Evaluating Morphological Compositional Generalization in Large Language Models

• URL: http://arxiv.org/abs/2410.12656v2
• Date: Wed, 06 Nov 2024 14:14:58 GMT
• Title: Evaluating Morphological Compositional Generalization in Large Language Models
• Authors: Mete Ismayilzada, Defne Circi, Jonne Sälevä, Hale Sirin, Abdullatif Köksal, Bhuwan Dhingra, Antoine Bosselut, Lonneke van der Plas, Duygu Ataman,
• Abstract summary: We investigate the morphological generalization abilities of large language models (LLMs) through the lens of compositionality. We focus on agglutinative languages such as Turkish and Finnish. Our analysis shows that LLMs struggle with morphological compositional generalization particularly when applied to novel word roots. While models can identify individual morphological combinations better than chance, their performance lacks systematicity, leading to significant accuracy gaps compared to humans.
• Score: 17.507983593566223
• License:
• Abstract: Large language models (LLMs) have demonstrated significant progress in various natural language generation and understanding tasks. However, their linguistic generalization capabilities remain questionable, raising doubts about whether these models learn language similarly to humans. While humans exhibit compositional generalization and linguistic creativity in language use, the extent to which LLMs replicate these abilities, particularly in morphology, is under-explored. In this work, we systematically investigate the morphological generalization abilities of LLMs through the lens of compositionality. We define morphemes as compositional primitives and design a novel suite of generative and discriminative tasks to assess morphological productivity and systematicity. Focusing on agglutinative languages such as Turkish and Finnish, we evaluate several state-of-the-art instruction-finetuned multilingual models, including GPT-4 and Gemini. Our analysis shows that LLMs struggle with morphological compositional generalization particularly when applied to novel word roots, with performance declining sharply as morphological complexity increases. While models can identify individual morphological combinations better than chance, their performance lacks systematicity, leading to significant accuracy gaps compared to humans.

Related papers

• Large Language Models as Neurolinguistic Subjects: Identifying Internal Representations for Form and Meaning [49.60849499134362]
  This study investigates the linguistic understanding of Large Language Models (LLMs) regarding signifier (form) and signified (meaning) Traditional psycholinguistic evaluations often reflect statistical biases that may misrepresent LLMs' true linguistic capabilities. We introduce a neurolinguistic approach, utilizing a novel method that combines minimal pair and diagnostic probing to analyze activation patterns across model layers.
  arXiv  Detail & Related papers  (2024-11-12T04:16:44Z)
• Analyzing The Language of Visual Tokens [48.62180485759458]
  We take a natural-language-centric approach to analyzing discrete visual languages. We show that higher token innovation drives greater entropy and lower compression, with tokens predominantly representing object parts. We also show that visual languages lack cohesive grammatical structures, leading to higher perplexity and weaker hierarchical organization compared to natural languages.
  arXiv  Detail & Related papers  (2024-11-07T18:59:28Z)
• Holmes: A Benchmark to Assess the Linguistic Competence of Language Models [59.627729608055006]
  We introduce Holmes, a new benchmark designed to assess language models (LMs) linguistic competence. We use computation-based probing to examine LMs' internal representations regarding distinct linguistic phenomena. As a result, we meet recent calls to disentangle LMs' linguistic competence from other cognitive abilities.
  arXiv  Detail & Related papers  (2024-04-29T17:58:36Z)
• LinguAlchemy: Fusing Typological and Geographical Elements for Unseen Language Generalization [35.12566667582262]
  LinguAlchemy is a regularization method that incorporates various linguistic information covering typological, geographical, and phylogenetic features. Our LinguAlchemy significantly improves the performance of mBERT and XLM-R on low-resource languages.
  arXiv  Detail & Related papers  (2024-01-11T16:48:00Z)
• Explicit Morphological Knowledge Improves Pre-training of Language Models for Hebrew [19.4968960182412]
  We investigate the hypothesis that incorporating explicit morphological knowledge in the pre-training phase can improve the performance of PLMs for morphologically rich languages. We propose various morphologically driven tokenization methods enabling the model to leverage morphological cues beyond raw text. Our experiments show that morphologically driven tokenization demonstrates improved results compared to a standard language-agnostic tokenization.
  arXiv  Detail & Related papers  (2023-11-01T17:02:49Z)
• A Survey of Large Language Models [81.06947636926638]
  Language modeling has been widely studied for language understanding and generation in the past two decades. Recently, pre-trained language models (PLMs) have been proposed by pre-training Transformer models over large-scale corpora. To discriminate the difference in parameter scale, the research community has coined the term large language models (LLM) for the PLMs of significant size.
  arXiv  Detail & Related papers  (2023-03-31T17:28:46Z)
• Cross-Lingual Transfer of Cognitive Processing Complexity [11.939409227407769]
  We use sentence-level eye-tracking patterns as a cognitive indicator for structural complexity. We show that the multilingual model XLM-RoBERTa can successfully predict varied patterns for 13 typologically diverse languages.
  arXiv  Detail & Related papers  (2023-02-24T15:48:23Z)
• Language Embeddings Sometimes Contain Typological Generalizations [0.0]
  We train neural models for a range of natural language processing tasks on a massively multilingual dataset of Bible translations in 1295 languages. The learned language representations are then compared to existing typological databases as well as to a novel set of quantitative syntactic and morphological features. We conclude that some generalizations are surprisingly close to traditional features from linguistic typology, but that most models, as well as those of previous work, do not appear to have made linguistically meaningful generalizations.
  arXiv  Detail & Related papers  (2023-01-19T15:09:59Z)
• Morphology Matters: A Multilingual Language Modeling Analysis [8.791030561752384]
  Prior studies disagree on whether inflectional morphology makes languages harder to model. We compile a larger corpus of 145 Bible translations in 92 languages and a larger number of typological features. Several morphological measures are significantly associated with higher surprisal when LSTM models are trained with BPE-segmented data.
  arXiv  Detail & Related papers  (2020-12-11T11:55:55Z)
• Linguistic Typology Features from Text: Inferring the Sparse Features of World Atlas of Language Structures [73.06435180872293]
  We construct a recurrent neural network predictor based on byte embeddings and convolutional layers. We show that some features from various linguistic types can be predicted reliably.
  arXiv  Detail & Related papers  (2020-04-30T21:00:53Z)
• Evaluating Transformer-Based Multilingual Text Classification [55.53547556060537]
  We argue that NLP tools perform unequally across languages with different syntactic and morphological structures. We calculate word order and morphological similarity indices to aid our empirical study.
  arXiv  Detail & Related papers  (2020-04-29T03:34:53Z)

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.