Interference-aware Edge Runtime Prediction with Conformal Matrix Completion

Tianshu Huang, Arjun Ramesh, Emily Ruppel, Anthony Rowe, Carlee Joe-Wong

Conference on Machine Learning and Systems 2025 · Day 2 · Session 4: Reliable and Scalable Systems

This talk, "Interference-aware Edge Runtime Prediction with Conformal Matrix Completion," presented by Tianshu Huang, delves into the critical and complex problem of accurately predicting software workload execution times on novel hardware platforms, particularly within the challenging domain of edge and cyber-physical systems. The core innovation lies in framing this prediction challenge as a **matrix completion problem**, augmented with **conformal prediction** to provide statistical bounds on runtime, and specifically designed to account for **interference effects** when multiple tasks run concurrently.

AI review

Genuinely interesting framing of a hard problem — runtime prediction on heterogeneous edge hardware as matrix completion with conformal bounds — and the research appears real and carefully executed. But the article leaves too many implementation gaps to be actionable for engineers who want to build something similar, and the 'novelty vs. cloud prior work' story is undersold. Worth watching if you're in this exact domain; not must-see for the broader ML systems audience.