Indexing World's Knowledge
We develop algorithms to make machines understand text, images and location data. For instance, we have indexed the US census data and have built algorithms that given a location information can make fast and accurate predictions about the demographics of the person at that location (their likely age, gender, education, income, ethnicity, etc.). We have also been indexing Wikipedia (together with Wikidata, Freebase and DBPedia). We have designed a battery of Natural Language Processing algorithms that use this knowledge index to understand the meaning of short, unstructured and idiosyncratic text such as social media content.