Explore a groundbreaking video presentation from PLDI 2024 that introduces a novel program analysis technique for adaptive data analysis. Delve into the challenges of controlling generalization errors in complex, multi-query data analyses and discover how this innovative approach formalizes adaptivity as a quantitative property of programs. Learn about the implementation of a weighted dependency graph representation and path search strategy to establish upper bounds on adaptivity. Gain insights into how this analysis can be applied to various data analyses with different adaptivity structures, potentially revolutionizing the way researchers approach and optimize adaptive data analysis techniques. Access supplementary materials, including the full article and reusable artifacts, to further enhance your understanding of this cutting-edge research in program analysis and data science.
Overview
Syllabus
[PLDI24] Program Analysis for Adaptive Data Analysis
Taught by
ACM SIGPLAN