cleaning up my "bio" for the ML webpage

[gusl]

One of the Machine Learning managers just asked me to write a few sentences about my interests, to go on this webpage.

Here's what I came up with:

Gustavo Lacerda is interested in "automating psychology", i.e. learning cognitive models from data; and more generally, in problems of (empirical and theoretical) scientific discovery. Tackling these tasks involves integrating machine learning with conceptual, logical reasoning.

Recently, he has been working on:
* learning linear causal models (SEMs) by combining ICA with conditional-independence-based methods and regression.
* learning to classify the political orientations of bloggers, by looking at the sentiment expressed in links between them.
* unsupervised learning of SAT-style analogies from large corpora.
* SimStudent, which learns to simulate plausible student behavior in multi-step problems.

He is also interested in argument maps as a medium for argumentation, logical formalization, and decision-making. If you are aware of a non-trivial corpus of argument maps, he would appreciate being contacted.

It seems a bit longish, the style very different from other people (go to the "people" menu), and my stated interests have little to do with the recent projects shown. Any suggestions?


Here's another idea:

Gustavo Lacerda is interested in generative models of natural phenomena, especially in cognitive science. He's fond of models that can be interpreted naturally in terms of existing scientific concepts, thereby making a direct contribution to science (rather than just engineering).

One problem: "natural phenomena" sounds redundant. But just "models of phenomena" also sounds redundant.