Limitless possibilities with a search engine that can think like us
Recently in the 3 geeks and a law blog there was a post about IBM’s new project called “Watson”. This project is the next step in trying to get computers to understand how humans think. How is this relevant to research? Well the project’s major implication is to get the computer to understand “natural language”. In other words to get computers to understand a sentence or phrase that might be ambiguous in various ways but that a person from inferences and assumption will easily be able to decipher and understand. Google, Lexis Nexis and Westlaw all have a section on searching using “natural language” but for those of us who have tried it before we know that such a search is not intuitive. A natural language search in Lexis Nexis or West Law will normally get you very little or no hits if you type a sentence or a question that is too complex or ambiguous. One of the reasons that Lexis Nexis and West Law is so hard to navigate is because there is just so much information and its hard for sift through and find relevant information from such a huge amount of data. Yet if the Watson project were to succeed it will be an amazing advancement when applied to research engines. Why? because the computer will be able to quickly figure out the relevant data for you instead of you having to perform multiple searches on key words and hope you get relevant results. Not only will computers understanding of natural language make it much easier for lawyers to research, as asking specific legal questions would become more plausible, but it would also expand the possibility of searching for legal and non legal information that would be relevant for your case in the world wide web with search engines like google. The option of not using complicated and often confusing terms and connectors in Lexis Nexis and West Law search engines will also make the process of researching less stressful (for those paying for these services) and more time efficient. More importantly, however, is the fact that normal people will be able to perform legal research more efficiently with a search engine that is more intuitive to a normal person’s way of thinking regardless of having a JD.
Thus far the project has been able to get the computer to play Jeopardy. It is a huge step in getting the computer to understand natural language because Jeopardy is a game that has questions that are made for humans to answer not computers. It is a game that you have to understand the English language, including slang and colloquial words. Jeopardy is also based on a broad domain of knowledge and these answers must be formulated in a quick amount of time in order to beat other players in the game, in their simulation Watson faced off with Jeopardy champions. The success of the Watson project will definantly lead to a amazing computer capable of advanced question and answer that can revolutionize not only research but perhaps our daily life. For more information about the article just discussed and a video of the Watson project made by IBM go to: