國立虎尾科技大學 |

Jointly Studying Linguistic Structure and Language Models : = Methods for a Bilateral Science.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Jointly Studying Linguistic Structure and Language Models :/
其他題名:	Methods for a Bilateral Science.
作者:	Papadimitriou, Isabel Victoria.
面頁冊數:	1 online resource (144 pages)
附註:	Source: Dissertations Abstracts International, Volume: 86-05, Section: B.
Contained By:	Dissertations Abstracts International86-05B.
標題:	Logic. -
電子資源:	click for full text (PQDT)
ISBN:	9798346388548

Jointly Studying Linguistic Structure and Language Models : = Methods for a Bilateral Science.
Papadimitriou, Isabel Victoria.

Jointly Studying Linguistic Structure and Language Models :Methods for a Bilateral Science. - 1 online resource (144 pages)

Source: Dissertations Abstracts International, Volume: 86-05, Section: B.

Thesis (Ph.D.)--Stanford University, 2024.

Includes bibliographical references

In this thesis, I propose methods for a bilateral science connecting language models and human language. My goal is to apply what we know about the human language system to understand language models, and conversely use language models as experimental testbeds and functional theoretical artifacts to expand the horizons of language science. I propose four empirical approaches for analyzing language technology to address fundamental questions in language science. First, I develop a new experimental method, injecting language models with structural information before training, that allows me to investigate a fundamental theoretical question: what makes language learning structurally possible? I find that complex context-sensitive structures are important in scaffolding language learning, even in the absence of a context-free recursive structural bias. Secondly, I develop analysis methods for examining the representation of subjecthood and case in statistical language models. I show how abstract high-order grammatical systems like grammatical role can arise from pure statistical learning. Thirdly, I introduce an analysis methodology to show how language models use syntactic word order in combination with lexical semantics. My analysis shows that language models are a functional theoretical model for how structural and prototype theories can co-exist and function in a single learner. The fourth and final methodology proposes a linguistic syntactic evaluation approach for multilingual language models, showing how English syntactic influence can bias syntactic structure in other languages.

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

Mode of access: World Wide Web

ISBN: 9798346388548Subjects--Topical Terms:

558909
Logic.
Index Terms--Genre/Form:

554714
Electronic books.

Jointly Studying Linguistic Structure and Language Models : = Methods for a Bilateral Science.
LDR:02896ntm a22003617 4500 001 1152796
005 20241213095603.5
006 m o d
007 cr mn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798346388548
035 $a (MiAaPQ)AAI31643357
035 $a (MiAaPQ)Stanfordqf786tt0752
035 $a AAI31643357
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Papadimitriou, Isabel Victoria. $3 1479883
245 1 0 $a Jointly Studying Linguistic Structure and Language Models : $b Methods for a Bilateral Science.
264 0 $c 2024
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: 86-05, Section: B.
500 $a Advisor: Jurafsky, Dan.
502 $a Thesis (Ph.D.)--Stanford University, 2024.
504 $a Includes bibliographical references
520 $a In this thesis, I propose methods for a bilateral science connecting language models and human language. My goal is to apply what we know about the human language system to understand language models, and conversely use language models as experimental testbeds and functional theoretical artifacts to expand the horizons of language science. I propose four empirical approaches for analyzing language technology to address fundamental questions in language science. First, I develop a new experimental method, injecting language models with structural information before training, that allows me to investigate a fundamental theoretical question: what makes language learning structurally possible? I find that complex context-sensitive structures are important in scaffolding language learning, even in the absence of a context-free recursive structural bias. Secondly, I develop analysis methods for examining the representation of subjecthood and case in statistical language models. I show how abstract high-order grammatical systems like grammatical role can arise from pure statistical learning. Thirdly, I introduce an analysis methodology to show how language models use syntactic word order in combination with lexical semantics. My analysis shows that language models are a functional theoretical model for how structural and prototype theories can co-exist and function in a single learner. The fourth and final methodology proposes a linguistic syntactic evaluation approach for multilingual language models, showing how English syntactic influence can bias syntactic structure in other languages.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Logic. $3 558909
650 4 $a Bilingual education. $3 1148431
650 4 $a Natural language. $3 1468611
650 4 $a Semantics. $3 555362
650 4 $a Role models. $3 1413525
650 4 $a Linguistics. $3 557829
650 4 $a Multilingualism. $3 556374
650 4 $a Probability. $3 667790
650 4 $a Syntax. $3 671699
655 7 $a Electronic books. $2 local $3 554714
690 $a 0290
690 $a 0800
690 $a 0282
690 $a 0395
710 2 $a Stanford University. $3 1184533
710 2 $a ProQuest Information and Learning Co. $3 1178819
773 0 $t Dissertations Abstracts International $g 86-05B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31643357 $z click for full text (PQDT)