基于深度學(xué)習(xí)的學(xué)術(shù)搜索引擎——Semantic Scholar
發(fā)布時(shí)間：2019-03-17 12:20
【摘要】：[目的/意義]Alpha Go戰(zhàn)勝李世石后,人工智能的研究與發(fā)展備受關(guān)注。在此之前不久,基于深度學(xué)習(xí)的Semantic Scholar免費(fèi)學(xué)術(shù)搜索引擎的問(wèn)世,也為科研工作者們搜索和篩選學(xué)術(shù)文獻(xiàn)資源帶來(lái)了新的體驗(yàn)。[方法 /過(guò)程]在介紹人工智能、機(jī)器學(xué)習(xí)和深度學(xué)習(xí)之間關(guān)系的基礎(chǔ)上,介紹了Semantic Scholar的檢索功能,重點(diǎn)就該引擎基于系統(tǒng)在理解文獻(xiàn)內(nèi)容基礎(chǔ)上的學(xué)術(shù)影響力評(píng)價(jià)功能作了分析,并將Semantic Scholar與現(xiàn)行主流學(xué)術(shù)搜索引擎Google Scholar、Microsoft Academic、必應(yīng)學(xué)術(shù)和百度學(xué)術(shù)進(jìn)行比較研究。[結(jié)果/結(jié)論]Semantic Scholar通過(guò)機(jī)器學(xué)習(xí)可以使系統(tǒng)理解不同引用之間的影響力差異,提出了基于引用內(nèi)容分析的學(xué)術(shù)影響力評(píng)價(jià)指標(biāo),但在信息來(lái)源、學(xué)科范圍、檢索功能和個(gè)性化服務(wù)功能方面還有待進(jìn)一步完善。最后提出今后學(xué)術(shù)搜索引擎的發(fā)展展望。
[Abstract]:Objective: after Alpha Go defeated Li Shi-Shi, the research and development of artificial intelligence (AI) has been paid more and more attention. Not long ago, Semantic Scholar, a free academic search engine based on in-depth learning, brought new experiences to researchers in searching and screening academic literature resources. [methods / processes] on the basis of introducing the relationship among artificial intelligence, machine learning and in-depth learning, this paper introduces the retrieval function of Semantic Scholar. This paper focuses on the analysis of the academic influence evaluation function of the engine based on the understanding of the contents of the literature, and compares Semantic Scholar with the current mainstream academic search engine Google Scholar,Microsoft Academic, Bing academic and Baidu academic research. [results / conclusion] through machine learning, Semantic Scholar can make the system understand the difference of influence among different citations, and put forward the evaluation index of academic influence based on citation content analysis, but in the information source and subject scope, The retrieval function and personalized service function need to be further improved. Finally, the development prospect of academic search engine in the future is put forward.
【作者單位】： 北京化工大學(xué)圖書(shū)館;
【分類(lèi)號(hào)】：G252.7;G434
，
本文編號(hào)：2442296
[Abstract]:Objective: after Alpha Go defeated Li Shi-Shi, the research and development of artificial intelligence (AI) has been paid more and more attention. Not long ago, Semantic Scholar, a free academic search engine based on in-depth learning, brought new experiences to researchers in searching and screening academic literature resources. [methods / processes] on the basis of introducing the relationship among artificial intelligence, machine learning and in-depth learning, this paper introduces the retrieval function of Semantic Scholar. This paper focuses on the analysis of the academic influence evaluation function of the engine based on the understanding of the contents of the literature, and compares Semantic Scholar with the current mainstream academic search engine Google Scholar,Microsoft Academic, Bing academic and Baidu academic research. [results / conclusion] through machine learning, Semantic Scholar can make the system understand the difference of influence among different citations, and put forward the evaluation index of academic influence based on citation content analysis, but in the information source and subject scope, The retrieval function and personalized service function need to be further improved. Finally, the development prospect of academic search engine in the future is put forward.
【作者單位】： 北京化工大學(xué)圖書(shū)館;
【分類(lèi)號(hào)】：G252.7;G434
，
本文編號(hào)：2442296
本文鏈接：http://sikaile.net/kejilunwen/sousuoyinqinglunwen/2442296.html
最近更新
教材專著
