Jeff Dean, Chief Scientist at Google Deepmind, talks about how machine learning is changing the way we approach traditional computer systems problems. This talk, presented at NeurIPS 2024, covers practical applications of ML in areas like: Compiler optimization: Using learning techniques to make better compilation decisions. Memory allocation: Smarter ways to manage resources in real time. ASIC chip design: Applying ML across multiple stages of hardware design.
Jeff Dean discusses exciting trends in how machine learning can be used to tackle traditional computer systems problems. He provides an overview of the application of ML in several areas, including compiler optimization, memory allocation, learning-driven approaches to ASIC chip design, and techniques to integrate learned decisions into traditional software systems. This talk highlights the collaborative work of many contributors at Google.