Yongkang Du
duyongka[at]gmail[dot]com
Yongkang Du is a second-year PhD student at Penn State University, College of IST, advised by Prof. Lu Lin. He received his M.S. in computer science from University of Southern California advised by Prof. Jieyu Zhao and B.E. in computer science and technology from North China University of Technology. His cv is available here.
His research is driven by building capable and reliable AI systems. Specifically, two fundamental questions:
- How can we bridge the gap between human and AI reasoning? Investigating advanced cognitive capabilities in multimodal and agentic systems, including compositional logic, analogical reasoning, and long-horizon memory.
- How can we ensure the trustworthiness of AI in the real world? Auditing and optimizing models for fairness, alignment, and safety, particularly in the context of decision-making.
Open to collaboration and internship opportunities!
News
| Jun 2026 | Two papers accepted to the ICML Workshop on Forecasting! From Narrative to Auditable Forecasts: A Structured Scaffold for Agentic Forecasting and ForecastCompass: Guiding Agentic Forecasting with Adaptive Factor Memory (🌟 Spotlight) |
|---|---|
| May 2026 | One paper accepted to the ACL TrustNLP workshop! Controllable Pareto Trade-off between Fairness and Accuracy |
| Mar 2026 | New paper out! CARV: A Diagnostic Benchmark for Compositional Analogical Reasoning in Multimodal LLMs. |
Selected publications
- COLM’26CARV: A Diagnostic Benchmark for Compositional Analogical Reasoning in Multimodal LLMsCOLM, 2026Under Review
- ICML’26ForecastCompass: Guiding Agentic Forecasting with Adaptive Factor MemoryICML Workshop on Forecasting, 2026Spotlight
- ICML’26From Narrative to Auditable Forecasts: A Structured Scaffold for Agentic ForecastingICML Workshop on Forecasting, 2026
- KDD’26FairCoder: Probing LLM Bias in High-Stakes Decision Making via Coding TasksKDD, 2026Under Review
- ACL’26Controllable Pareto Trade-off between Fairness and AccuracyACL TrustNLP Workshop, 2026
- EMNLP