"The younger brother" IBM Watson engaged in the analysis of patents
On Habré, IBM’s self-taught erudite computer IBM Watson, who is able to answer questions by understanding an array of unstructured data, has been repeatedly discussed. He showed his best side by winning several rounds of the Jeopardy quiz (His Own Game) with the champions of this game.
Then Watson went to medicine : he began to analyze publications from the national medical database PubMed and dozens of textbooks. By May 2011, he had mastered the college and first-year university programs, that is, he was able to correctly answer exam questions that medical students take. According to the developers, in three to five years, Watson will be ready for the first pilot tests for treating real patients, and in 8-10 years such computers can be widely used in hospitals as a diagnostic tool. The first contract for the commercial use of Watson has already been signed with the medical company WellPoint.
The diagnostician is not the only profession where it is difficult for people to struggle with automated systems for analyzing an array of unstructured data. There is another profession, albeit not so honorable. This is a lawyer, a specialist in patent law. IBM has developed the Strategic IP Insight Platform (SIIP) system, which they offer to buy large corporations to analyze and manage their intellectual property portfolio. The system performs data mining in the entire patent database, analyzes the archive of scientific journals and other information on the topic, so that the owner of intellectual property can analyze and detect "problems" in his or someone else's patent database. ')
SIIP was used to create a database containing records of medical patents, as well as biomedical journals and information about pharmaceuticals. It is assumed that this base will help in the development of new drugs.
From each article / patent / description, the SIIP system extracted the names of chemical preparations, diagrams, names of scientists-inventors and other keywords. To date, 2.5 million drug names, 4.7 million patents, and 11 million journal articles from 1976 to 2000 have been processed. IBM donated some of the information gathered to the open source directory of drugs, PubChem .
According to some experts , the intelligent platform SIIP can be successfully used to combat patent trolls. Unfortunately, this option looks almost unreal, given that IBM itself obtains the most patents in the world and is one of the richest intellectual property owners, using its baggage of 10,000+ patents for absurd lawsuits. Worse, IBM is actively supporting software patents , and software patents are even more dangerous for IT companies than trolls .
So it is unlikely the artificial intelligence created by IBM will fight patent trolls. Quite the contrary - it can be used to generate even more patents. Or to create the most powerful patent troll with the use of AI, capable of generating an infinite number of "inventions". Hopefully, such predictions will never be realized.
The SIIP platform is based on the same technologies as IBM Watson. Recall that Watson is an IBM POWER7 server, Apache Hadoop and Apache UIMA are used for information processing, as well as a number of other programs created over four years by the IBM development team.
The car is named after IBM founder Thomas Watson. In the future, it will become part of an artificial intelligence system capable of understanding human language and answering any questions, drawing on information from the Internet. The ability to filter and analyze facts can be used in medicine (for making complex diagnoses empirically after analyzing the symptoms), automatic support services (for answering questions), tourist guides and ... for generating patents, maybe.