Controlled Evaluation of Syntactic Knowledge in Multilingual Language Models

• URL: http://arxiv.org/abs/2411.07474v1
• Date: Tue, 12 Nov 2024 01:26:41 GMT
• Title: Controlled Evaluation of Syntactic Knowledge in Multilingual Language Models
• Authors: Daria Kryvosheieva, Roger Levy,
• Abstract summary: This study develops targeted syntactic evaluation tests for three low-resource languages. We use them to evaluate five families of open-access multilingual Transformer LMs. We find that some syntactic tasks prove relatively easy for LMs while others are challenging.
• Score: 16.414150004715367
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Language models (LMs) are capable of acquiring elements of human-like syntactic knowledge. Targeted syntactic evaluation tests have been employed to measure how well they form generalizations about syntactic phenomena in high-resource languages such as English. However, we still lack a thorough understanding of LMs' capacity for syntactic generalizations in low-resource languages, which are responsible for much of the diversity of syntactic patterns worldwide. In this study, we develop targeted syntactic evaluation tests for three low-resource languages (Basque, Hindi, and Swahili) and use them to evaluate five families of open-access multilingual Transformer LMs. We find that some syntactic tasks prove relatively easy for LMs while others (agreement in sentences containing indirect objects in Basque, agreement across a prepositional phrase in Swahili) are challenging. We additionally uncover issues with publicly available Transformers, including a bias toward the habitual aspect in Hindi in multilingual BERT and underperformance compared to similar-sized models in XGLM-4.5B.

Related papers

• Who Relies More on World Knowledge and Bias for Syntactic Ambiguity Resolution: Humans or LLMs? [2.3749120526936465]
  This study explores how recent large language models (LLMs) navigate relative clause attachment ambiguity in six typologically diverse languages.
  arXiv  Detail & Related papers  (2025-03-13T19:44:15Z)
• Understanding and Mitigating Language Confusion in LLMs [76.96033035093204]
  We evaluate 15 typologically diverse languages with existing and newly-created English and multilingual prompts. We find that Llama Instruct and Mistral models exhibit high degrees of language confusion. We find that language confusion can be partially mitigated via few-shot prompting, multilingual SFT and preference tuning.
  arXiv  Detail & Related papers  (2024-06-28T17:03:51Z)
• Crosslingual Capabilities and Knowledge Barriers in Multilingual Large Language Models [62.91524967852552]
  Large language models (LLMs) are typically multilingual due to pretraining on diverse multilingual corpora. But can these models relate corresponding concepts across languages, effectively being crosslingual? This study evaluates six state-of-the-art LLMs on inherently crosslingual tasks.
  arXiv  Detail & Related papers  (2024-06-23T15:15:17Z)
• OMGEval: An Open Multilingual Generative Evaluation Benchmark for Large Language Models [59.54423478596468]
  We introduce OMGEval, the first Open-source Multilingual Generative test set that can assess the capability of LLMs in different languages. For each language, OMGEval provides 804 open-ended questions, covering a wide range of important capabilities of LLMs. Specifically, the current version of OMGEval includes 5 languages (i.e., Zh, Ru, Fr, Es, Ar)
  arXiv  Detail & Related papers  (2024-02-21T04:42:41Z)
• Evaluating Neural Language Models as Cognitive Models of Language Acquisition [4.779196219827507]
  We argue that some of the most prominent benchmarks for evaluating the syntactic capacities of neural language models may not be sufficiently rigorous. When trained on small-scale data modeling child language acquisition, the LMs can be readily matched by simple baseline models. We conclude with suggestions for better connecting LMs with the empirical study of child language acquisition.
  arXiv  Detail & Related papers  (2023-10-31T00:16:17Z)
• Are Structural Concepts Universal in Transformer Language Models? Towards Interpretable Cross-Lingual Generalization [27.368684663279463]
  We investigate the potential for explicitly aligning conceptual correspondence between languages to enhance cross-lingual generalization. Using the syntactic aspect of language as a testbed, our analyses of 43 languages reveal a high degree of alignability. We propose a meta-learning-based method to learn to align conceptual spaces of different languages.
  arXiv  Detail & Related papers  (2023-10-19T14:50:51Z)
• Democratizing LLMs for Low-Resource Languages by Leveraging their English Dominant Abilities with Linguistically-Diverse Prompts [75.33019401706188]
  Large language models (LLMs) are known to effectively perform tasks by simply observing few exemplars. We propose to assemble synthetic exemplars from a diverse set of high-resource languages to prompt the LLMs to translate from any language into English. Our unsupervised prompting method performs on par with supervised few-shot learning in LLMs of different sizes for translations between English and 13 Indic and 21 African low-resource languages.
  arXiv  Detail & Related papers  (2023-06-20T08:27:47Z)
• Assessing the Syntactic Capabilities of Transformer-based Multilingual Language Models [4.590895888766304]
  We explore the syntactic generalization capabilities of the monolingual and multilingual versions of BERT and RoBERTa. More specifically, we evaluate the syntactic generalization potential of the models on English and Spanish tests. For English, we use the available SyntaxGym test suite; for Spanish, we introduce SyntaxGymES, a novel ensemble of targeted syntactic tests in Spanish.
  arXiv  Detail & Related papers  (2021-05-10T22:06:53Z)
• AM2iCo: Evaluating Word Meaning in Context across Low-ResourceLanguages with Adversarial Examples [51.048234591165155]
  We present AM2iCo, Adversarial and Multilingual Meaning in Context. It aims to faithfully assess the ability of state-of-the-art (SotA) representation models to understand the identity of word meaning in cross-lingual contexts. Results reveal that current SotA pretrained encoders substantially lag behind human performance.
  arXiv  Detail & Related papers  (2021-04-17T20:23:45Z)
• Finding Universal Grammatical Relations in Multilingual BERT [47.74015366712623]
  We show that subspaces of mBERT representations recover syntactic tree distances in languages other than English. We present an unsupervised analysis method that provides evidence mBERT learns representations of syntactic dependency labels.
  arXiv  Detail & Related papers  (2020-05-09T20:46:02Z)
• Cross-Linguistic Syntactic Evaluation of Word Prediction Models [25.39896327641704]
  We investigate how neural word prediction models' ability to learn syntax varies by language. CLAMS includes subject-verb agreement challenge sets for English, French, German, Hebrew and Russian. We use CLAMS to evaluate LSTM language models as well as monolingual and multilingual BERT.
  arXiv  Detail & Related papers  (2020-05-01T02:51:20Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.