Learning to solve arithmetic problems with a virtual abacus
This work addresses the problem of acquiring mathematical skills in AI, which is a key challenge for the field, though it appears incremental as it builds on existing reinforcement learning methods applied to a specific domain.
The paper tackled the challenge of teaching AI systems mathematical skills by developing a deep reinforcement learning framework that simulates learning to solve arithmetic problems using a virtual abacus. The model achieved an error rate below 1% on multi-digit additions and subtractions, even with operands longer than those in training.
Acquiring mathematical skills is considered a key challenge for modern Artificial Intelligence systems. Inspired by the way humans discover numerical knowledge, here we introduce a deep reinforcement learning framework that allows to simulate how cognitive agents could gradually learn to solve arithmetic problems by interacting with a virtual abacus. The proposed model successfully learn to perform multi-digit additions and subtractions, achieving an error rate below 1% even when operands are much longer than those observed during training. We also compare the performance of learning agents receiving a different amount of explicit supervision, and we analyze the most common error patterns to better understand the limitations and biases resulting from our design choices.