FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3186095502> ?p ?o ?g. }

• W3186095502 abstract "Transformer-based language models (LMs) continue to advance state-of-the-art performance on NLP benchmark tasks, including tasks designed to mimic human-inspired “commonsense” competencies. To better understand the degree to which LMs can be said to have certain linguistic reasoning skills, researchers are beginning to adapt the tools and concepts of the field of psychometrics. But to what extent can the benefits flow in the other direction? I.e., can LMs be of use in predicting what the psychometric properties of test items will be when those items are given to human participants? We gather responses from numerous human participants and LMs (transformer- and non-transformer-based) on a broad diagnostic test of linguistic competencies. We then use the responses to calculate standard psychometric properties of the items in the diagnostic test, using the human responses and the LM responses separately. We then determine how well these two sets of predictions match. We find cases in which transformer-based LMs predict psychometric properties consistently well in certain categories but consistently poorly in others, thus providing new insights into fundamental similarities and differences between human and LM reasoning." @default.
• W3186095502 created "2021-08-02" @default.
• W3186095502 creator A5009772358 @default.
• W3186095502 creator A5037647150 @default.
• W3186095502 creator A5046323511 @default.
• W3186095502 creator A5064336148 @default.
• W3186095502 date "2021-01-01" @default.
• W3186095502 modified "2023-10-17" @default.
• W3186095502 title "Can Transformer Language Models Predict Psychometric Properties?" @default.
• W3186095502 cites W1555938369 @default.
• W3186095502 cites W1602117322 @default.
• W3186095502 cites W1807286188 @default.
• W3186095502 cites W1840435438 @default.
• W3186095502 cites W1987971958 @default.
• W3186095502 cites W1988709327 @default.
• W3186095502 cites W2010746547 @default.
• W3186095502 cites W2017966270 @default.
• W3186095502 cites W2031649430 @default.
• W3186095502 cites W2062500851 @default.
• W3186095502 cites W2064675550 @default.
• W3186095502 cites W2095705004 @default.
• W3186095502 cites W2104218272 @default.
• W3186095502 cites W2109553965 @default.
• W3186095502 cites W2110515480 @default.
• W3186095502 cites W2119175660 @default.
• W3186095502 cites W2127374659 @default.
• W3186095502 cites W2148579496 @default.
• W3186095502 cites W2151967815 @default.
• W3186095502 cites W2160920329 @default.
• W3186095502 cites W2173959313 @default.
• W3186095502 cites W2250539671 @default.
• W3186095502 cites W2296073425 @default.
• W3186095502 cites W2305467646 @default.
• W3186095502 cites W2566111950 @default.
• W3186095502 cites W2611669587 @default.
• W3186095502 cites W2742643783 @default.
• W3186095502 cites W2794524118 @default.
• W3186095502 cites W2891340257 @default.
• W3186095502 cites W2906152891 @default.
• W3186095502 cites W2910243263 @default.
• W3186095502 cites W2912981918 @default.
• W3186095502 cites W2921387835 @default.
• W3186095502 cites W2963310665 @default.
• W3186095502 cites W2963341956 @default.
• W3186095502 cites W2963403868 @default.
• W3186095502 cites W2963544815 @default.
• W3186095502 cites W2965373594 @default.
• W3186095502 cites W2970161131 @default.
• W3186095502 cites W2970184386 @default.
• W3186095502 cites W2970476646 @default.
• W3186095502 cites W2990704537 @default.
• W3186095502 cites W2996428491 @default.
• W3186095502 cites W3003519915 @default.
• W3186095502 cites W3004561114 @default.
• W3186095502 cites W3011411500 @default.
• W3186095502 cites W3015468748 @default.
• W3186095502 cites W3033187248 @default.
• W3186095502 cites W3034850762 @default.
• W3186095502 cites W3035656956 @default.
• W3186095502 cites W3037273551 @default.
• W3186095502 cites W3082274269 @default.
• W3186095502 cites W3091565975 @default.
• W3186095502 cites W3098068882 @default.
• W3186095502 cites W3098371923 @default.
• W3186095502 cites W3098467034 @default.
• W3186095502 cites W3098683121 @default.
• W3186095502 cites W3099080236 @default.
• W3186095502 cites W3100748148 @default.
• W3186095502 cites W3101278968 @default.
• W3186095502 cites W3115314063 @default.
• W3186095502 cites W3118485687 @default.
• W3186095502 cites W3162012630 @default.
• W3186095502 cites W3168987555 @default.
• W3186095502 cites W3188691716 @default.
• W3186095502 cites W38422447 @default.
• W3186095502 cites W50626113 @default.
• W3186095502 cites W80679475 @default.
• W3186095502 doi "https://doi.org/10.18653/v1/2021.starsem-1.2" @default.
• W3186095502 hasPublicationYear "2021" @default.
• W3186095502 type Work @default.
• W3186095502 sameAs 3186095502 @default.
• W3186095502 citedByCount "2" @default.
• W3186095502 countsByYear W31860955022022 @default.
• W3186095502 crossrefType "proceedings-article" @default.
• W3186095502 hasAuthorship W3186095502A5009772358 @default.
• W3186095502 hasAuthorship W3186095502A5037647150 @default.
• W3186095502 hasAuthorship W3186095502A5046323511 @default.
• W3186095502 hasAuthorship W3186095502A5064336148 @default.
• W3186095502 hasBestOaLocation W31860955021 @default.
• W3186095502 hasConcept C119599485 @default.
• W3186095502 hasConcept C119857082 @default.
• W3186095502 hasConcept C127413603 @default.
• W3186095502 hasConcept C138496976 @default.
• W3186095502 hasConcept C154945302 @default.
• W3186095502 hasConcept C15744967 @default.
• W3186095502 hasConcept C165801399 @default.
• W3186095502 hasConcept C169760540 @default.
• W3186095502 hasConcept C169900460 @default.
• W3186095502 hasConcept C171606756 @default.
• W3186095502 hasConcept C180747234 @default.

• next