Research overview

My research is in machine learning/data mining and natural language processing, with an emphasis on applications in health informatics.

For example, one of my major ongoing research aims concerns optimizing the processes of evidence-based medicine using novel natural language processing and machine learning methods. The aim is to reduce the (human) workload involved in conducting systematic reviews (i.e., making sense of the biomedical literature), so that we can realize the aim of evidence-based care in an era of information overload. An ongoing project in this direction is RobotReviewer. Here is an article written for a general audience about some of this work. I have also talked about this project at length on the NLP Highlights podcast, available here.

More broadly, I am interested in core machine learning and natural language processing issues: e.g., structured and unstructured classification techniques; neural models; semi-supervised learning methods; learning from imbalanced data; and learning from alternative forms of supervision. I tend to be most excited by interdisciplinary research that motivates technical questions by way of interesting applications.

A random sample of semi-recent publications

Iain J. Marshall, Rachel Marshall, Byron C. Wallace, Jon Brassey, and James Thomas. Rapid reviews may produce different results to systematic reviews: a meta-epidemiological study Journal of Clinical Epidemiology; 2019.

Ye Zhang, Matt Lease and Byron C. Wallace. Active Discriminative Text Representation Learning AAAI Conference on Artificial Intelligence (AAAI); 2017.

Silvio Amir, Glen Coppersmith, Paula Carvalho, Mário J. Silva and Byron C. Wallace. Quantifying Mental Health from Social Media with Neural User Embeddings Machine Learning in Health Care (MLHC); 2017.

News

08/15/2019 NIH/NLM R01 Renewed

The NIH has renewed the R01 grant that supports our work on RobotReviewer and related methods!

06/24/2019 NSF Grant

The NSF has awarded Jan-Willem van de Meent and I a grant to study disentangled representations for text!

05/20/2019 Early career spotlight @ IJCAI

I'll be giving an "early career spotlight" talk at IJCAI in August.

07/12/2018 ARO grant

The Army Research Office (ARO) has has selected my proposal, "Neural Models for Text: Improving Efficiency, Interpretability and Accuracy" for funding.

Support

My work has been supported with grants from the National Institutes of Health, National Science Foundation (including a CAREER grant), the Army Research Office, Seton hospital, Amazon and seed funds from Brown University.