Towards AutoAI @ ICML 2024

Optimizing an ML System with Black-box and Differentiable Components

Towards AutoAI @ ICML 2024


The AutoAI problem is about the optimization of a complex ML system (e.g., LLM prompt engineering pipeline, healthcare system) with many differentiable ML components & black-box components.



AutoAI presents a paradigm shift!


Complex ML systems are challenging to optimize --- they are very high-dimensional and have no analytical form.



For more details, watch our ICML 2024 video presentation below:


Or catch Zhiliang in action during his poster presentation at ICML 2024: