I lead a research group at the Institute for Infocomm Research (I2R), A*STAR, focusing on data-efficient deep learning. Our work is inspired by applications in healthcare and manufacturing where collecting large, well-annotated datasets is often time and cost-prohibitive due to the need for careful expert labelling.
I earned a BS ('08), MS ('12) and PhD ('17) in Computer Science from Stanford, where I had the privilege to work with Professors Anshul Kundaje, Daphne Koller, Andrew Ng and Serafim Batzoglou on machine learning and deep learning for biology. I took a year off during my PhD to work for Counsyl, where I developed software (variant calling and LIMS integration) for clinically validated sequencing-based genetic tests.
Semi-supervised Audio Classification with Consistency-Based Regularization
INTERSPEECH (2019)
[paper (coming soon)]TEA-DNN: the Quest for Time-Energy-Accuracy Co-optimized Deep Neural Networks
ACM/IEEE International Symposium on Low Power Electronics and Design (2019)
[paper (coming soon)]MaxpoolNMS: Getting Rid of NMS Bottlenecks in Two-Stage Object Detectors
IEEE Conference on Computer Vision and Pattern Recognition (2019)
[paper]Encoding Knowledge Graph with Graph CNN for Question Answering
International Conference on Learning Representations Workshop on Representation Learning on Graphs and Manifolds (2019)
[paper]Optimistic mirror descent in saddle-point problems: Going the extra(-gradient) mile
International Conference on Learning Representations (2019)
[paper, code (coming soon)]The Unusual Effectiveness of Averaging in GAN Training
International Conference on Learning Representations (2019)
[paper, code]Semi-Supervised Deep Learning for Abnormality Classification in Retinal Images
NeurIPS Workshop on Machine Learning for Health (2018)
[paper, code (coming soon)]Predicting thermoelectric properties from crystal graphs and material descriptors – first application for functional materials
NeurIPS Workshop on Machine Learning for Molecules and Materials (2018)
[paper]Adversarially Learned Anomaly Detection
IEEE International Conference on Data Mining (2018)
[paper, arXiv, code]Semi-Supervised Learning With GANs: Revisiting Manifold Regularization
International Conference on Learning Representations, Workshop Track (2018)
[paper, arXiv, code]Learning RNA secondary structure (only) from structure probing data
ICML Workshop on Computational Biology (2017)
[paper (bioRxiv)]The C2H2-ZF transcription factor Zfp335 recognizes two consensus motifs using separate zinc finger arrays
Genes & Development 30: 1509-1514 (2016)
[paper]Zinc finger protein Zfp335 is required for the formation of the naïve T cell compartment
eLife 3: e03549 (2014)
[paper]A majorization-minimization algorithm for (multiple) hyperparameter learning
Proceedings of the 26th International Conference on Machine Learning, 321–328 (2009)
[pdf, slides]Proximal regularization for online and batch learning
Proceedings of the 26th International Conference on Machine Learning, 257–264 (2009)
[pdf, code]Searching for rising stars in bibliography networks
Proceedings of the 14th International Conference on Database Systems for Advanced Applications (2009)
[pdf]A max-margin model for efficient simultaneous alignment and folding of RNA sequences
Bioinformatics, 24(13):i68–76 (2008)
[pdf, code]Efficient multiple hyperparameter learning for log-linear models
Advances in Neural Information Processing Systems 20:377–384 (2007)
[pdf]Discovering protein complexes in dense reliable neighborhoods of protein interaction networks
Computational Systems Bioinformatics: CSB 2007 Conference Proceedings, 6:157–68 (2007)
[pdf, website]Interaction graph mining for protein complexes using local clique merging
[ Best Paper Award, GIW 2005 ]Genome Informatics, 16(2):260–9 (2005)
[pdf]
Deep learning the relationship between chromatin architecture, chromatin state, and transcription factor binding [ Platform Talk ]
65th Annual Meeting of The American Society of Human Genetics
ATAC-seq is predictive of chromatin state
2015 Cold Spring Harbor Laboratory meeting on Systems Biology: Global Regulation of Gene Expression
ATAC-seq is predictive of chromatin state [ Oral Presentation ]
2014 RECOMB/ISCB Conference on Regulatory and Systems Genomics, with DREAM Challenges and Cytoscape Workshops
Ab-initio identification of chromatin states from chromatin accessibility data with CASCADE
2014 Cold Spring Harbor Laboratory meeting on The Biology of Genomes
Validation of and initial experience with a next-generation sequencing-based 98-gene expanded carrier screening assay
Association for Molecular Pathology 2013 Annual Meeting
Gene expression deconvolution
Biomedical Computation at Stanford 2010