TY - GEN
T1 - Do Neural Language Models Compose Concepts the Way Humans Can?
AU - Rodriguez, Amilleah
AU - Wang, Shaonan
AU - Pylkkanen, Liina
N1 - Publisher Copyright:
© 2024 ELRA Language Resource Association: CC BY-NC 4.0.
PY - 2024
Y1 - 2024
N2 - While compositional interpretation is the core of language understanding, humans also derive meaning via inference. For example, while the phrase “the blue hat” introduces a blue hat into the discourse via the direct composition of “blue” and “hat,” the same discourse entity is introduced by the phrase “the blue color of this hat” despite the absence of any local composition between “blue” and “hat.” Instead, we infer that if the color is blue and it belongs to the hat, the hat must be blue. We tested the performance of neural language models and humans on such inferentially driven conceptual compositions, eliciting probability estimates for a noun in a syntactically composing phrase, "This blue hat", following contexts that had introduced the conceptual combinations of those nouns and adjectives either syntactically or inferentially. Surprisingly, our findings reveal significant disparities between the performance of neural language models and human judgments. Among the eight models evaluated, RoBERTa, BERT-large, and GPT-2 exhibited the closest resemblance to human responses, while other models faced challenges in accurately identifying compositions in the provided contexts. Our study reveals that language models and humans may rely on different approaches to represent and compose lexical items across sentence structure. All data and code are accessible at https://github.com/wangshaonan/BlueHat.
AB - While compositional interpretation is the core of language understanding, humans also derive meaning via inference. For example, while the phrase “the blue hat” introduces a blue hat into the discourse via the direct composition of “blue” and “hat,” the same discourse entity is introduced by the phrase “the blue color of this hat” despite the absence of any local composition between “blue” and “hat.” Instead, we infer that if the color is blue and it belongs to the hat, the hat must be blue. We tested the performance of neural language models and humans on such inferentially driven conceptual compositions, eliciting probability estimates for a noun in a syntactically composing phrase, "This blue hat", following contexts that had introduced the conceptual combinations of those nouns and adjectives either syntactically or inferentially. Surprisingly, our findings reveal significant disparities between the performance of neural language models and human judgments. Among the eight models evaluated, RoBERTa, BERT-large, and GPT-2 exhibited the closest resemblance to human responses, while other models faced challenges in accurately identifying compositions in the provided contexts. Our study reveals that language models and humans may rely on different approaches to represent and compose lexical items across sentence structure. All data and code are accessible at https://github.com/wangshaonan/BlueHat.
KW - Composition
KW - Dataset Construction
KW - Inference
KW - Neural Language Models
UR - http://www.scopus.com/inward/record.url?scp=85195977666&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85195977666&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85195977666
T3 - 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation, LREC-COLING 2024 - Main Conference Proceedings
SP - 5309
EP - 5314
BT - 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation, LREC-COLING 2024 - Main Conference Proceedings
A2 - Calzolari, Nicoletta
A2 - Kan, Min-Yen
A2 - Hoste, Veronique
A2 - Lenci, Alessandro
A2 - Sakti, Sakriani
A2 - Xue, Nianwen
PB - European Language Resources Association (ELRA)
T2 - Joint 30th International Conference on Computational Linguistics and 14th International Conference on Language Resources and Evaluation, LREC-COLING 2024
Y2 - 20 May 2024 through 25 May 2024
ER -