Model for Bayesian Optimization on Large Graphs
Jwala Dhamala, Sandesh Ghimire, John L. Sapp, Milan Horacek, Linwei Wang
MICCAI 2019 (early accept, finalist Young Scientist Award)
[Paper][Code]

We present a novel graph convolutional VAE to allow generative modeling of non-Euclidean data, and utilize it to embed Bayesian optimization of large graphs into a small latent space. This approach bridges the gap of previous works by introducing an expressive generative model that is able to incorporate the knowledge of spatial proximity and hierarchical compositionality of the underlying geometry. It further allows transferring of the learned features across different geometries.

Multivariate Time-series Similarity Assessment via Unsupervised Representation Learning and Stratified Locality Sensitive Hashing
Jwala Dhamala, Emmanuel Azuh, Abdullah Al-Dujaili, Jonathan Rubin, Una-May O'Reilly
IEEE Sensors Letters, NeurIPS ML4H 2018
[Paper]

We learn the representations of multivariate time-series of physiologic signals with a sequence-to-sequence auto-encoder. We then hash the learned multivariate time-series representations of labeled dataset to enable signal similarity assessment. This methodological framework is evaluated to predict Acute Hypotensive Episodes (AHE) on vital signal recordings extracted from eICU Collaborative Research Database.

High-dimensional Bayesian Optimization via an Embedded Generative Model
Jwala Dhamala, Sandesh, Ghimire, John L. Sapp, B. Milan Horacek, Linwei Wang
MICCAI 2018 (oral presentation, finalist young scientist award), WiML 2018
[Paper][Code]

We devise a novel concept that embeds a generative variational auto-encoder (VAE) into the objective function of Bayesian optimization, providing an implicit low-dimensional (LD) search space that represents the generative code of the HD spatially-varying tissue properties. In addition, the VAE-encoded knowledge about the generative code is used to guide the exploration of the search space. It is applied to estimating high-dimensional tissue excitability in a cardiac electrophysiological model.

Quantifying the Uncertainty in Model Parameters using Gaussian Process-Based Markov Chain Monte Carlo
Jwala Dhamala, Hermenegild J. Arevalo, John L. Sapp, Milan Horacek, Katherine C. Wu, Natalia A. Trayanova, Linwei Wang
IPMI 2017 (acceptance rate~30%), Medical Image Analysis

The quantification of uncertainty in model parameters is challenging because the posterior distribution of the parameters given the measurement data is non-Gaussian and the evaluation of the model is computationally expensive. In this project, we l earn a surrogate of this complicated and computationally expensive posterior distribution and utilize it to obtain a MCMC sampling with higher acceptance rate. The surrogate posterior pdf is used to accelerate the sampling of the true posterior pdf and not as a replacement.

Spatially-Adaptive Multi-scale Optimization for Local Parameter Estimation
Jwala Dhamala, Hermenegild J. Arevalo, John L. Sapp, Milan Horacek, Katherine C. Wu, Natalia A. Trayanova, Linwei Wang
MICCAI 2016 (early accept, acceptance rate~25%), IEEE TMI

We present a novel framework that, going beyond a uniform low-resolution approach, is able to obtain a higher resolution estimation of tissue properties represented by spatially non-uniform resolution. This is achieved by two central elements: 1) a multi-scale coarse-to-fine optimization that facilitates higher resolution optimization using the lower resolution solution, and 2) a spatially adaptive decision criterion that retains lower resolution in homogeneous tissue regions and allows higher resolution in heterogeneous tissue regions. The presented framework is evaluated in estimating the local tissue excitability properties of a cardiac EP model.

Activities

Student co-organizer
Hackathon on Premature Ventricular Contraction
Consortium of ECG Imaging

Relevant Publications:
[1] Sandesh, Ghimire, Jwala Dhamala, Jaume Coll-Font, Jess D Tate, Maria S Guillem, Dana H Brooks, Rob S MacLeod, and Linwei Wang. “Overcoming Barriers to Quantification and Comparison of Electrocardiographic Imaging Methods: A Community-Based Approach”. In: Proc. CinC: Computing in Cardiology, IEEE (2017) To Appear
[2] Jaume Coll-Font*, Dana H Brooks, Peter M van Dam, Jwala Dhamala, Olaf Dössel, Maria de la Salud Guillem Sánchez, Rob MacLeod, Danila Potyagaylo, Walther Schulze, Jess D Tate and Linwei Wang, “The Consortium on Electrocardiographic Imaging”, Computing in Cardiology (CINC), Vancouver, Canada, September 2016

2015 - 2017
Student co-organizer
Pre-orientation Program, Women in Computing
Rochester Institute of Technology
2018