Learning From Natural Instructions
Dan Goldwasser and Dan Roth
In this paper we suggest to view learning a decision function as a natural language lesson interpretation problem instead of learning from labeled examples as done traditionally. This interpretation of machine learning is motivated by human learning processes, in which the learner is given a lesson describing the target concept directly and a few instances exemplifying it. We introduce a learning algorithm for the lesson interpretation problem that relies on lesson performance as feedback, and learns both tasks jointly. This approach alleviates the supervision burden of traditional machine learning by focusing on human-level task expertise for learning. We evaluate our approach by applying it to the rules of the Freecell solitaire card game. We show that our learning approach can learn the target concept and play the game. Furthermore, we show that the learned semantic interpreter also generalizes to previously unseen instructions.