LACS 2026

Overview

Recent advances in machine learning—particularly foundation models, graph neural networks, and reinforcement learning—are transforming how systems are designed, optimized, and deployed.

Modern systems are no longer purely engineered using heuristics; instead, they increasingly incorporate learned decision-making components.

LACS (Learning-Augmented Compilers & Systems) aims to bring together researchers and practitioners working at the intersection of:

• Machine Learning
• Compilers and Programming Systems
• Computer Architecture and Hardware
• Large-scale Infrastructure and Distributed Systems

The workshop focuses on learning-augmented system design, where machine learning models are integrated into the core of system decision-making pipelines.

Venue & Date

Venue: Illini Center (University of Illinois)

Date: Monday, October 19 (afternoon)

Topics of Interest

Learning-Augmented Compilers

• ML-guided compiler optimizations (LLVM, MLIR, Halide, TVM, XLA)
• Learned cost models and scheduling
• Kernel selection and operator optimization
• ML-based profile-guided optimization (ML-PGO)
• Transfer learning across architectures

LLMs for Systems

• LLM-driven optimization and autotuning
• Code generation for systems
• Agent-based system debugging
• Performance modeling

Systems for ML / LLMs

• Efficient training and inference
• Serving systems (batching, caching)
• Distributed systems
• Memory optimization

Hardware–ML–Compiler Co-Design

• ML-driven hardware design
• Compiler–hardware co-optimization
• Cross-layer optimization

Submission Types

• Extended Abstract (2 pages)
• Position Papers (2–4 pages)

Important Dates

• Submission Deadline: August 14, 2026
• Author Notification: August 28, 2026

Organizers

• Eun Jung (EJ) Park, Qualcomm Inc.
• Riyadh Baghdadi, NYU Abu Dhabi & MIT
• Joseph Manzano, Pacific Northwest National Laboratory
• Keren Zhou, George Mason University & OpenAI

Program Committee

TBD