v256: Proceedings of AutoML 2024

2026-06-10 · Source: Proceedings of Machine Learning Research · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics · Depth: Expert, short

Summary

Volume 256 compiles the proceedings of the Third International Conference on Automated Machine Learning (AutoML), held from 9-12 September 2024 at Sorbonne Université, Paris, France. The research presented spans critical advancements in AutoML methodologies, including techniques for speeding up and analyzing Neural Architecture Search (NAS), such as adaptive subset selection, few-shot NAS, and evolutionary or gradient-based approaches. Significant contributions are also made in Hyperparameter Optimization (HPO), with methods for unsupervised outlier detection, reinforcement learning, and asynchronous multi-fidelity optimization. Furthermore, papers address model efficiency through mixed-precision quantization search (FLIQS), early stopping cross-validation, and memory-limited training for ML pipelines. The proceedings also highlight practical applications and tools like AutoGluon-Multimodal for foundation models, ASML for data streams, HoNCAML for no-code AutoML, and TabRepo for tabular model evaluations, alongside exploring capabilities such as Mamba's "in-context learning" and automated deep learning for load forecasting.

Key takeaway

Volume 256 of the AutoML conference presents significant advancements in automating machine learning, focusing on optimizing Neural Architecture Search (NAS) and Hyperparameter Optimization (HPO). Papers introduce techniques like adaptive subset selection for faster NAS, HPOD for unsupervised outlier detection, and FLIQS for one-shot mixed-precision quantization, alongside integrating foundation models for multimodal AutoML. This collection provides critical tools and insights for ML engineers and researchers to accelerate model development, enhance efficiency, and deploy robust, scalable systems in diverse and resource-constrained environments.

Topics

• Neural Architecture Search
• Hyperparameter Optimization
• In-Context Learning
• Model Quantization
• Multimodal AutoML

Code references

• review2024/HPOD
• Data-Science-Eurecat/HoNCAML
• autogluon/tabrepo

Best for: Research Scientist, AI Scientist, Machine Learning Engineer, AI Engineer

Related on AIssential

• v224: Proceedings of AutoML 2023 — Volume 224 of the Proceedings of the Second International Conference on Automated Machine Learning (AutoML), held from 12-15 November 2023 at the Hasso Plattner Institute in Potsdam, Germany, compiles…
• v188: Proceedings of AutoML-Conf 2022 — Volume 188 compiles the proceedings of the First International Conference on Automated Machine Learning (AutoML), held from July 25-27, 2022, at Johns Hopkins University in Baltimore, MD, USA, with ed…
• Auto-FP: An Experimental Study of Automated Feature Preprocessing for Tabular Data — Automating feature preprocessing for tabular data can be framed as HPO or NAS, with evolution-based algorithms performing best.
• AutoAdapt: Automated domain adaptation for large language models — AutoAdapt automates LLM domain adaptation, transforming a manual, costly process into a reproducible, constraint-aware engineering discipline.
• A Reproducible Log-Driven AutoML Framework for Interpretable Pipeline Optimization in Healthcare Risk Prediction — Reproducible, log-driven AutoML reveals healthcare risk prediction pipeline component importance and search space redundancy.
• Simplifying Image Classification with Azure AutoML for Images: A Practical Guide — Azure AutoML for Images automates complex computer vision tasks, enabling rapid deployment of production-ready models.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Proceedings of Machine Learning Research.