Pablo Caceres

Credentials: Ph.D. Student, Psychology

Cognitive Science and Machine Learning

Research Advisors: Timothy T. Rogers and Joseph Austeirweil

I’m a PhD student in Psychology at the University of Wisconsin-Madison, working under the supervision of Timothy T. Rogers and Joseph Austerweil, and supported by a Fulbright Scholarship.

My research interests lie at the intersection of cognitive science, machine learning, and data science.

At LUCID, my research is focused in the use of machine learning and cognitive science to explore questions about the structure of human conceptual representations.

Here is an example: I can give you a concept, like animal, and then ask you to enumerate as many animals as you can in 60 seconds. Now, from a cognitive science perspective, the question is what kind of structure will better represent the way in which the words that you mentioned, in this case animals, are organized in your mind. For instance, one way to approach this could be by using a network graph, where words are nodes, like little points, and the relations or semantic associations between words are represented as lines connecting those nodes or words.

What I find interesting about this research is that we can use machine learning techniques to create such structures and then to assess which kind representation makes a better job predicting human behavior in a variety of tasks. Such research is potentially beneficial from a variety of perspectives: from scientific point of view, it can help to illuminate the understanding of how knowledge is structured and represented in the mind and brain; from an applied perspective, it can help to design systems that are able to assist human decision-making processes in a variety of areas, like the medical diagnosis, recommender systems, or legal processes; and from a social and human well-being perspective, it can help to improve educational processes and the treatment of neuro-psychological disorders like Alzheimer disease, Broca’s aphasia, and other disorders, where language, memory, and semantic impairments are prominent.

In the long term, my aim is to combine the best of cognitive science and a variety of machine learning techniques, to tackle scientific, technological and social issues, in a flexible, innovative and creative manner.