VAL: Interactive Task Learning with GPT Dialog Parsing
This addresses the usability issue in ITL systems for users who need to teach machines tasks incrementally with limited instruction, though it is incremental as it builds on existing LLM and symbolic integration approaches.
The paper tackled the problem of brittle language parsing in interactive task learning (ITL) systems by introducing VAL, which integrates large language models (LLMs) for specific parsing tasks within a symbolic framework, enabling incremental learning of hierarchical task knowledge from natural language. In a video game study, most users successfully taught VAL using natural language, and the acquired knowledge was human-interpretable and generalized to novel tasks without additional training.
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.