Wenlong Zhao
wenlongzhao@cs.umass.edu
I’m a second-year Computer Science PhD student at UMass Amherst, working with Prof. Andrew McCallum at Information Extraction and Synthesis Lab (IESL). I completed a Computer Science master’s degree also at UMass, prior to which I graduated with a major in Mathematics and Economics, a major with distinction in Cognitive Science, a minor in Computer Science, and a minor in Philosophy from UC San Diego.
My recent research focuses on probabilistic box embeddings and energy-based models, their learning and inference algorithms, and efficiency-accuracy trade-offs. I am also enthusiastic about building systems. I am part of a multi-university AI-for-science project series led by Prof. Subhransu Maji and Prof. Dan Sheldon on developing machine learning algorithms to detect and track bird and bat roosts in weather radar data for conservation biology and ecology research. I worked with Prof. Mohit Iyyer and Prof. Brendan O’Connor on developing a crowdsourcing platform for collecting coreference resolution annotations. I have also contributed to AI-for-sociology research about using computational models to study peer review processes.
I am excited to see AI become useful tools in industry. I have worked with Adobe collaborators to develop an authoring assistant that generates texts according to intent guidance. I have collaborated with Amazon Alexa as a Summer 2022 intern and via the UMass Industry Mentorship Program to investigate computational efficiency of large language models.
In a longer term, I am interested in both AI applications and the possibility to build AI that learns to interact with the world based on reliable common sense, knowledge, and reasoning. Relevant work include my Spring 2023 collaboration with Allen AI mentors about editing memory in transformer models.
Publications
- NLPCollecting Crowdsourced Annotations for Coreference ResolutionFindings of ACL: EACL 2023
- CVLong-term analysis of persistence and size of swallow and martin roosts in the US Great LakesRemote Sensing in Ecology and Conservation 2023
- CVUsing Spatio-Temporal Information in Weather Radar Data to Detect and Track Communal Bird RoostUnder review 2022
- CVQuantifying long-term phenological patterns of aerial insectivores roosting in the Great Lakes region using weather surveillance radarGlobal Change Biology 2022
- MLStructured Energy Network As a LossNeurIPS 2022
- MLToward Compact Parameter Representations for Architecture-Agnostic Neural Network CompressionarXiv preprint arXiv:2111.10320 2021
- NLPIGA: An intent-guided authoring assistantEMNLP 2021
- NLPCompressing transformer-based semantic parsing models using compositional code embeddingsFindings of ACL: EMNLP 2020
- NLPA Thorough Examination of the RACE Machine Reading Comprehension TaskUC San Diego Cognitive Science Undergrad Honors Thesis 2019
- NLPRethinking exposure bias in language modelingarXiv preprint arXiv:1910.11235 2019
Teaching
- Mentor for UMass Amherst COMPSCI 696: Independent Study, Fall 2022
- Mentor for UMass Amherst COMPSCI 696DS : Industry Mentorship Program, Spring 2022
- Grader for UMass Amherst COMPSCI 611: Advanced Algorithm, Spring 2021
- Teaching assistant for UCSD COGS 118A: Introduction to Machine Learning, Winter 2018
- Teaching assistant for UCSD COGS 10: Cognitive Consequences of Technology, Spring 2017
Service
- Organizer: UMass Machine Learning and Friends Lunch (MLFL) series.
- Reviewer: COLING 2022.