Welcome!

Research

Our research is motivated by the following thesis: background world knowledge is key to intelligent decision making. While we humans routinely use such background knowledge (e.g., rose-hasColor-red, Bangalore-isACityIn-India, etc.) in making decisions in our daily lives, intelligent machines (e.g., systems using Machine Learning) unfortunatly don't have access to such knowledge. Our research is focused on bridging this knowledge bottleneck and in making broad-coverage world knowledge available to machines (and humans) at the right granularity and at the right time.

We view unstructured Web data (e.g., Webpages, tweets, blogs, etc) as one rich source of such knowledge. One of our primary research goals is to extract, orgazine, and make readily available the knowledge trapped inside such unstructured text data on a large scale. To achieve these goals, our research spans the areas of Machine Learning and Natural Language Processing.

In addition to Web-scale unstructured text corpus, we are also interested in understanding how knowledge is organized and processed in the human brain. We are of the opinion that text corpus and brain imaging data (e.g., fMRI, MEG, etc.) offer complementary views of the same latent phenomenon, and we would like to benefit by reconciling these two modalities.

MALL Lab actively collaborates with the Read the Web (NELL) and the Brain Research Group at CMU. Additional collaborators include Polo Chau (Georgia Tech), Christos Faloutsos (CMU), and Nikos Sidiropoulos (UMinnesota).