Learning Program Synthesis for Integer Sequences from Scratch
This addresses the challenge of automated program synthesis for integer sequences, which is incremental as it builds on existing methods but introduces self-learning from scratch.
The paper tackles the problem of synthesizing programs from integer sequences using a self-learning approach with tree search and a learned policy, achieving solutions for 27,987 sequences from the On-Line Encyclopedia of Integer Sequences without human training examples.
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.