$L^*LM$: Learning Automata from Examples using Natural Language Oracles
This work addresses the challenge of efficiently extracting formal specifications from demonstrations for tasks in AI and programming, representing an incremental improvement by combining existing techniques with natural language.
The paper tackles the problem of learning deterministic finite automata (DFA) from expert demonstrations, which is often not sample efficient, by introducing L*LM, an algorithm that uses both demonstrations and natural language to improve data efficiency, resulting in significant gains in few-shot learning.
Expert demonstrations have proven an easy way to indirectly specify complex tasks. Recent algorithms even support extracting unambiguous formal specifications, e.g. deterministic finite automata (DFA), from demonstrations. Unfortunately, these techniques are generally not sample efficient. In this work, we introduce $L^*LM$, an algorithm for learning DFAs from both demonstrations and natural language. Due to the expressivity of natural language, we observe a significant improvement in the data efficiency of learning DFAs from expert demonstrations. Technically, $L^*LM$ leverages large language models to answer membership queries about the underlying task. This is then combined with recent techniques for transforming learning from demonstrations into a sequence of labeled example learning problems. In our experiments, we observe the two modalities complement each other, yielding a powerful few-shot learner.