AlphaD3M: An Open-Source AutoML Library for Multiple ML Tasks

Roque Lopez, Raoni Lourenço, Remi Rampin, Sonia Castelo, Aécio Santos, Jorge Ono, Claudio Silva, Juliana Freire

Research output: Contribution to journalConference articlepeer-review


We present AlphaD3M, an open-source Python library that supports a wide range of machine learning tasks over different data types. We discuss the challenges involved in supporting multiple tasks and howAlphaD3M addresses them by combining deep reinforcement learning and meta-learning to construct pipelines over a large collection of primitives effectively. To better integrate the use of AutoML within the data science lifecycle, we have built an ecosystem of tools around AlphaD3M that support user-in-the-loop tasks, including selecting suitable pipelines and developing custom solutions for complex problems. We present use cases that demonstrate some of these features. We report the results of a detailed experimental evaluation showing that AlphaD3M is effective and derives highquality pipelines for a diverse set of problems with performance comparable or superior to state-of-the-art AutoML systems.

Original languageEnglish (US)
JournalProceedings of Machine Learning Research
StatePublished - 2023
Event2nd International Conferenceon Automated Machine Learning, AutoML 2023 - Potsdam, Germany
Duration: Nov 12 2023Nov 15 2023

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software
  • Control and Systems Engineering
  • Statistics and Probability


Dive into the research topics of 'AlphaD3M: An Open-Source AutoML Library for Multiple ML Tasks'. Together they form a unique fingerprint.

Cite this