Title : Essentials for Contextual Interpretation of Language
| Presenter | Filip Ilievski |
| Abstract | Despite the inherent ambiguity of natural language, humans have almost no difficulty communicating and resolving this ambiguity on-the-fly within a specific context. How could we teach machines to leverage context in order to deal with ambiguity of language? In my Minutes of Fame, I will discuss the essential ingredients for contextual interpretation of natural language: algorithms, contextual knowledge, evaluation data and evaluation metrics. I will look into the state-of-the-art disambiguation approaches in terms of their consideration of the listed ingredients. I will also introduce my contribution to this research direction so far, concerning evaluation, access to contextual knowledge, and tooling/algorithms. |
Title : Increasing Reward in Biased Natural Selection Decreases Task Performance
| Presenter | Evert Haasdijk |
| Abstract | My talk will present an investigation into a population of robots that evolves through embodied evolution --- an evolutionary process that is not centrally controlled, but emerges from robot interactions just as natural evolution does. The robots select their partners randomly, without reference to any assessment of task performance, but the environment is biased to promote task behaviour by awarding additional lifetime to robots that pick up pucks. The experiments show that the robots do learn to pick up pucks in such a setting. Contrary to what one might expect, increasing the amount of additional lifetime awarded decreases task performance for all settings considered. Closer analysis shows that this decrease is in part due to the fact that the increased lifespan decreases the number of opportunities to spread a robot's genome, but that increasing the award level also negatively affects selection pressure when there is opportunity for robots to spread their genome. We conclude that higher rewards overly emphasise one aspect of robot behaviour and in doing so prevent evolution from exploring the behaviour space. |