國立虎尾科技大學 |

Design and Data Mining Techniques for Large-Scale Scholarly Digital Libraries and Search Engines.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Design and Data Mining Techniques for Large-Scale Scholarly Digital Libraries and Search Engines./
作者:	Rohatgi, Shaurya.
面頁冊數:	1 online resource (144 pages)
附註:	Source: Dissertations Abstracts International, Volume: 85-05, Section: B.
Contained By:	Dissertations Abstracts International85-05B.
標題:	Internships. -
電子資源:	click for full text (PQDT)
ISBN:	9798380735773

Design and Data Mining Techniques for Large-Scale Scholarly Digital Libraries and Search Engines.
Rohatgi, Shaurya.

Design and Data Mining Techniques for Large-Scale Scholarly Digital Libraries and Search Engines. - 1 online resource (144 pages)

Source: Dissertations Abstracts International, Volume: 85-05, Section: B.

Thesis (Ph.D.)--The Pennsylvania State University, 2023.

Includes bibliographical references

The exponential growth of digital libraries and the proliferation of scholarly content in electronic formats have made data mining and information retrieval essential tools for effectively managing, organizing, and disseminating knowledge. This thesis provides a comprehensive analysis of the advancements and challenges in these fields, with a focus on mathematical information retrieval from scholarly documents, figure captioning and classification of scientific images, and searching and re-ranking techniques for largescale scholarly documents. We also explore the future of scholarly search, considering the potential roles of generative artificial intelligence and scientific question-answering systems in these domains. In the initial section of this thesis, we delve into the complex design and implementation aspects involved in building a large-scale digital library. We discuss various challenges and critical design decisions that were made to ensure the library's long-term sustainability and ease of maintenance. Furthermore, we provide an in-depth analysis of the unique and robust components of CiteSeerX, including its advanced crawling, extraction, ingestion, and production-ready capabilities. Our focus then shifts to the investigation of search and re-ranking techniques specifically tailored for large-scale scholarly documents. We delve into various approaches for indexing, searching, and ranking vast collections of scientific literature, proposing inventive methods for optimizing their performance and scalability. After successfully implementing our system and achieving an impressive index of over 15 million academic papers, we explore the numerous potential applications and opportunities that can arise from this extensive collection of scholarly articles. Following this, we conduct a thorough examination of cutting-edge mathematical information retrieval techniques for extracting and processing mathematical expressions from scholarly documents. We present an exhaustive review of existing approaches, shedding light on their strengths and weaknesses, and propose innovative methods that significantly enhance the accuracy and efficiency of mathematical information retrieval systems. Subsequently, we discuss a subset of CiteSeerX data that is focused on Computational Linguistics (CL) - The ACL Anthology Corpus. We provide the metadata, full-text, and citation graph for the CL domain. This dataset is then analyzed for deeper insights into the evolving direction of the field and the potential applications that can be developed from it. One such application is addressing the challenges of figure captioning and classification of scientific images. We analyze state-of-the-art methods for extracting and processing image data from scholarly documents and propose a groundbreaking approach that effectively combines advanced image processing techniques with cutting-edge machine learning algorithms for highly accurate and reliable figure captioning and classification. Lastly, we discuss the future of scholarly search and the role of generative AI in scientific question answering. We envision a question answering system, which looks at the relevant literature and formulates an answer for the researcher's information need. To this end, we investigate the potential of large language models and search for enabling such capabilities and outline the challenges and opportunities that lie ahead in this exciting domain.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798380735773Subjects--Topical Terms:

1413484
Internships.
Subjects--Index Terms:

Electronic formatsIndex Terms--Genre/Form:

554714
Electronic books.

Design and Data Mining Techniques for Large-Scale Scholarly Digital Libraries and Search Engines.
LDR:04879ntm a22003857 4500 001 1142869
005 20240429063853.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798380735773
035 $a (MiAaPQ)AAI30720673
035 $a (MiAaPQ)PennState_22325szr207
035 $a AAI30720673
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Rohatgi, Shaurya. $3 1467374
245 1 0 $a Design and Data Mining Techniques for Large-Scale Scholarly Digital Libraries and Search Engines.
264 0 $c 2023
300 $a 1 online resource (144 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-05, Section: B.
500 $a Advisor: Giles, C. Lee.
502 $a Thesis (Ph.D.)--The Pennsylvania State University, 2023.
504 $a Includes bibliographical references
520 $a The exponential growth of digital libraries and the proliferation of scholarly content in electronic formats have made data mining and information retrieval essential tools for effectively managing, organizing, and disseminating knowledge. This thesis provides a comprehensive analysis of the advancements and challenges in these fields, with a focus on mathematical information retrieval from scholarly documents, figure captioning and classification of scientific images, and searching and re-ranking techniques for largescale scholarly documents. We also explore the future of scholarly search, considering the potential roles of generative artificial intelligence and scientific question-answering systems in these domains. In the initial section of this thesis, we delve into the complex design and implementation aspects involved in building a large-scale digital library. We discuss various challenges and critical design decisions that were made to ensure the library's long-term sustainability and ease of maintenance. Furthermore, we provide an in-depth analysis of the unique and robust components of CiteSeerX, including its advanced crawling, extraction, ingestion, and production-ready capabilities. Our focus then shifts to the investigation of search and re-ranking techniques specifically tailored for large-scale scholarly documents. We delve into various approaches for indexing, searching, and ranking vast collections of scientific literature, proposing inventive methods for optimizing their performance and scalability. After successfully implementing our system and achieving an impressive index of over 15 million academic papers, we explore the numerous potential applications and opportunities that can arise from this extensive collection of scholarly articles. Following this, we conduct a thorough examination of cutting-edge mathematical information retrieval techniques for extracting and processing mathematical expressions from scholarly documents. We present an exhaustive review of existing approaches, shedding light on their strengths and weaknesses, and propose innovative methods that significantly enhance the accuracy and efficiency of mathematical information retrieval systems. Subsequently, we discuss a subset of CiteSeerX data that is focused on Computational Linguistics (CL) - The ACL Anthology Corpus. We provide the metadata, full-text, and citation graph for the CL domain. This dataset is then analyzed for deeper insights into the evolving direction of the field and the potential applications that can be developed from it. One such application is addressing the challenges of figure captioning and classification of scientific images. We analyze state-of-the-art methods for extracting and processing image data from scholarly documents and propose a groundbreaking approach that effectively combines advanced image processing techniques with cutting-edge machine learning algorithms for highly accurate and reliable figure captioning and classification. Lastly, we discuss the future of scholarly search and the role of generative AI in scientific question answering. We envision a question answering system, which looks at the relevant literature and formulates an answer for the researcher's information need. To this end, we investigate the potential of large language models and search for enabling such capabilities and outline the challenges and opportunities that lie ahead in this exciting domain.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Internships. $3 1413484
650 4 $a Digital libraries. $3 561243
650 4 $a Search engines. $3 641795
650 4 $a Anthologies. $3 813772
650 4 $a Information retrieval. $3 563691
650 4 $a Linux. $3 1467375
650 4 $a Information technology. $3 559429
650 4 $a Information science. $3 561178
653 $a Electronic formats
653 $a Digital libraries
653 $a Computational Linguistics
653 $a Mathematical information
655 7 $a Electronic books. $2 local $3 554714
690 $a 0489
690 $a 0723
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a The Pennsylvania State University. $3 845556
773 0 $t Dissertations Abstracts International $g 85-05B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30720673 $z click for full text (PQDT)