Now Computer will say “Try this”: An automated Hypothesis Generation
Now software will help the researchers to decide for what they should look, more than 90 groups of researchers are developing hypothesis-generation software with a hope to use it not for recipe books, but on the vast corpus of scientific literature that has piled up in public databases.
Apples, pork and mushrooms sounds like a important ingredient for recipe of a kebab, but the average someone rarely think of to add strawberries. According to John Gordon from IBM, the result is delicious. Dr Gordon from the team of cognitive-computing and responsible for a machine called Watson; this machine is able to digest and analyze large number of English text and than it can draw interface from it. In March 2014, Watson fed reams of recipes and texts about the food, and it seems that these 4 ingredients would complement each and it will base on sharing a number of flavorsome chemical compounds. And at least Dr Gordon thinks Watson’s suggestion is a winner.
Working for new recipes sounds a trivial and seems to like the use of a multimillion-dollar piece of kit, but Dr. Gordon’s culinary experiment neatly explains the unique idea of Hypothesis Generation and the possible uses of same are certainly not trivial. Researcher was demonstrated the power of the technique published by Olivier Lichtarge of Baylor College of Medicine in August 2014, in collaboration with Dr Gordon’s group they employed it to hunt for kinases, which can activate another protein P53, as it curbs the growth of cancers.
They use the software to read the 186,879 abstract papers and produced a list of the most promising experiments on kinases. The twist was that question sin papers were published before 2003. It means that Dr Lichtarge could check that Watson-based approach came to the same conclusions as those arrived at by human researchers over the subsequent 10 years, in which top 9 kinases the software picked and out of which 7 have subsequently been shown to activate P53. Anne Poupon from French National Institute for Agricultural Research, heads the another group working on automated Hypothesis Generation and her software, Méthode d’Inférence crunches research on hormones and they interact with the 1,500 types of receptor molecules. It recommended to look more closely at certain of these interactions, as the literature on them contains contradictory results, which need to be resolved.
BrainSCANr, devised by Bradley Voytek from University of California, San Diego and his wife Jessica presents the third example of automated hypothesis generation. BrainSCANr is designed to help the neuroscientists, where it will help in the selection of research projects. By sifting more than 3.5m papers, software suggested that clues to the origin of migraines may be found in the levels of serotonin, it is signaling molecule which released by neurons in a region.
Advantages of Hypothesis:
According to Dr Lichtarge, hypothesis generation software works as a part because science writing tends to be free of humor and sarcasm that could trip it up. The source of text can be analyzed in search of hypotheses to test. Web searches rarely lack complex grammar and verbs that can be confusing for software’s. By analyzing the typed words in the web browser of Microsoft’s Internet Explorer and Bing search engine by people wondering why they feel ill while using Microsoft Research, in Redmond, Washington produced the hypotheses on potentially harmful pairings of medications. Microsoft Research’s head, Eric Horvitz, says America’s Food and Drug Administration has formed the team to use “early warning” hypotheses for producing better designs for laboratory experiments.
It all looks rather than promising that both IBM and science, it launched a commercial version in August for automated hypothesis generation software. Discovery Advisor, Dr Gordon hopes, as this service is known to be a money-spinner for the firm.
