Product Lead Former Siri silence his apple products sold raises $ 1M to start Unsilo a semantic search engine
The Apple's Siri acquisition in 2010 gave the company the technology you need to build a voice-activated personal assistant for your iPhone and iPad . A year later , Mads Rydahl - one of the first employees of Siri as director of product design - Apple sold something : a set of patents , nine in all , from the start , he founded before joining Siri.
Today Rydahl working on a new beginning : A semantic search engine called Unsilo , which is prepared for launch in November supported with $ 1 million Danes Oei and Scale Capital incubator , a small business co VC - based in the United States (Palo high ) and Denmark.
Not Wolfram Alpha , and not as Enigma , Unsilo offers users - business customers initially in areas such as scientific research - the ability to find answers to your questions by aggregating data from dozens of different sources . Unsilo The idea behind is that good answers can not be linked to the keywords in the search query , but how those keywords , juxtaposed laterally , as it would be in the spirit of a human being can help you think about the correct answer , using essentially the concept of natural language processing to search .
"We invest a bit of [ processing power ] NLP content index us," said Rydahl . "We have a statistical and semantic approach . We are not trying to go to Q & A as Wolfram Alpha , or facets and filters in the way that Enigma , and although we also have patent data index , which are very different from Lens. org "he said. "They are , in the end , traditional techniques for indexing and classification. We have the idea that we can make a completely different type to match what has been done before. " He describes it as a query expansion to collect changes in the wording and terminology. " Then detect semantic models around the words that match the user's query . "
The end result is pages that give users not only lists of documents, but similar to the query, grouped by the focus , ie how they provide answers to the initial question posed by the user. You can drill a low level through the interface to see good results.
There is also a cunning animation screen that looks like a table or graph variable rate , which gives an idea of how different concepts are literally chained . The display is interactive : if you are looking for vaccines against the cold , and certainly not want to see a homeopathic related results , can eliminate this concept . Or if that is what you want to prioritize, you give priority to the above results . The results of the investigation to rearranged accordingly.y.
The Apple's Siri acquisition in 2010 gave the company the technology you need to build a voice-activated personal assistant for your iPhone and iPad . A year later , Mads Rydahl - one of the first employees of Siri as director of product design - Apple sold something : a set of patents , nine in all , from the start , he founded before joining Siri.
Today Rydahl working on a new beginning : A semantic search engine called Unsilo , which is prepared for launch in November supported with $ 1 million Danes Oei and Scale Capital incubator , a small business co VC - based in the United States (Palo high ) and Denmark.
Not Wolfram Alpha , and not as Enigma , Unsilo offers users - business customers initially in areas such as scientific research - the ability to find answers to your questions by aggregating data from dozens of different sources . Unsilo The idea behind is that good answers can not be linked to the keywords in the search query , but how those keywords , juxtaposed laterally , as it would be in the spirit of a human being can help you think about the correct answer , using essentially the concept of natural language processing to search .
"We invest a bit of [ processing power ] NLP content index us," said Rydahl . "We have a statistical and semantic approach . We are not trying to go to Q & A as Wolfram Alpha , or facets and filters in the way that Enigma , and although we also have patent data index , which are very different from Lens. org "he said. "They are , in the end , traditional techniques for indexing and classification. We have the idea that we can make a completely different type to match what has been done before. " He describes it as a query expansion to collect changes in the wording and terminology. " Then detect semantic models around the words that match the user's query . "
The end result is pages that give users not only lists of documents, but similar to the query, grouped by the focus , ie how they provide answers to the initial question posed by the user. You can drill a low level through the interface to see good results.
There is also a cunning animation screen that looks like a table or graph variable rate , which gives an idea of how different concepts are literally chained . The display is interactive : if you are looking for vaccines against the cold , and certainly not want to see a homeopathic related results , can eliminate this concept . Or if that is what you want to prioritize, you give priority to the above results . The results of the investigation to rearranged accordingly.y.
ليست هناك تعليقات:
إرسال تعليق