The Applied Artificial Intelligence Laboratory (AAIL) is part of the Departamento de Computación, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires and the Instituto de Investigación en Ciencias de la Computación, CONICET-UBA. https://icc.fcen.uba.ar/
The AAIL is an interdisciplinary environment, which combines different aspects from Computational Neuroscience, as complexity and randomness perception in humans, computational linguistics, data mining in big text corpus and source code, interactive dialogue systems, speech recognition and real-time analysis of brain signals. Our main research projects are:
« Massive-scale neuroscience. »
With the development of information technology and the consequent everyday activity in the Internet, many human products are represented in the same digital environment. Current research can capitalize this unprecedented opportunity: a virtually infinite repository of human computation. This change of scale allows the exploration of procedures in high dimensional spaces which were previously intractable with data generated in the laboratory, opening a privileged window to understand human cognition. We focus on Education, Decision Making and Social Psychology.
« Spoken Language Processing. »
This branch of AI aims at developing systems capable of effectively manipulating spoken language. Its ultimate goal consists in creating a talking machine that passes the Turing test: a machine that can engage in open-domain conversation with a performance indistinguishable from that of a human being. The main challenges of SLP include speech recognition (translating spoken words into text), speech synthesis (transforming text into spoken words), dialogue management (interpreting the input and deciding how to respond), and prosody modeling (understanding the relation between ‘what’ is said and ‘how’ it is said). We focus on two languages of study: English and Spanish.
« Signal Processing. »
GPU-based Independent Component Analysis (ICA), Real-time spike-sorting, Brain-Computer Interfaces.
« Randomness, Computability and Logic. »
Algorithmic randomness, computability theory, Kolmogorov complexity, Algorithmic information theory, Model theory, Modal logics.
« Parsing human behavior in natural scenarios: Neural correlates and computational models. »
Any everyday visual behaviour, such as searching for a friend in a crowded place or reading this paragraph, involves a sequence or routine of unitary operations (i.e. fixations) embedded in a larger mental global program. Combining magneto/electroencephalography (M/EEG) and eye tracking techniques we study brain signatures of those unitary operations and the ones occurring during the sequence of fixations. Moreover, we develop computational models that help us to understand how the available information is processed and integrated. By doing so, we aim to understand how local and global information is combined in implementing visual routines in natural tasks.
« Development of Brain Processes of Cognitive Control in Contexts of Social Vulnerability. »
We study the impact of socioeconomic variables on brain processes underlying cognitive control in children, collaborating in the evaluation of brain changes after interventions aimed to stimulate cognitive development in children living in social vulnerability conditions. Also, we develop new tools to evaluate and improve the protocols performed outside the lab.