• Home
• Publications
• CV

Wenbo Zhang (张文博)

Welcome to my homepage!

Biography

I’m a 5th-year Ph.D. student in the College of Information Sciences and Technology at Pennsylvania State University, advised by Dr. Amulya Yadav. Prior to Penn State, I was an algorithm engineer in the AI lab at Kingsoft. While at Kingsoft, I primarily worked on projects that focused on natural language processing (NLP) and speech processing (especially speech synthesis). Previously, I got my M.S. degree in Electrical Engineering from the University of Southern California, under the supervision of Prof. Cauligi Raghavendra in 2018. I obtained the B.E. degree in Renewable Energy Materials and Devices at the University of Electronic Science and Technology of China in 2015.

I am broadly interested in multilingual and cross-lingual NLP, model alignment, AI for social impact, and speech processing. My Ph.D. research primarily focuses on: (1) adapt large language models (LLMs) to handle code-switching (also known as code-mixed languages); (2) work extensively at the intersection of AI and society to examine the sociotechnical challenges of LLMs such as bias and misinformation. Beyond these, I have collaborated on real-world AI for social impact projects, such as TRIM-AI, a multilingual NLP system deployed in Kenya to support maternal health through SMS-based tele-triage.

My research has successfully been deployed in the real-world with substantial impact on the ground. Since July 2022, an AI tool that I developed (TRIM-AI) has been in continuous deployment at Jacaranda Health, an NGO that aids the maternal care of over 1.25 millions pregnant mothers in Kenya.

Research Interests

• Multilingual and cross-lingual NLP
• Model Alignment
• Code-switching
• AI for Social Impact
• Data Mining
• Speech Processing

Recent News

We win the 2022 AI Societal Impact Award held by the Center for Artificial Intelligence Foundations and Engineered Systems (CAFÉ) at Pennsylvania State University.

Education

• Doctor of Philosophy in Informatics
• Advisor: Dr. Amulya Yadav

• Master of Science in Electrical Engineering
• Advisor: Dr. Cauligi Raghavendra

• Bachelor of Engineering in Renewable Energy Materials and Devices

Recent Highlights

A Continual Pre-training Approach to Tele-Triaging Pregnant Women in Kenya
Wenbo Zhang, Hangzhi Guo, Prerna Ranganathan, Jay Patel, Sathyanath Rajasekharan, Nidhi Danayak, Manan Gupta, Amulya Yadav. In 37th AAAI Conference on Artificial Intelligence (AAAI), 2023.
• Full-time deployed @ Jacaranda Health
PDF Abstract

×

A Continual Pre-training Approach to Tele-Triaging Pregnant Women in Kenya

Access to high-quality maternal health care services is limited in Kenya, which resulted in ~36,000 maternal and neonatal deaths in 2018. To tackle this challenge, our collaborating non-governmental organization (NGO) which works on maternal health in Kenya, developed PROMPTS, an SMS based tele-triage system for pregnant and puerperal women which has more than 350,000 active users in Kenya. PROMPTS empowers pregnant women living far away from doctors and hospitals to send SMS messages to get quick answers (through human helpdesk agents) to questions about their medical symptoms and pregnancy status. Unfortunately, ~1.1 million SMS messages are received by PROMPTS every month, which makes it challenging for helpdesk agents to ensure that these messages can be interpreted correctly and evaluated by their level of emergency to ensure timely responses and/or treatments for women in need. This paper reports on a collaborative effort with an NGO to develop a state-of-the-art natural language processing (NLP) framework, TRIM-AI (TRIage for Mothers using AI), which can automatically predict the emergency level (or severity of medical condition) of a pregnant mother based on the content of their SMS messages. TRIM-AI leverages recent advances in multi-lingual pre-training and continual pre-training to tackle code-mixed SMS messages (between English and Swahili), and achieves a weighted F_1 score of 0.774 on real-world datasets. TRIM-AI has been successfully deployed in the field since June 2022, and is being used by our collaborating NGO to prioritize the provision of services and care to pregnant women with the most critical medical conditions. Our preliminary A-B tests in the field show that TRIM-AI is ~17% more accurate at predicting high-risk medical conditions from SMS messages sent by pregnant Kenyan mothers, which reduces the helpdesk's workload by ~12%.

Industrial Experience

• Machine Learning Engineer | Kingsoft Co., Ltd.
  Topic: Natural Language Processing and Speech Processing
  Jan 2019 - Jul 2021
• Research Intern | Comprehend Information Technology Co., Ltd.
  Topic: Data Mining on traffic data
  Jun 2017 - Aug 2017

Thank you for visiting.