Pickens writes: "Geographers report that fundamental principles of geography place Osama bin Laden in one of three buildings in the Pakistan town of Parachinar. "If he's still alive, he honestly could be sitting there right now," said Thomas W. Gillespie, the study's lead author. The UCLA findings rely on two principles used in geography: distance-decay theory and island biogeographic theory. Distance-decay theory holds that as one travels farther away from a precise location with a specific composition of species — or, in this case, a specific composition of cultural and physical factors — the probability of finding spots with that same specific composition decreases exponentially. "The farther bin Laden moves from his last reported location into the more secular parts of Pakistan or into India, the greater the probability that he will be in an area with a different cultural composition, thereby increasing the probability of his being captured or eliminated," Gillespie said. Island biogeographic theory holds that large and close islands have larger immigration rates and will support more species than smaller, more isolated islands. "Island biology theory predicts that he would find his way to the largest but least isolated city of that area," said Gillespie. After the team determined that Parachinar best met the criteria, they then used satellite photos to look for buildings in Parachinar that met six specific criteria and identified three structures in the city where bin Laden is most likely to be hiding. The team proposes that it is past time "to embrace this technology and create a public database concerning models or hypotheses about Bin Laden's current location.""
Quoting from the summary:
That is, he explicitly says that _most_ (that is, not all) of their work is at the word/phrase level. This implies that some is at levels of abstraction above that. They may not be "concentrating on the sentence" but that doesn't mean that they're ignoring it entirely. Furthermore, there are well-known ways of creating good approximations of the meaning of a document that don't consider word order at all. The classic is the TF-IDF model, but there are others (Latent Semantic Analysis, other types of topic models) that are richer and more descriptive. No, they don't capture everything about the semantics or pragmatics of a document, but they do well enough to (for instance) provide good predictors of the grade of an essay as assigned by a panel of human graders.'We think what's important about natural language is the mapping of words onto the concepts that users are looking for. But we don't think it's a big advance to be able to type something as a question as opposed to keywords
... understanding how words go together is important ... That's a natural-language aspect that we're focusing on. Most of what we do is at the word and phrase level; we're not concentrating on the sentence.'"