Learning Program Synthesis for Integer Sequences from Scratch

Abstract

We present a self-learning approach for synthesizing programs from integer sequences. Our method relies on a tree search guided by a learned policy. Our system is tested on the On-Line Encyclopedia of Integer Sequences. There, it discovers, on its own, solutions for 27987 sequences starting from basic operators and without human-written training examples.

Cite

Text

Gauthier and Urban. "Learning Program Synthesis for Integer Sequences from Scratch." AAAI Conference on Artificial Intelligence, 2023. doi:10.1609/AAAI.V37I6.25930

Markdown

[Gauthier and Urban. "Learning Program Synthesis for Integer Sequences from Scratch." AAAI Conference on Artificial Intelligence, 2023.](https://mlanthology.org/aaai/2023/gauthier2023aaai-learning/) doi:10.1609/AAAI.V37I6.25930

BibTeX

@inproceedings{gauthier2023aaai-learning,
  title     = {{Learning Program Synthesis for Integer Sequences from Scratch}},
  author    = {Gauthier, Thibault and Urban, Josef},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2023},
  pages     = {7670-7677},
  doi       = {10.1609/AAAI.V37I6.25930},
  url       = {https://mlanthology.org/aaai/2023/gauthier2023aaai-learning/}
}