Ting-Kuei Hu
San Diego,California
Summary
Enthusiastic and self-motivated imaging engineer with a broad interest for learning-based computer vision challenges. Experienced in a variety of projects,including super-resolution, adversarial robustness, efficient inference, multi-modal dynamic gesture recognition, and learning to control. Possesses a deep understanding of the field, both theoretically and practically, built through over four years of focused research in machine learning, resulting in publications in top-tier conferences and journals such as AAAI, ICLR, NIPS, to name a few. Further, brings three years of industry expertise as an Imaging Algorithm Engineer in Qualcomm's Camera system team, showcasing a consistent track record of innovation through the filing of 3+ IDF/patent applications for camera imaging solutions. Proficient in programming languages like C/C++, Python, and well-versed in common deep learning frameworks such as Pytorch and Tensorflow.
Overview
9
9
years of professional experience
Work History
Senior Camera Imaging Algorithm Engineer
Qualcomm Technologies
01.2021 - Current
- Explored and improved real-world super-resolution (SR) algorithms through research and development. Introduced creative training concepts like noise pool, learning-based downscaler, and content-aware perceptual loss to reduce differences between training and test sets. Achieved high-quality images based on both PSNR and subjective assessments.
- Explored various model optimization techniques for the deployment of hardware-efficient neural networks, including, efficient architecture, quantization, pruning, etc. Implementing these techniques resulted in the model reducing half of its parameters without significant degradation.
Graduate Research Assistant
Texas A&M University
08.2018 - 12.2020
- Robust Dynamic Inference Network: a new defense strategy against adversarial attacks by simultaneously optimizing accuracy, robustness, and efficiency. This defense method for multi-exit networks has become a widely accepted standard for evaluating robustness.
- Learning to Control: Introduced a learning-based decentralized controller for coordinating swarms. This approach, for the first time, demonstrated performance comparable to a centralized controller while relying solely on locally observed visual features.
- Authored and Co-authored top-tier AI/ML conferences and journals such as ICLR/NIPS/ICASSP, etc.
Full-Time Research Assistant
Academia Sinica
07.2014 - 08.2018
- Developed Multi-Modal Visual System: Using 3D-CNN + LSTM with a newly introduced a creative network design to fuse information from different modals, such as RGB/Depth or RGB/accelerometers to enhance the quality of video gesture recognition. Prototyped system is Implemented on a practical mobile phone (Samsung Note 7) with minimal delay and high accuracy. Research outcome is publish in AAAI, a top tier AI conference.
Education
Master of Engineering - Computer Science
Texas A&M University
College Station, TX12.2020
Bachelor of Arts - Electronics Engineering
National Chiao Tung University
Hsinchu07.2014
Skills
- Deep Learning
- Artificial Intelligence / Machine Learning
- C/C
- Python
- Pytorch
- Computer Vision
- Image Processing
SELECTED PUBLICATION
1.W. Zheng, T. Chen, T. Hu and Z. Wang “Symbolic learning to optimize: Towards interpretability and scalability”, International Conference on Learning Representations (ICLR), 2022.
2.T. Hu, F. Gama, T. Chen, W. Zheng, Z. Wang, A. Ribeiro and B. M. Sadler “Scalable Perception-Action-Communication Loops with Convolutional and Graph Neural Networks”, IEEE Transaction on Signal and Information Processing over Networks (TISPN), 2022.
3. T. Hu, F. Gama, T. Chen, Z. Wang, A. Ribeiro and B. M. Sadler “VGAI: A Vision-Based Decentralized Controller Learning Framework for Robot Swarms”, The international Conference on Acoustics, Speech, & Signal Processing (ICASSP), 2021.
4.H. Ma, T. Chen, T. Hu, C. You, X. Xie and Z. Wang “Undistillable: Making A Nasty Teacher That CANNOT teach students.”, International Conference on Learning Representations (ICLR), 2021.
5. H. Wang, T. Chen, S. Gui, T. Hu, J. Liu and Z. Wang “Once-for-All Adversarial Training: In-Situ Tradeoff between Robustness and Accuracy for Free”, Advances in Neural Information Processing Systems (NIPS), 2020.
6. T. Hu, T. Chen, H. Wang and Z. Wang “Triple Wins: Boosting Accuracy, Robustness and Efficiency Together by Enabling Input-Adaptive Inference”, International Conference on Learning Representations (ICLR), 2020.
7. Yue. Wang, J. Shen, T. Hu, P. Xu, T. Nguyen, R. G. Baraniuk, Z. Wang and Y. Lin “Dual Dynamic Inference: Enabling More Efficient, Adaptive and Controllable Deep Inference”, IEEE Journal of Selected Topics in Signal Processing (JSTSP), 2020.
8. T. Hu, Y. Lin, and P. Hsiu, “Learning Adaptive Hidden Layers for Mobile Gesture Recognition”, AAAI Conference on Artificial Intelligence (AAAI), 2018.
9. C. Cheng, P. Hsiu, T. Hu, and T. Kuo, “Oasis: A Mobile Cyber-Physical System for Accessible Location Exploration”, Proceedings of the IEEE (PIEEE), 2018.
10. C. Feng, T. Hu, J. Chang, and W. Fang, “A Reliable Brain Computer Interface Implemented on an FPGA for a Mobile Dialing System”, IEEE International Symposium on Circuits & Systems (ISCAS), 2014.
Timeline
Senior Camera Imaging Algorithm Engineer
Qualcomm Technologies
01.2021 - Current
Graduate Research Assistant
Texas A&M University
08.2018 - 12.2020
Full-Time Research Assistant
Academia Sinica
07.2014 - 08.2018
Master of Engineering - Computer Science
Texas A&M University
Bachelor of Arts - Electronics Engineering
National Chiao Tung University
Similar Profiles
Santhosh DyavanapallySanthosh Dyavanapally
Senior Lead Software Engineer at Qualcomm TechnologiesSenior Lead Software Engineer at Qualcomm Technologies
CAMILA FLETCHERCAMILA FLETCHER
LOGISTICS SPECIALIST at Qualcomm TechnologiesLOGISTICS SPECIALIST at Qualcomm Technologies
Julia HarmonJulia Harmon
Senior Program Manager, Extended Reality (XR) at Qualcomm Technologies, Inc.Senior Program Manager, Extended Reality (XR) at Qualcomm Technologies, Inc.