Towards Multi-Agent Communication-Based Language Learning
This addresses the challenge of enabling AI agents to learn language from scratch through interaction, though it is incremental as it builds on existing multi-agent and referential game research.
The paper tackles the problem of language learning by proposing a multi-agent communication framework where neural networks develop their own language through cooperative referential games, with preliminary experiments showing promising results but highlighting risks of adhoc communication codes.
We propose an interactive multimodal framework for language learning. Instead of being passively exposed to large amounts of natural text, our learners (implemented as feed-forward neural networks) engage in cooperative referential games starting from a tabula rasa setup, and thus develop their own language from the need to communicate in order to succeed at the game. Preliminary experiments provide promising results, but also suggest that it is important to ensure that agents trained in this way do not develop an adhoc communication code only effective for the game they are playing