Teaching Multiple Inverse Reinforcement Learners

In this paper, we propose the first machine teaching algorithm for multiple inverse reinforcement learners. As our initial contribution, we formalize the problem of optimally teaching a sequential task to a heterogeneous class of learners. We then contribute a theoretical analysis of such problem, identifying conditions under which it is possible to conduct such teaching using the same demonstration for all learners. Our analysis shows that, contrary to other teaching problems, teaching a sequential task to a heterogeneous class of learners with a single demonstration may not be possible, as the differences between individual agents increase. We then contribute two algorithms that address the main difficulties identified by our theoretical analysis. The first algorithm, which we dub SplitTeach, starts by teaching the class as a whole until all students have learned all that they can learn as a group; it then teaches each student individually, ensuring that all students are able to perfectly acquire the target task. The second approach, which we dub JointTeach, selects a single demonstration to be provided to the whole class so that all students learn the target task as well as a single demonstration allows. While SplitTeach ensures optimal teaching at the cost of a bigger teaching effort, JointTeach ensures minimal effort, although the learners are not guaranteed to perfectly recover the target task. We conclude by illustrating our methods in several simulation domains. The simulation results agree with our theoretical findings, showcasing that indeed class teaching is not possible in the presence of heterogeneous students. At the same time, they also illustrate the main properties of our proposed algorithms: in all domains, SplitTeach guarantees perfect teaching and, in terms of teaching effort, is always at least as good as individualized teaching (often better); on the other hand, JointTeach attains minimal teaching effort in all domains, even if sometimes it compromises the teaching performance.

Eye Tracking Using Nonverbal Tasks Could Contribute to Diagnostics of Developmental Dyslexia and Developmental Language Disorder

There are not many studies dealing with a comparison of the eye movements of individuals with dyslexia and developmental language disorder (DLD). The aim of this study is to compare the eye movements in the two most common language disorders, dyslexia and DLD and to consider their contribution to diagnostics. In the research the oculomotor test was administered to 60 children with the clinical diagnosis of dyslexia or DLD and 58 typically developing children (controls). The test included a prosaccadic task, antisaccadic task and a nonverbal sequential task with self-regulation of the pace. Controls could be singled out from other two clinical groups by means of the oculomotor imaging. Both of the clinical groups in comparison with the controls were characterized by worse overall performance. Through the employment of the oculomotor it was possible to differentiate between both of the clinical groups. The dyslexics had an overall worse oculomotor performance than the DLD group. The results of the study show that the oculomotor test has the potential to contribute to diagnostics of dyslexia and DLD and the screening of these disorders at pre-school age.

Angle-Awareness Based Joint Cooperative Positioning and Warning for Intelligent Transportation Systems

In future intelligent vehicle-infrastructure cooperation frameworks, accurate self-positioning is an important prerequisite for better driving environment evaluation (e.g., traffic safety and traffic efficiency). We herein describe a joint cooperative positioning and warning (JCPW) system based on angle information. In this system, we first design the sequential task allocation of cooperative positioning (CP) warning and the related frame format of the positioning packet. With the cooperation of RSUs, multiple groups of the two-dimensional angle-of-departure (AOD) are estimated and then transformed into the vehicle’s positions. Considering the system computational efficiency, a novel AOD estimation algorithm based on a truncated signal subspace is proposed, which can avoid the eigen decomposition and exhaustive spectrum searching; and a distance based weighting strategy is also utilized to fuse multiple independent estimations. Numerical simulations prove that the proposed method can be a better alternative to achieve sub-lane level positioning if considering the accuracy and computational complexity.