Larry Zhang
Dual PhD Student in Complex Network Systems and Intelligent Systems Engineering
Indiana University Bloomington
Biography
I am a Dual PhD Student in the Department of Informatics (specializing in Complex Network Systems), and the Department of Intelligent Systems Engineering (specializing in Neuroengineering) at Indiana University Bloomington (IU) as a Fellow of the CNS-NRT Program funded by the National Science Foundation (NSF). I am co-advised by Yong-Yeol Ahn and Gregory Lewis.
In 2023, I am working at the National Institute for Informatics (NII) in Tokyo, Japan from January to May on psychophysiological applications in information interaction and search, Samsung Digital Health Lab in Mountain View, California from May to September on stress sensing, and Central European University in Vienna, Austria from September to December on higher-order interactions in temporal brain fingerprinting research.
My research focuses on applying computational and cognitive models to measurements from multimodal behavioral sensing technologies (e.g. voice recordings, heart rate, pose, language) to enable and accelerate theory building around neuropsychiatric health and general wellbeing. I am particularly interested in developing cheap, non-invasive measures of neuropsychiatric state to make psychiatric healthcare more accessible to all. Currently, I am studying computational methods for assessing cross-modal synchrony in behavioral and neuroscientific studies.
Prior to IU, I worked on research at the Institute of Creative Technologies at the University of Southern California, where I worked with Mohammad Soleymani and Stefan Scherer studying multimodal behavioral sensing in substance abuse disorders and psychological distress. I also worked with Reza Hosseini Ghomi and Trevor Cohen at the University of Washington studying vocal biomarkers and the clinical application of semantic memory networks for cognitive decline. I continue to study semantic memory search with Michael Jones at IU.
Download Larry’s resumé.
- Affective Computing
- Behavioral Signal Processing
- Behavioral and Physiological Synchrony
- Digital Neuropsychiatry Applications
- Cognitive Science
- Accessible Mental Healthcare
- Multimodal Machine Learning
-
Dual PhD in Complex Network Systems and Intelligent Systems Engineering, 2020 - Present
Indiana University Bloomington
-
BS in Electrical Engineering (Specialization in Machine Learning and Controls), 2013 - 2017
University of California, San Diego
Skills
Acoustic, Linguistic, Physiology, Body & Head Pose
Plotly, Altair, Seaborn, Vega
MySQL, PostgreSQL, AWS
Digital Biomarkers, Clinical Study Design, Ethics Approval
Journal articles, Grants & Fellowships, Reviewing
Fury.gl, Blender, VTK, trimesh
Machine Learning, Deep Learning, Network Science, Dynamical Systems
Scientific Computing and Package Development
Conference Presentations and Public Speaking
Experience
- Conducting Research with Faculty across Intelligent Systems Engineering, Cognitive Science, and Informatics.
- Researching information theoretic and tensor factorization methods on time-varying synchronization phenomena including behavioral, physiological, and neural synchronization studies.
- Proposed new predictive model of semantic memory search outperforming existing methodologies.
- Developed multimodal cross-atlas reference brain organ for user studies in HuBMAP interface.
- Investigated multimodal sensing on vocal behaviors for clinical interviewing applications.
- Published paper on automatic coding in motivational interviewing via representation learning.
- Led study linking multimodal vocal behaviors to psychological distress symptoms in virtual human interviews.
- Led research efforts and data science projects; manage and develop projects for 20 tribe members.
- Developed computational analysis direction for clients including Harvard, Stanford, Biogen, LEO I-Lab.
- Explored machine learning and signal processing methodology for internal product improvement.
- Published research on vocal markers of cognitive decline in Framingham Heart Study Cognitive Aging Cohort
- Lead machine learning analysis across voice and survey data in adolescent depression study.
- Piloted development of robust ML models on language and acoustic data.
- Devised data analysis approach to identify key features with predictive power in remission detection.
- Co-first author abstract accepted to Society for Biological Psychiatry Conference.
- Lead developer of lab’s data science team, supervised 12 undergraduate and graduate students.
- Built acoustic models of depression with near-SOTA performance on non-curated clinical datasets.
- Established extensive data-featurization pipeline to collect thousands of features to model voice data.
- Published two manuscripts on voice biomarkers for depression on psychomotor disturbances.
- Deployed automation framework to accelerate data collection and analysis from key GPU endpoints.
- Established end-to-end analytics framework for circuit verification insights from millions of datapoints.
- Achieved 94% accuracy classifying silicon health of data; predicted parameters within ±0.5 loss range.
- Tuned parameters for a Neural Collaborative Filtering model on corpus of resumes and job postings.
- Implemented Bi-recurrent LSTM on IMDB Sentiment Dataset in BigDL framework as proof of concept.
- Implemented LSTM in a time series-based model on Pollution Dataset in BigDL as proof of concept.
- Investigated implementation of convolutional neural networks on the TrueNorth neuromorphic chip.
- Implemented EEDN spiking network implementation of convolutional neural networks on TrueNorth.
- Presented on backprop methodology utilizing binary weights in spiking network to discussion group.
