VAL: Interactive Task Learning with GPT Dialog Parsing

TL;DR

VAL is an interactive task learning system that combines large language models and symbolic methods to enable human-interpretable, incremental learning of hierarchical tasks from natural language, demonstrated effectively in a video game setting.

Contribution

The paper introduces VAL, a novel ITL system integrating LLMs with symbolic algorithms for interpretable, incremental learning of hierarchical tasks from natural language.

Findings

Most users successfully taught VAL using natural language.

VAL's knowledge generalizes to new tasks without retraining.

The system improves usability of interactive task learning.

Abstract

Machine learning often requires millions of examples to produce static, black-box models. In contrast, interactive task learning (ITL) emphasizes incremental knowledge acquisition from limited instruction provided by humans in modalities such as natural language. However, ITL systems often suffer from brittle, error-prone language parsing, which limits their usability. Large language models (LLMs) are resistant to brittleness but are not interpretable and cannot learn incrementally. We present VAL, an ITL system with a new philosophy for LLM/symbolic integration. By using LLMs only for specific tasks--such as predicate and argument selection--within an algorithmic framework, VAL reaps the benefits of LLMs to support interactive learning of hierarchical task knowledge from natural language. Acquired knowledge is human interpretable and generalizes to support execution of novel tasks without additional training. We studied users' interactions with VAL in a video game setting, finding that most users could successfully teach VAL using language they felt was natural.