Researchers at the Australian National University (ANU) will have their research published at the International Joint Conference on Artificial Intelligence (IJCAI), the main international gathering of researchers in Artificial Intelligence (AI).
The paper features research relating to artificial curiosity and the ability to encourage and train an IA agent to explore and speed up learning. The paper is titled: Count-Based Exploration in Feature Space for Reinforcement Learning. Authored by: Jarryd Martin, Suraj Narayanan, Tom Everitt and Professor Marcus Hutter.
“Artificial curiosity has been explored theoretically, but we are one of the very first groups to develop a curious agent that outperforms state-of-the-art algorithms on large practical problems. We test agents by letting them play video games, and our agent’s curiosity helped it outperform non-curious agents from leading labs like Google DeepMind, OpenAI, UC Berkeley and Alberta. Our research lends weight to the hypothesis that curiosity is crucial for learning and general intelligence” said Jarryd Martin.
The International Joint Conference on Artificial Intelligence will be held in Melbourne this month and brings together the best, brightest and most influential members of the AI community.