Researchers from the Australian National University have been recognised for their research in Artificial Intelligence.
Computing PhD candidate Tom Everitt has been awarded the Kurzweil Prize for best AGI Paper at the Ninth Conference on Artificial General Intelligence in New York.
The Kurzweil Prize is given to research that makes exceptional contribution to the artificial general inteligence field.
Tom is exited his paper was recognised at the conference as it validates the community concerns over the development of this technology.
"Given the rapid developement of artificial intelligence, it becomes increasingly important to formalise a theoretical argument which controls agents with any level of intelligence. The award is of course generally inspiring, and will make it easier for me to justify further research on AI safety. It also increases my faith in that the community is starting to take safety questions seriously. I very much doubt that a similar paper could have won even a few years ago."
The paper was written by a group of computer scientists from the Research School of Computer Science, including Honours student Daniel Filan, Dr Mayank Daswani, and Professor Marcus Hutter.
"The intelligent agents group at ANU is world leading in developing foundational mathematical models for artificial intelligence. For me, it is really exciting to apply these techniques AI safety, and (hopefully) contribute to the long-term happiness of humanity" says lead researcher Tom Everitt.
You can watch Tom's presentation from AGI-16 here.
You can read Tom's paper titled Self-Modification of Policy and Utility Function in Rational Agents here.