Projects

First-Order Probabilistic Inference consists of performing inference with probabilistic models specified with an expressive language using quantified relational predicates, as opposed to, for example, graphical models, which specify dependencies for each (propositional) random variable.
It consists of two parts: BLOG/DBLOG (Bayesian LOGic/ Dynamic BLOG), and lifted inference.
Projects I have previously worked on:
Textual Entailment/Question Answering: This project deals with the problem of detecting the entailment between two snippets of text, using lexical, syntactical and semantic resources.
Multi-modal Computer Tutoring: This interdisciplinary project, involving research groups in the Beckman Institute working with education, speech recognition, facial recognition, vision and reasoning (our part), aims at creating a multimodal computer tutoring environment for children to interact and learn about physical objects (in this case, Lego gears).