Structured Apprenticeship Learning
نویسندگان
چکیده
We propose a graph-based algorithm for apprenticeship learning when the reward features are noisy. Previous apprenticeship learning techniques learn a reward function by using only local state features. This can be a limitation in practice, as often some features are misspecified or subject to measurement noise. Our graphical framework, inspired from the work on Markov Random Fields, allows to alleviate this problem by propagating information between states, and rewarding policies that choose similar actions in adjacent states. We demonstrate the advantage of the proposed approach on grid-world navigation problems, and on the problem of teaching a robot to grasp novel objects in simulation.
منابع مشابه
Comparison of Students’ Perception of Preparedness for Interprofessional learning readiness in apprenticeship and apprenticeship on site in Schools of Nursing and Midwifery of Islamic Azad Universities in Isfahan, Iran in 2018
Background & Objective: Interprofessional education (IPE) is one of the new approaches in the education of students in health-related disciplines. This type of training can increase interprofessional collaborations, thereby improving patient care quality. This study aimed to compare the perception of IPE in students apprenticeship and apprenticeship on site in schools of nursing and midwifery o...
متن کاملBoosted and reward-regularized classification for apprenticeship learning
This paper deals with the problem of learning from demonstrations, where an agent called the apprentice tries to learn a behavior from demonstrations of another agent called the expert. To address this problem, we place ourselves into the Markov Decision Process (MDP) framework, which is well suited for sequential decision making problems. A way to tackle this problem is to reduce it to classif...
متن کاملSituating Learning in the Workplace: Having Another Look at Apprenticeships
This article examines the acquisition of vocational skills through apprenticeship-type situated learning. Findings from a studies of skilled workers revealed that learning processes that were consonant with the apprenticeship model of learning were highly valued as a means of acquiring and maintaining vocational skills. Supported by current research and theorising, this article, describes some ...
متن کاملGeneralizing Apprenticeship Learning across Hypothesis Classes
This paper develops a generalized apprenticeship learning protocol for reinforcementlearning agents with access to a teacher who provides policy traces (transition and reward observations). We characterize sufficient conditions of the underlying models for efficient apprenticeship learning and link this criteria to two established learnability classes (KWIK and Mistake Bound). We then construct...
متن کاملTask sharing in rural Haiti: Qualitative assessment of a brief, structured training with and without apprenticeship supervision for community health workers.
Despite growing support for supervision after task sharing trainings in humanitarian settings, there is limited research on the experience of trainees in apprenticeship and other supervision approaches. Studying apprenticeships from trainees' perspectives is crucial to refine supervision and enhance motivation for service implementation. The authors implemented a multi-stage, transcultural adap...
متن کامل