Towards Efficient Neurally-Guided Program Induction for ARC-AGI
This addresses program induction efficiency and generalization for ARC-AGI, an open-world AI problem, but appears incremental in exploring existing paradigms.
The paper investigated neurally-guided program induction approaches for the ARC-AGI domain, finding that learning the program space was most effective for submission, while suggesting learning the transform space as a potential solution with preliminary experiments.
ARC-AGI is an open-world problem domain in which the ability to generalize out-of-distribution is a crucial quality. Under the program induction paradigm, we present a series of experiments that reveal the efficiency and generalization characteristics of various neurally-guided program induction approaches. The three paradigms we consider are Learning the grid space, Learning the program space, and Learning the transform space. We implement and experiment thoroughly on the first two, and retain the second one for ARC-AGI submission. After identifying the strengths and weaknesses of both of these approaches, we suggest the third as a potential solution, and run preliminary experiments.