Welcome to autoBOT’s documentation

autoBOT is an autoML system for automatic explainable classification of texts. It includes automatically generated features at different granularities, evolved to offer optimal performance/explainability tradeoff.

About:

• Key idea underlying autoBOTLib

Tutorial and first steps:

• Installation
• autoBOTLib library
• Suggested use
• Model persistence
• autoBOTLib CLI
• Feature types
• Using custom classifiers
• Using custom feature transformers
  • Using custom features
  • Extending existing feature spaces
• Obtaining underlying representations
• Singularity container for autoBOT
• Using custom background knowledge
• Using custom metrics
• Potentially interesting further work
  • Extensions with contextual neural representations
  • Improving the classification phase
  • Speeding up evolution
  • Meta-transfer learning
• Releases
• autoBOT API
• Scaling up (SLING)
• Performance on shared tasks

Contents:

• autoBOTLib
  • autoBOTLib package
    • Subpackages
      • autoBOTLib.features package
      • autoBOTLib.learning package
      • autoBOTLib.misc package
      • autoBOTLib.optimization package
    • Submodules
    • autoBOTLib.cli module
    • Module contents