第3部分 概括大意与完成句子
FirstImage-recognition Software
1.Dartmouth researchers and their colleagues have created an artificial 1 ,software that uses photos to locate documents on the Internet with far gre jthan ever before.
2.The new system, which was tested on photos and is now being applied to , Ishows for the first time that a machine learning algorithm (运算法则)or ,ma9e I recognition and retrieval is accurate and efficient enough toimprove large seaie , document searches online. The system uses pixel (像素)data in images and potentia y video — rather than just text — to locatedocuments. It learns to recognize the pixels associated with a search phrase bystudying the results from text-based image search engines. The knowledgegleaned (收集) from those results can then beapplied to other photos without tags or captions making for more accuratedocument search results.
3."Over the last 30 years," says Associate Professor Lorenzo Torresani,a co-author of the study, "the Web has evolved from a small collection ofmostly text documents to a modern, massive, fast-growing multimedia dataset,where nearly every page includes multiple pictures or videos. When a personlooks at a Web page, he immediately gets the gist (主旨)of it by looking at the pictures in it. Yet, surprisingly, all existing popularsearch engines, such as Google or Bing, strip away the information contained inthe photos and use exclusively the text of Web pages to perform the documentretrieval. Our study is the first to show that modern machine vision systemsare accurate and efficient enough to make effective use of the informationcontained in image pixels to improve document search."
4.The researchers designed and tested a machine vision system — a type ofartificial intelligence that allows computers to learn without being explicitlyprogrammed — that extracts semantic (语义的) information from thepixels of photos in Web pages. This information is used to enrich thedescription of the HTML page used by search engines for document retrieval. Theresearchers tested their approach using more than 600 search queries (查询)on a database of 50 million Web pages. They selected the text-retheval searchengine with the best performance and modified it to make use of the additionalsemantic information extracted by their method from the pictures of the Webpages. They found that this produced a 30 percent improvement in precision overthe original search engine purely based on text.
23. Paragraph 1 __B__
24. Paragraph 2 __C__
25. Paragraph 3 __E__
26. Paragraph 4 __D__
A.Popularity of the new system
B.Publication of the new discovery
C.Function of the new system
D.Artificial intelligence software created
E.Problems of the existing search engines
F.Improvement in document retrieval
27. The new system does documentretrieval by __C__.
28. The new system is expected toimprove precision in __B__.
29. When performing documentretrieval the existing search engines ignore __A__
30. The new system was found moreeffective in document search than the __E__
A.information in images
B.current popular search engines
C.using photos
D.machine vision systems
E.document search
F.description of the HTML page