Skip to content

SELA

Explore AutoML pipelines with tree search

Visit official site

What SELA is

The framework is described as a system that combines Monte Carlo Tree Search with LLM agents for AutoML.

What you can do with it

  • Represent a pipeline as a tree
  • Explore the solution space with MCTS
  • Refine strategy from experiment feedback

Application scenarios

  • Compare search strategies in an ablation study
  • Prepare a dataset from a link
  • Adjust rollouts and timeout values

Pros & cons

✓ What we like

  • Data and LLM configurations can be modified
  • Rollouts and timeout parameters are adjustable

! What to watch out for

  • Pipelines are explored as tree configurations
  • Experiments need dataset preparation and dependency installation

FAQ

What search modes are listed?

The material lists MCTS and random search (RS) experiment modes.

Can a run be resumed?

The material states an existing MCTS tree can be loaded after an interruption.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review