About me [Scholar, Twitter, DBLP]

I am a Senior Research Scientist with the Learning and Perception Research team led by Jan Kautz, at NVIDIA Research. I received my Ph.D. from Princeton University, advised by Prof. Niraj K. Jha.

I'm interested in data-/executition-efficient and secure deep learning.

We are always looking for highly-motivated Ph.D. research interns that have passion and experience in efficient and secure deep learning. If interested please reach me at: dannyy@nvidia.com

 

Publications - partial list (full list in scholar*: equal contrib.)

  • H. Yin, A. Vahdat, J. Alvarez, A. Mallya, J. Kautz, P. Molchanov, "A-ViT: Adaptive Tokens for Efficient Vision Transformer," in CVPR 2022. (Oral) [paper, project page, code]
  • A. Hatamizadeh*, H. Yin*, H. Roth, W. Li, J. Kautz, D. Xu, P. Molchanov, "GradViT: Gradient Inversion of Vision Transformers," in CVPR 2022. [paper, project page]
  • M. Shen, P. Molchanov, H. Yin, J. Alvarez, "When to Prune? A Policy towards Early Structural Pruning," in CVPR 2022. [paper]
  • X. Dong, H. Yin, J. Alvarez, J. Kautz, P. Molchanov, "Deep Neural Networks are Surprisingly Reversible: A Baseline for Zero-Shot Inversion," in CVPRW 2022. [paper, MIT Tech Review]
  • P. Molchanov*, J. Hall*, H. Yin*, J. Kautz, N. Fusi, A. Vahdat, "HANT: Hardware-Aware Network Transformation," in ECCV 2022. [paper]
  • H. Yang, H. Yin, P. Molchanov, H. Li, J. Kautz, "NViT: Vision Transformer Compression and Parameter Redistribution," arXiv 2021. [paper]
  • H. Yin, A. Mallya, A. Vahdat, J. Alvarez, J. Kautz, P. Molchanov, "See through Gradients: Image Batch Recovery via GradInversion," in CVPR 2021. [paper]
  • Y. Idelbayev, P. Molchanov, M. Shen, H. Yin, M. C.-Perpinan, J. Alvarez, "Optimal Quantization using Scaled Codebook," in CVPR 2021. [paper]
  • A. Chawla, H. Yin, P. Molchanov, J. Alvarez, "Data-Free Knowledge Distillation for Object Detection," in WACV, 2021. [paper, code]
  • W. Xia, H. Yin, X. Dai, N. K. Jha, "Fully Dynamic Inference with Deep Neural Networks," IEEE Trans. Emerging Topics in Computing, 2021. [paper]
  • H. Yin*, P. Molchanov*, J. Alvarez, Z. Li, A. Mallya, D. Hoiem, N. K. Jha, and J. Kautz, "Dreaming to Distill: Data-free Knowledge Transfer via DeepInversion," in CVPR, 2020. (Oral) [paper, code, video]
  • W. Xia, H. Yin, N. K. Jha, "Efficient Synthesis of Compact Deep Neural Networks," in DAC, 2020. [paper]
  • H. Yin, B. Mukadam, X. Dai, and N. K. Jha, “DiabDeep: Pervasive Diabetes Diagnosis based on Wearable Medical Sensors and Efficient Neural Networks,” IEEE Trans. Emerging Topics in Computing, 2019. [paper]
  • X. Dai, P. Zhang, B. Wu, H. Yin, F. Sun, Y. Wang, M. Dukhan, Y. Hu, Y. Wu, Y. Jia, P. Vajda, M. Uyttendaele, N. K. Jha, “ChamNet: Towards Efficient Network Design through Platform-Aware Model Adaptation,” in CVPR, 2019. [paper, code]
  • H. Yin, G. Chen, Y. Li, S. Che, W. Zhang, and N. K. Jha, “Hardware-guided symbiotic training for compact, accurate, yet execution-efficient LSTMs,” IEEE Trans. Emerging Topics in Computing, 2021. [paper]
  • X. Dai*, H. Yin*, and N. K. Jha, “Grow and prune compact, fast, and accurate LSTMs,” IEEE Trans. Computers, 2019. [paper]
  • X. Dai, H. Yin, and N. K. Jha, “NeST: A neural network synthesis tool based on a grow-and-prune paradigm,” IEEE Trans. Computers, 2018. [paper]
  • H. Yin, B. Gwee, Z. Lin, A. Kumar, S. Razul, C. See, "Novel Real-time System Design for Floating-point sub-Nyquist Multi-coset Signal Blind reconstruction," in ISCAS2015. (Oral)

 

Patents

  • NeST: A neural network synthesis tool based on a grow-and-prune paradigm, US 62/580525
  • Grow and prune compact, fast, yet accurate LSTMs, US 62/677232
  • Incremental learning using a grow-and-prune paradigm with efficient neural networks, US 62/851740
  • DiabDeep: pervasive diabetes diagnosis based on wearable medical sensors and efficient neural networks, US 62/862354
  • A hierarchical health decision support system based on wearable medical sensors and machine learning ensembles, US 16/475879
  • Smart, secure, yet energy-efficient Internet-of-Things sensors, US 62/615475
  • Hardware-guided symbiotic training for compact, accurate, yet execution-efficient LSTMs, US 16/374738industrial
  • Synthesizing data for training one or more neural networks, US 16/682967industrial
  • (out-dated, updated till 2019)

 

Invited Talk

  • 'Towards Efficient and Secure Deep Learning', invited keynote, Design & Automation Conference (DAC'59), San Francisco, July, 2022.
  • 'Towards Efficient and Secure Deep Nets', invited seminar, University of British Columbia ECE department, virtual, May, 2022.
  • 'Inverting Deep Nets', invited seminar, Princeton University research group, virtual, Aug., 2021.
  • 'See through Gradients', invited guest speaker, Europe ML meeting, virtual, Apr., 2021.
  • 'Dreaming to Distill', invited talk, Synced AI (机器之心), virtual, Jun., 2020.
  • 'Dreaming to Distill', invited seminar, Facebook AR/VR, virtual, Jun., 2020.
  • 'Efficient Neural Networks', invited talk, Alibaba Cloud / Platform AI, virtual, Feb., 2020.
  • 'Efficient Neural Networks', research seminar, NVIDIA Research, Facebook AI, Alibaba, Baidu Research, ByteDance AI Lab, etc., California, USA, Dec., 2019.
  • 'Applied Machine Learning: From Theory to Practice', invited talk, IEEE Circuits and Systems Society (Singapore Chapter), Singapore, Feb., 2018.
  • 'A Health Decision Support System for Disease Diagnosis based on Wearable Medical Sensors and Machine Learning Ensembles', invited poster, New Jersey Tech Concil's 'What's Next in Medical Devices' Forum, Princeton, NJ, USA, Jun., 2016.

 

Reviewer & committee

  • Conferences: CVPR, NeurIPS, ICML, ICCV, ECCV, WACV, AAAI, HPCA, DAC, etc.
  • Journals: TPAMI, IJCV, JBHI, JSTSP, SPL, JHTM, etc.

 

Princeton Mentees

  • (19-20) Joe Zhang 20', EE, (Stanford)
  • (19-20) Hari Santhanam 20', EE (UPenn, Robotics)
  • (18-19) Frederick Hertan 19', Math (SIG trading)
  • (18-19) Bilal Mukadam 19', EE (Microsoft) 
  • (18-19) Kyle Johnson 21', EE
  • (17-18) Chloe Song 18', EE (Astra)

 

Teaching Assistant

  • (17-18 F') ELE364 Machine Learning for Predictive Data Analytics
  • (16-17 S') ELE464 Embedded Computing

 

 

<embed>
Copy and paste this code to your website.
Copy and paste this code to your website.