PhD position - Abstraction and reasoning challenge: Create an AI capable of solving reasoning tasks it has never seen before
Can a computer learn complex, abstract tasks from just a few examples? Current machine learning techniques are data-hungry and brittle—they can only make sense of patterns they've seen before. Using current methods like reinforcement learning, an algorithm can gain new skills by exposure to large amounts of data, but cognitive abilities that could broadly generalize to many tasks remain elusive. This makes it very challenging to create systems that can handle the variability and unpredictability of the real world, such as domestic robots or self-driving cars. However, alternative approaches, like inductive programming, offer the potential for more human-like abstraction and reasoning.
The abstraction and reasoning corpus (ARC) provides a benchmark to measure AI skill-acquisition on unknown tasks, with the constraint that only a handful of demonstrations are shown to learn a complex task (example). This competition was initially created by the creator of the Keras neural networks library and it’s explained in this paper. The idea is to move beyond the competition timeframe to create an AI that can solve reasoning tasks it has never seen before and set up a path toward a PhD in AI. It is expected that novel work in terms of a paper should be produced during this period.
Barcelona, Spain, Barcelona Biomedical Research park. The laboratory is part of the Barcelona Biomedical Research Park which, with a privileged location on the shoreline of the Mediterranean sea, constitutes one of the most exciting interdisciplinary research centres in Southern Europe with more than 1000 scientists in the building alone.
Information on the application process:
Please send an email to gianni.defabritiis[at]upf.edu with subject “JOB PhD1 2020” with a CV and a cover letter together with the names of up to three contacts for requesting recommendations
Deadline to submit applications 15/6/2020