v293: Proceedings of AutoML 2025

2026-06-04 · Source: Proceedings of Machine Learning Research · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Automated Machine Learning · Depth: Expert, short

Summary

Volume 293 presents the proceedings of the Fourth International Conference on Automated Machine Learning (AutoML), held from September 8-11, 2025, at Cornell Tech in New York. This collection features 25 papers exploring diverse advancements in AutoML. Key research areas include optimizing machine learning workflows using LLM agents, such as PiML and AutoPDL, and various approaches to hyperparameter optimization, including memory-efficient methods like "Frozen Layers" and cost-aware prompt optimization (CAPO). Papers also detail innovations in Neural Architecture Search (NAS) through techniques like Iterative Monte Carlo Tree Search and evolutionary algorithms (EG-ENAS), alongside advancements in Bayesian optimization, model ensembling for time series forecasting, and classifier probability calibration with SmartCal. The volume also introduces new benchmarking suites like CATBench and platforms for adaptive experimentation such as Ax, covering applications from generative AI to medical image segmentation.

Key takeaway

For AI Scientists and Machine Learning Engineers seeking to enhance model development efficiency, these proceedings highlight critical advancements in AutoML. You should investigate new methods for hyperparameter optimization, such as "Frozen Layers" for memory efficiency or CAPO for cost-aware prompt tuning, to improve your model training pipelines. Consider exploring LLM-agent-driven workflow automation tools like PiML or platforms like Ax for adaptive experimentation to streamline your development cycles and achieve better performance.

Key insights

AutoML research focuses on enhancing efficiency, automation, and performance across diverse ML tasks.

Principles

• Automate ML workflows using intelligent agents.
• Optimize hyperparameters for efficiency and performance.
• Explore diverse architectures via Neural Architecture Search.

In practice

• Utilize Ax for adaptive experimentation.
• Apply EG-ENAS for image classification NAS.
• Implement PiML for LLM-driven workflow optimization.

Topics

• Automated Machine Learning
• Hyperparameter Optimization
• Neural Architecture Search
• LLM Agents
• Bayesian Optimization
• Time Series Forecasting
• Medical Image Segmentation

Code references

• datarobot/syftr
• ankilab/EG-ENAS
• automl/ConfigurableOptimizer
• automl/AutoNNUnet

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

Related on AIssential

• Import AI 439: AI kernels; decentralized training; and universal representations — AI systems can automate and accelerate complex software optimization tasks, yielding significant performance gains.
• Explainable Iterative Data Visualisation Refinement via an LLM Agent — An LLM-driven agent automates and explains data visualization refinement by iteratively optimizing DR hyperparameters.
• How Autoresearch will change Small Language Models adoption — Autoresearch enables autonomous, iterative optimization of ML models by an AI agent editing training code.
• 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.
• Capacity Efficiency at Meta: How Unified AI Agents Optimize Performance at Hyperscale — Meta's AI agent platform automates performance optimization and regression fixing at hyperscale, encoding expert knowledge for efficiency.
• Feedback-Driven Tool-Use Improvements in Large Language Models via Automated Build Environments — Automated environment generation and verifiable rewards significantly boost LLM tool-use performance via reinforcement learning.

Open in AIssential →

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