Title : Improving RL Power for On-Line Evolution of Gaits in Modular Robots
| Presenter | Milan Jelisavcic |
| Abstract | This presentation addresses the problem of on-line gait learning in modular robots whose shape is not known in advance. The best algorithm for this problem known to us is a reinforcement learning method, called RL Power. In our study we revisited the original RL Power algorithm and observed that in essence it is a specific evolutionary algorithm. Based on this insight we propose two modifications of the main search operators and compare the quality of the evolved gaits when either or both of these modified operators are employed. The results show that using mutation with self-adaptive step-sizes can significantly improve the performance of the original algorithm. |
Title : Capturing the Ineffable: Collecting, Analysing, and Automating Web Document Quality Assessments
| Presenter | Davide Ceolin |
| Abstract | Automatic estimation of the quality of Web documents is a challenging task, especially because the definition of quality heavily depends on the individuals who define it, on the context where it applies, and on the nature of the tasks at hand. In this talk, I will present the current status of the ADS project QuPiD, which aims at automating the process of quality assessment of Web documents. In particular, I will illustrate two use cases we recently run with the aim of nichesourcing app we developed to collect quality assessments. These user studies involved Journalism students and Media scholars and allowed us to shed a light both on how these assessments are characterized and on how they relate to documents features. |