Yue Zhao
Avatar of Yue Zhao
Assistant Professor
Thomas Lord Department of Computer Science

University of Southern California

Los Angeles, CA, USA
Email:

Collaboration with Me. I am open to external opportunities for invited talks, research collaborations, and employment (on the part-time/advising/visiting basis). Let us have a chat via email. I frequently visit major cities, e.g., Seattle, NYC, Chicago, Boston, and Bay Area to meet people and give in-person talks.

Lab Openings. We are recruiting 2 Ph.D. students for Fall 2026. I only admit students with prior collaboration, so please reach out as early as possible. Applicants should have at least one paper published in a top ML, System, or LLM conference.
We also welcome undergraduate and graduate interns from USC and other institutions any time, all year around, preferably in North America for time zone compatibility. To apply, please complete this Google Form: Application Form and email me after submitting the form. For more information, please review the FORTIS Lab website before reaching out.

Research Interests: My research focuses on building trustworthy, knowledge-driven, and generative AI systems to address real-world challenges. By integrating robustness, graph learning, generative AI, and scalable AI systems, my work ensures reliability, automates processes, and drives innovation across domains. I also create open-source tools and frameworks, such as PyOD, to accelerate efficient and automated AI adoption in fields like finance, healthcare, and AI4Science.

  1. Robust and Trustworthy AI Across Domains: Developing reliable AI systems that detect outliers, anomalies, and out-of-distribution (OOD) data to ensure trust, fairness, and transparency across diverse domains, including finance, security, and healthcare.
    Keywords: OOD Detection, Outlier Detection, Anomaly Detection, Fairness, Trustworthiness
  2. Graph Learning and Structured Knowledge for Decision-Making: Applying graph-based models to analyze interconnected data, enabling tasks like OOD detection, neural architecture search (NAS), and anomaly detection on graphs. These methods power AI applications in healthcare, financial risk modeling, and molecular science.
    Keywords: GNNs, Graph Open Set Learning, Anomaly Detection on Graphs, Graph-based Knowledge Discovery
  3. Generative AI and Foundation Models for AI for Science (AI4Science): Leveraging generative AI, LLMs, and foundation models to address scientific and societal challenges. Applications include synthetic clinical trials, drug discovery, and political forecasting, driving breakthroughs in AI4Science.
    Keywords: LLMs, Foundation Models, AI4Science, Drug Discovery, LLMs for Political Science
  4. Scalable and Open-Source AI Systems: Building scalable and accelerated AI tools to automate tasks like model selection, hyperparameter optimization, and anomaly detection. As the creator of PyOD (25M+ downloads, used by NASA, Tesla, etc.), I lead 10+ open-source ML projects, including PyGOD, TDC, and ADBench. Collectively, these projects have earned 20,000+ GitHub stars (top 0.002%), accelerating AI adoption and impact.
    Keywords: Automated ML, GPU Acceleration, Open-source AI, Scalable AI Systems

Biography.

Copy/paste from here :)


✈ News and Travel

[Nov 2024] We have a new paper on dynamic prototype updating for multimodal out-of-distribution detection; see our Preprint!

[Nov 2024] We have a new paper on data augmentation for anomaly detection accepted to IEEE TNNLS; see our Preprint!

[Oct 2024] We have a new paper on label-efficient graph learning for OOD; see our Preprint!

[Oct 2024] Received Capital One Research Awards with Prof. Jieyu Zhao responsible AI for Finance!

[Oct 2024] We have a new paper on AI x Protein (Spotlight at NeurIPS Workshop on AI for New Drug Modalities): see our Preprint!

[Oct 2024] We have a new paper on automated OOD model selection (the best paper at KDD Resource Efficient Learning Workshop): see our Preprint!

[Sep 2024] MultiOOD: Scaling Out-of-Distribution Detection for Multiple Modalities, led by Hao Dong @ ETHZ, will appear at NeurIPS 2024 as a spotlight paper!

🏅 Awards and Grants

As Principal Investigator (August 2023 onwards)
Prior to Principal Investigator Role (Before August 2023)