Chen Liu's Homepage

Ph.D. Student, Yale University

chen.liu.cl2482 at yale.edu

New Haven, CT & Mountain View, CA.

Google Scholar Twitter GitHub
LinkedIn Resume CV

Acknowledgements Many thanks to Zhuang Liu for kindly providing this website template, which was adapted from Zhe Cao's website.

Chen Liu

I am a Ph.D. student in computer science at Yale University (2022~) advised by Prof. Smita Krishnaswamy [Google Scholar]. I received my M.S. from Columbia University (2018~2020), and I did my undergraduate studies at Bucknell University, a liberal arts college featuring engineering education (2014~2018).

Research Areas (Plain) I use machine learning to uncover hidden structure in complex high-dimensional data: images, videos, time series, genes, proteins, and spatial tissue maps. My works aims to unveil how diseases unfold over time, how cells organize in space, and how brain activity relates to perception, among many others.

Research Areas (Technical) I primarily work on manifold learning, a subfield of deep learning that studies how to best organize the representations in the latent space of neural networks.

Most of my research involve direct applications to healthcare or medicine, via analysis of medical images and omics (genomics, transcriptomics, proteomics) data. I also work on fusing multiple modalities, modeling time-varying dynamics, and learning from limited or no labels.

In the remaining years of my Ph.D., I would like to spend more time on taming neural networks, including generative models, to better comply with constraints such as physical laws, with applications in modeling the natural progression of diseases in time-varying medical images.

Full-Time Experience

[2021-2022] I worked as a Senior Research Scientist at GE Healthcare on deep learning in medical imaging applications.

[2021] I went to the industry and joined a startup company Matic and developed SLAM algorithms for housekeeping robots.

[2019-2020] I worked as a full-time research assistant at Columbia University Medical Center in a medical imaging lab.

News

[06/2025] C-ManifoldWalk is accepted to ICML 2025 GenBio workshop as a Spotlight (links to be added).

[05/2025] Introducing Brainteaser, a benchmark to assess creativity of LLM reasoning.

[03/2025] Our paper on clinical time-series prediction for organ transplantation was accepted to Scientific Reports.

[01/2025] Geometry-Aware Generative Autoencoder (GAGA) was accepted to AISTATS 2025.

[12/2024] 3/3 papers (2 first-authored, both Oral) were accepted to ICASSP 2025.

[11/2024] I was recognized as a Top Reviewer at NeurIPS 2024.

[07/2024] I developed the first free tool to generate your Google Scholar Citation World Map.

[06/2024] My first project during my Ph.D. was accepted to MICCAI 2024.

[08/2022] I started my Ph.D. journey at Krishnaswamy Lab, Yale University.

[06/2022] I was recognized as an Outstanding Reviewer at ICML 2022.

Selected Recent Publications (Asterisk denotes equal contribution)

Manifold learning Dynamics modeling Self-supervised learning

⏳ ImageFlowNet: Forecasting Multiscale Image-Level Trajectories of Disease Progression with Irregularly-Sampled Longitudinal Medical Images

I designed a position-parameterized neural ODE that flows the multiscale latent representations, so that we can predict a future image given an earlier image and the change in time. For example: ``Predict how this patient's eye will look like if we leave the disease untreated for 2 years.''

Chen Liu*, Ke Xu*, Liangbo L. Shen, Guillaume Huguet, Zilong Wang, Alexander Tong, Danilo Bzdok, Jay Stewart, Jay C. Wang, Lucian V. Del Priore, Smita Krishnaswamy

ICASSP 2025 Oral Presentation 🔊

Manifold learning Multimodal fusion Self-supervised learning

🧬 ImmunoStruct: a multimodal neural network framework for immunogenicity prediction from peptide-MHC sequence, structure, and biochemical properties

ImmunoStruct predicts immunogenicity of protein MHC complexes by fusing information from multiple biological modalities: sequence, structure and biochemical properties. I designed a novel cancer-wildtype contrastive learning objective to establish a new state of the art in the field, by encouraging immunogenicity-aware pairwise similarity and suppressing feature space collapse.

Kevin Bijan Givechian, Joao Felipe Rocha*, Edward Yang*, Chen Liu*, Kerrie Greene, Rex Ying, Etienne Caron, Akiko Iwasaki, Smita Krishnaswamy

bioRxiv 2024, under review at Nature Machine Intelligence

Manifold learning Self-supervised learning

✂️ CUTS: A Deep Learning and Topological Framework for Multigranular Unsupervised Medical Image Segmentation

Chen Liu*, Matthew Amodio*, Liangbo L. Shen, Feng Gao, Arman Avesta, Sanjay Aneja, Jay C. Wang, Lucian V. Del Priore, Smita Krishnaswamy

I introduced CUTS, a novel multiscale unsupervised segmentation framework. It first uses intra-image contrastive learning and local patch reconstruction to organize a meaningful pixel-level embedding space, and then produces multiscale assignments with diffusion condensation. On datasets with few training samples, CUTS performs on par or better than Segment Anything methods.

MICCAI 2024

Ph.D. Student, Yale University

chen.liu.cl2482 at yale.edu

New Haven, CT & Mountain View, CA.

Google Scholar Twitter GitHub LinkedIn Resume CV

Acknowledgements Many thanks to Zhuang Liu for kindly providing this website template, which was adapted from Zhe Cao's website.

Chen Liu

Full-Time Experience

News

Selected Recent Publications (Asterisk denotes equal contribution)

⏳ ImageFlowNet: Forecasting Multiscale Image-Level Trajectories of Disease Progression with Irregularly-Sampled Longitudinal Medical Images

🧬 ImmunoStruct: a multimodal neural network framework for immunogenicity prediction from peptide-MHC sequence, structure, and biochemical properties

✂️ CUTS: A Deep Learning and Topological Framework for Multigranular Unsupervised Medical Image Segmentation

Conference Papers (Asterisk denotes equal contribution)

To Optimize, Not to Invent: Controlled On-Manifold Diffusion Walk for mRNA Sequence Generation and Optimization Without de novo Design

⏳ ImageFlowNet: Forecasting Multiscale Image-Level Trajectories of Disease Progression with Irregularly-Sampled Longitudinal Medical Images

⚔️ DiffKillR: Killing and Recreating Diffeomorphisms for Cell Annotation in Dense Microscopy Images

🌐 Geometry-Aware Generative Autoencoders for Warped Riemannian Metric Learning and Generative Modeling on Data Manifolds

Hyperedge Representations with Hypergraph Wavelets: Applications to Spatial Transcriptomics

✂️ CUTS: A Deep Learning and Topological Framework for Multigranular Unsupervised Medical Image Segmentation

🎲 Assessing Neural Network Representations During Training Using Noise-Resilient Diffusion Spectral Entropy

Segmentation with Residual Attention U-Net and an Edge-Enhancement Approach Preserves Cell Shape Features

Deep Learning Identifies Neuroimaging Signatures of Alzheimer's Disease Using Structural and Synthesized Functional MRI Data

Substituting Gadolinium in Brain MRI Using DeepContrast

Journal Publications (Asterisk denotes equal contribution)

🏥 Deep learning unlocks the true potential of organ donation after circulatory death with accurate prediction of time-to-death

Deep learning of MRI contrast enhancement for mapping cerebral blood volume from single-modal non-contrast scans of aging and Alzheimer's disease brains

In vivo γ-aminobutyric acid increase as a biomarker of the epileptogenic zone: An unbiased metabolomics approach

Reduced Hippocampal GABA is associated with Poorer Episodic Memory in Healthy Older Women: A Pilot Study

Understanding and Modeling Climate Impacts on Photosynthetic Dynamics with FLUXNET Data and Neural Networks

Preprints (Asterisk denotes equal contribution)

🧠 Creativity or Brute Force? Using Brainteasers as a Window into the Problem-Solving Abilities of Large Language Models

🧬 ImmunoStruct: a multimodal neural network framework for immunogenicity prediction from peptide-MHC sequence, structure, and biochemical properties

Adversarial Focal Loss: Asking Your Discriminator for Hard Examples

Patents (Asterisk denotes equal contribution)

System and Method for Obtaining Accurate Measurements and Quantification of X-Ray Image from Estimation of Key Anatomical Locations

X-Ray Lead Marker Detection System for X-Ray Imaging System

Conference Abstracts, excluding those published elsewhere (Asterisk denotes equal contribution)

DL-BET-A Deep Learning Based Tool for Automatic Brain Extraction from Structural Magnetic Resonance Images in Mice

JET-A Matlab Toolkit for Automated J-Difference-Edited MR Spectra Processing of In Vivo Mouse MEGA-PRESS Study at 9.4 T

Predicting Gadolinium Contrast Enhancement for Structural Lesion Analysis using DeepContrast

Blogs and Side Projects (Asterisk denotes equal contribution)

🌏 CitationMap: A Python Tool to Identify and Visualize Your Google Scholar Citations Around the World

A Technical Deep Dive into Drag Your GAN (DragGAN)

Towards a Large-Scale Unbiased Machine Learning Benchmark for Cell Instance Segmentation: Final Report for CPSC 537 Intro to Database Systems

Others

Google Scholar Twitter GitHub
LinkedIn Resume CV