@inproceedings{Amado2022b,
author = {Leonardo R. Amado and Reuth Mirsky and Felipe Meneguzzi},
title = {{Learning Goal Recognition Models using Human Examples}},
booktitle = {2022 Workshop on Human-Interactive Robot Learning (HIRL@HRI)},
year = {2022},
abstract = {Most approaches for goal recognition rely on specifications of the possible dynamics of the actor in the environment when pursuing a goal. However, encoding these dynamics requires careful design by a human expert, a design which is often not robust to noise at recognition time. 
In this paper, we present our recent framework that combines learning and goal recognition to alleviate the need for careful, manual domain design. This framework consists of two main stages: Offline learning of policies for each potential goal, and online inference. 
In this short paper, we focus on the first stage of learning the needed policies, and propose an approach to use behavioral cloning to elicit these policies. The aim of this extended abstract is to share the new goal recognition framework with the Human-Interactive Robot Learning (HIRL) community and obtain feedback on the best practices to promoting its implementation using human examples.}
}