Résumé

Wei Wang

I am currently a Ph.D. student in Computer Science at University at Buffalo, the State University of New York (SUNY).

Research Interests

Mobile and acoustic sensing, human-computer interaction, machine learning, biometrics

Education

University at Buffalo, SUNY Buffalo, NY Spring 2018 – Present
Ph.D. Student in Computer Science

Binghamton University, SUNY Binghamton, NY Fall 2015 – Spring 2018
Ph.D. Student in Electrical and Computer Engineering

Northwestern Polytechnical University Xi’an, China Fall 2008 – Fall 2012
B.S. in Automation

Experience

Rochester Data Science Consortium Rochester, NY Summer 2018
Research Intern
Implemented deep learning models with high-pass filters for image forensics. Developed machine learning models for ECG signal quality detection.

Baidu Ventures Beijing, China Fall 2015 – Spring 2018
Intern
Assessed and analyzed the technical strengths, limitations, feasibility, market potentials and core technology competencies of AI startup companies; conducted literature survey and comparative analysis of state-of-the-art AI technologies.

University at Buffalo, SUNY Buffalo, NY
Teaching Assistant
CSE 341 Computer Organization Spring 2018, Spring 2019
CSE 522 Object Oriented Analysis, Design, and Implementation Fall 2018, Fall 2019

Binghamton University, SUNY Binghamton, NY Fall 206 – Spring 2018
Research Assistant

Research Projects

Mobile Sensing, Biometrics and Human-Computer Interaction
My research seeks to explore and investigate emerging and innovative approaches for mobile sensing and interaction, specifically the acquisition, processing and application of ear canal acoustic echo, PPG, EEG, IMU sensor data and Wi-Fi.

EarEcho: Using Ear Canal Echo for Content-Independent Wearable (Earphone) Authentication:

Performed event detection, interference elimination on captured ear canal acoustic characteristics.
Extracted canal-structure-related features based on Welch’s power spectral density estimation.
Used for user authentication, achieving 97.6% accuracy and robustness in diverse environments.

EEG Based Emotion-Aware Music Recommendation System:

Proposed an algorithm to predict users’ emotional reactions towards different music utilizing EEG.
Designed a music recommendation system that considers current and target emotional states.

Camera-Based Non-Contact Continuous PPG Monitoring & Authentication System:

Carried out signal amplification and PPG extraction based on facial videos.
Calculated heart rate and blood pressure estimation based on extracted PPG signal.
Implemented PPG-based user authentication with an average accuracy of 86.67%.

Continuous Authentication System Based on Keystroke, Mouse and Wrist Motions:

Feature extraction from keystroke (temporal), mouse movement (spatial) and IMU sensor data.
Ensemble-classifier-based classification and sequential-sampling-analysis-based decision making.
High-confidence, low-latency, non-intrusive, and continuous authentication on the fly.

Indoor Person Localization Based on Wi-Fi Signal:

Developed CNN-based real-time person detection and localization using Wi-Fi signals.

Machine Learning and Neuromorphic Computing
My research seeks to discover the characteristics of fundamental bio-inspired cognitive process for next-generation AI systems, investigate neural networks’ structural plasticity/adaptability, evaluate the most popular deep learning approaches as well as reinforcement learning on selected applications.

Object Recognition for Autonomous Driving Using Spiking Neural Networks

Developed multi-layer SNN and spiking CNN with temporal coding for object recognition.
Tested on LiDAR (KITTI) and dynamic visual sensor data with 96.62% and 99.52% accuracy.

ECG Based Person Identification System and Personalized Heart Wave Generation:

Designed comprehensive wavelet distance based authentication with replay attack detection.
Implemented black box lifelike fake ECG templates generation using GAN and VAE, achieving 99% and 95% success rate on attacking the ECG authentication system.

Object Tracking in Autonomous Systems:

Developed object trackers based on particle filter and different neural networks: stacked denoising autoencoder (SDAE), convolutional neural network (CNN) and echo state network (ESN).
Benchmarked the tracking performance and algorithm efficiency on object tracking dataset.

Evolutionary Neural Networks Based on Reservoir Computing:

Investigated the feasibility and potential of multi-reservoir network.
Explored bio-inspired evolutionary neural networks based on reservoir computing paradigm.
Proposed dynamic multi-reservoir network leveraging structural evolution and synaptic plasticity.

Brain Activity Modeling Using Recurrent Network:

Fitted EEG/ERP signals with echo state network (ESN) - like recurrent neural networks.
Investigated the links between biological and artificial neural networks, explore related structural characteristics in neural networks based on correlation analysis of fitted models.

Publications

[P1] Yang Gao, Wei Wang, Vir V Phoha, Zhanpeng Jin. “EarEcho: Using Ear Canal Echo for Wearable Authentication.” Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 3, no. 3 (2019): 81.

[P2] Wei Wang, Yang Gao and Zhanpeng Jin, “Interpretive Reservoir: A Preliminary Study on the Association Between Artificial Neural Network and Biological Neural Network,” 2018 International Joint Conference on Neural Networks (IJCNN). IEEE, 2018.

[P3] Yang Gao, Wei Wang, Borui Li and Zhanpeng Jin, “Replicating Your Heart: Exploring Presentation Attacks on ECG Biometrics Using Synthesized Heart Waves,” 2018 IEEE Conference on Communications and Network Security (CNS). IEEE, 2018.

[P4] Borui Li, Wei Wang, Yang Gao, Vir Phoha and Zhanpeng Jin, “Hand in Motion: Enhanced Authentication Through Wrist and Mouse Movement,” The 9th IEEE International Conference on Biometrics: Theory, Applications, and Systems (BTAS). IEEE, 2018.

[P5] Omkar Patil, Wei Wang, Yang Gao and Zhanpeng Jin, “A Non-Contact PPG Biometric System Based on Deep Neural Network,” The 9th IEEE International Conference on Biometrics: Theory, Applications, and Systems (BTAS). IEEE, 2018.

[P6] Omkar R. Patil, Wei Wang, Yang Gao, Wenyao Xu, Zhanpeng Jin. “A low-cost, camera-based continuous PPG monitoring system using Laplacian pyramid.” Smart Health 9 (2018): 2-11.

[P7] Yang Gao, Borui Li, Wei Wang, Wenyao Xu, Chi Zhou, Zanpeng Jin. “Watching and Safeguarding Your 3D Printer: Online Process Monitoring Against Cyber-Physical Attacks.” Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 2.3 (2018): 108.

[P8] Wei Yang, Wei Wang and Zhanpeng Jin, “A Comparative Study of Object Tracking using CNN and SDAE,” 2018 International Joint Conference on Neural Networks (IJCNN). IEEE, 2018.

[P9] Wei Yang, Wei Wang and Zhanpeng Jin, “An Embedded Tracking System with Neural Network Accelerator,” 2018 International Joint Conference on Neural Networks (IJCNN). IEEE, 2018.

[P10] Xiaoliang Wang, Wei Wang and Zhanpeng Jin, “Context-Aware Reinforcement Learning-Based Mobile Cloud Computing for Telemonitoring,” 2018 IEEE EMBS International Conference on Biomedical & Health Informatics (BHI). IEEE, 2018.

[P11] Wei Wang, Hsiao-Tien Fan and Zhanpeng Jin, “Structure optimization of dynamic reservoir ensemble using genetic algorithm,” 2017 International Joint Conference on Neural Networks (IJCNN). IEEE, 2017.

[P12] Hsiao-Tien Fan, Wei Wang and Zhanpeng Jin, “Performance optimization of echo state networks through principal neuron reinforcement,” 2017 International Joint Conference on Neural Networks (IJCNN). IEEE, 2017.

[P13] Omkar Patil, Wei Wang, Yang Gao and Zhanpeng Jin. “A Camera-Based Pulse Transit Time Approach for Non-Intrusive Blood Pressure Monitoring.” Submitted to the 7th IEEE Int. Conf. Healthcare Informatics (ICHI’19). [under review]

[P14] Wei Wang, Shibo Zhou, Jingxi Li, Xiaohua Li, Junsong Yuan and Zhanpeng Jin. “Temporal Pulses Driven Spiking Neural Network for Fast Object Recognition in Autonomous Driving.” Submitted to the 34th AAAI Conference on Artificial Intelligence. [under review]

[P15] Jiyang Li, Wei Wang, Kratika Bhagtani, Yincheng Jin, Zhanpeng Jin. “A Delicate EEG-guided Music Recommendation System Understands Your Versatile Emotion” Submitted to the 2020 ACM CHI Conference on Human Factors in Computing Systems (SIGCHI’ 20). [under review]