東華大學圖書館 |

Revealing and Exploring the Literature's Known Unknowns : = Ignorance and How It Drives Science.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Revealing and Exploring the Literature's Known Unknowns :/
其他題名:	Ignorance and How It Drives Science.
作者:	Boguslav, Mayla Rachel.
面頁冊數:	1 online resource (242 pages)
附註:	Source: Dissertations Abstracts International, Volume: 84-11, Section: A.
Contained By:	Dissertations Abstracts International84-11A.
標題:	Bioinformatics. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30311522click for full text (PQDT)
ISBN:	9798379585341

Revealing and Exploring the Literature's Known Unknowns : = Ignorance and How It Drives Science.
Boguslav, Mayla Rachel.

Revealing and Exploring the Literature's Known Unknowns :Ignorance and How It Drives Science. - 1 online resource (242 pages)

Source: Dissertations Abstracts International, Volume: 84-11, Section: A.

Thesis (Ph.D.)--University of Colorado Denver, Anschutz Medical Campus, 2023.

Includes bibliographical references

Background: Research progresses through accumulating knowledge such that a previously unexplored subject (an unknown unknown) becomes an active research area exploring the questions (known unknowns), until a body of established facts emerges (known knowns). This work aims to help illuminate this process using biomedical natural language processing (BioNLP) to identify, categorize, classify, and explore known unknowns or ignorance statements from the scientific literature. The goal is to help researchers, students, funders, and publishers find the most pertinent research questions or scientific goals for knowledge from the literature based on a topic or experimental results. To do this, researchers must have foundations in both knowledge and questions. However, staying up-to-date on both of them is difficult because of the exponential growth of the scientific literature. Many tools exist to help with knowledge including information extraction systems and knowledge-bases that can be explored by topic or help contextualize experimental results. For the questions, there are some information extraction systems focused on hedging, uncertainty, speculation, factuality, epistemics, and meta-knowledge and a search engine focused on directions and challenges based on a topic (it cannot support queries by experimental results). This prior work mainly focused on the phenomenon in relation to knowledge (i.e., how hedged, certain, speculative, factual, or meta the knowledge is), with only the search engine explicitly focused on new knowledge. With the importance of finding pertinent questions or goals for scientific knowledge, there is a need to go a step further and categorize these statements based on their entailed goal for scientific knowledge (i.e., actionable next steps). Further, no knowledge-bases of such statements, i.e., ignorance-bases, exist that provide summaries and visualizations of such statements based on a topic or experimental results. Thus, we aim to rectify this to help students, researchers, funders, and publishers better understand the state of our collective scientific ignorance (known unknowns or knowledge goals) at scale and across disciplines, hopefully resulting in an accelerated research process. To determine the feasibility of these general computational ideas, we apply them to the prenatal nutrition field to help find pertinent questions that could affect mothers and offspring globally.Methods: To capture goals for scientific knowledge, we created a novel BioNLP task to identify, characterize, and classify statements of ignorance based on their entailed knowledge goal. Through manual annotation, we created a taxonomy of ignorance, annotation guidelines, a corpus, and classification models. We also identified the biomedical concepts (ontology concepts) in the ignorance statements to understand their biomedical subjects. We systematically characterized the factors that contributed to the accuracy and efficiency of several approaches to biomedical concept recognition, while aiming to improve performance. Together, the ignorance and biomedical concept classification formed the first ignorance-base on the prenatal nutrition literature. To demonstrate its power, we present two methods of exploration: (1) exploration by topic (e.g., vitamin D) to show that ignorance statements can provide new ideas for future research, and (2) exploration by experimental results (e.g., vitamin D and spontaneous preterm birth gene list) to help a researcher contextualize their results in the ignorance landscape providing questions that their results may bear on potentially from other disciplines.Results: We show that it is possible to characterize known unknowns as knowledge goals (ignorance taxonomy), that humans can identify statements of ignorance in the literature (annotation task to create a corpus), and that they can be automatically identified (ignorance classification). For the biomedical concepts, we present an automatic biomedical concept recognition system that performed comparably with state-of-the-art systems with some substantial efficiencies in the time and computational resources required for tuning and training. Combining these, we created the first ignorance-base on the prenatal nutrition literature. Exploring it by the topic vitamin D showed that it is possible to find other areas of research (immune system, respiratory system, and brain development) with lots of questions (ignorance statements) that are ripe for future research. Exploring it by a gene list, provided a novel research area (the brain) that implied a different discipline (neuroscience) of which could be explored for answers. Overall, the ignorance-base provided a question foundation rooted in knowledge goals and ideas for future research.Conclusion: Our goal is to help students, researchers, funders, and publishers better understand the state of our collective scientific ignorance in order to help accelerate translational research through illuminating the known unknowns and their respective goals for scientific knowledge.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2023

Mode of access: World Wide Web

ISBN: 9798379585341Subjects--Topical Terms:

553671
Bioinformatics.
Subjects--Index Terms:

Biomedical concept recognitionIndex Terms--Genre/Form:

542853
Electronic books.

Revealing and Exploring the Literature's Known Unknowns : = Ignorance and How It Drives Science.
LDR:06635nmm a2200421K 4500 001 2358377
005 20230731112634.5
006 m o d
007 cr mn ---uuuuu
008 241011s2023 xx obm 000 0 eng d
020 $a 9798379585341
035 $a (MiAaPQ)AAI30311522
035 $a AAI30311522
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Boguslav, Mayla Rachel. $3 3698909
245 1 0 $a Revealing and Exploring the Literature's Known Unknowns : $b Ignorance and How It Drives Science.
264 0 $c 2023
300 $a 1 online resource (242 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: 84-11, Section: A.
500 $a Advisor: Hunter, Lawrence E.; Leach, Sonia; Costello, James.
502 $a Thesis (Ph.D.)--University of Colorado Denver, Anschutz Medical Campus, 2023.
504 $a Includes bibliographical references
520 $a Background: Research progresses through accumulating knowledge such that a previously unexplored subject (an unknown unknown) becomes an active research area exploring the questions (known unknowns), until a body of established facts emerges (known knowns). This work aims to help illuminate this process using biomedical natural language processing (BioNLP) to identify, categorize, classify, and explore known unknowns or ignorance statements from the scientific literature. The goal is to help researchers, students, funders, and publishers find the most pertinent research questions or scientific goals for knowledge from the literature based on a topic or experimental results. To do this, researchers must have foundations in both knowledge and questions. However, staying up-to-date on both of them is difficult because of the exponential growth of the scientific literature. Many tools exist to help with knowledge including information extraction systems and knowledge-bases that can be explored by topic or help contextualize experimental results. For the questions, there are some information extraction systems focused on hedging, uncertainty, speculation, factuality, epistemics, and meta-knowledge and a search engine focused on directions and challenges based on a topic (it cannot support queries by experimental results). This prior work mainly focused on the phenomenon in relation to knowledge (i.e., how hedged, certain, speculative, factual, or meta the knowledge is), with only the search engine explicitly focused on new knowledge. With the importance of finding pertinent questions or goals for scientific knowledge, there is a need to go a step further and categorize these statements based on their entailed goal for scientific knowledge (i.e., actionable next steps). Further, no knowledge-bases of such statements, i.e., ignorance-bases, exist that provide summaries and visualizations of such statements based on a topic or experimental results. Thus, we aim to rectify this to help students, researchers, funders, and publishers better understand the state of our collective scientific ignorance (known unknowns or knowledge goals) at scale and across disciplines, hopefully resulting in an accelerated research process. To determine the feasibility of these general computational ideas, we apply them to the prenatal nutrition field to help find pertinent questions that could affect mothers and offspring globally.Methods: To capture goals for scientific knowledge, we created a novel BioNLP task to identify, characterize, and classify statements of ignorance based on their entailed knowledge goal. Through manual annotation, we created a taxonomy of ignorance, annotation guidelines, a corpus, and classification models. We also identified the biomedical concepts (ontology concepts) in the ignorance statements to understand their biomedical subjects. We systematically characterized the factors that contributed to the accuracy and efficiency of several approaches to biomedical concept recognition, while aiming to improve performance. Together, the ignorance and biomedical concept classification formed the first ignorance-base on the prenatal nutrition literature. To demonstrate its power, we present two methods of exploration: (1) exploration by topic (e.g., vitamin D) to show that ignorance statements can provide new ideas for future research, and (2) exploration by experimental results (e.g., vitamin D and spontaneous preterm birth gene list) to help a researcher contextualize their results in the ignorance landscape providing questions that their results may bear on potentially from other disciplines.Results: We show that it is possible to characterize known unknowns as knowledge goals (ignorance taxonomy), that humans can identify statements of ignorance in the literature (annotation task to create a corpus), and that they can be automatically identified (ignorance classification). For the biomedical concepts, we present an automatic biomedical concept recognition system that performed comparably with state-of-the-art systems with some substantial efficiencies in the time and computational resources required for tuning and training. Combining these, we created the first ignorance-base on the prenatal nutrition literature. Exploring it by the topic vitamin D showed that it is possible to find other areas of research (immune system, respiratory system, and brain development) with lots of questions (ignorance statements) that are ripe for future research. Exploring it by a gene list, provided a novel research area (the brain) that implied a different discipline (neuroscience) of which could be explored for answers. Overall, the ignorance-base provided a question foundation rooted in knowledge goals and ideas for future research.Conclusion: Our goal is to help students, researchers, funders, and publishers better understand the state of our collective scientific ignorance in order to help accelerate translational research through illuminating the known unknowns and their respective goals for scientific knowledge.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Bioinformatics. $3 553671
650 4 $a Systematic biology. $3 3173492
650 4 $a Biomedical engineering. $3 535387
650 4 $a Epistemology. $3 896969
653 $a Biomedical concept recognition
653 $a Information extraction
653 $a Knowledge representation
653 $a Knowledge-bases
653 $a Natural language processing
653 $a BioNLP
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0715
690 $a 0423
690 $a 0541
690 $a 0393
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a University of Colorado Denver, Anschutz Medical Campus. $b Computational Bioscience. $3 3563666
773 0 $t Dissertations Abstracts International $g 84-11A.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30311522 $z click for full text (PQDT)