Вестник Волгоградского государственного университета. Серия 2. Языкознание1998-99112409-197910.15688/jvolsu2.2024.5.8Интерпретация метафорического языка: вызов искусственному интеллектуInterpreting Metaphorical Language: A Challenge to Artificial IntelligenceСкрынниковаИнна ВалериевнаСкрынниковаИнна ВалериевнаSkrynnikovaInnai.skrynnikova@volsu.ru0000-0002-2390-7866Volgograd State University (Volgograd, Russian Federation)Волгоградский государственный университет (Волгоград, Российская Федерация)27122024235991073001202424062024CC BY 4.0

Актуальность работы обусловлена тем, что в последние годы многочисленные исследования указывают на способность искусственного интеллекта (ИИ) генерировать и анализировать выражения естественного языка, однако вопрос о том, способен ли ИИ действительно интерпретировать человеческий язык, а не имитировать его понимание, до сих пор остается открытым. Дополнительную сложность для систем ИИ составляют метафоры как неотъемлемая часть человеческого языка, которые являются не только распространенной фигурой речи, но и преобладающим когнитивным механизмом человеческого мышления. На основе обзора результатов существующих исследований компьютерной лингвистики и смежных областей в статье выделен ряд проблем, связанных с интерпретацией небуквальных выражений языка большими языковыми моделями (LLM). Показано, что в науке нет четкого представления о способах обучения языковых моделей автоматическому распознаванию и пониманию метафор, способных приблизить их к уровню «интерпретационных компетенций» человека. Цель исследования – выявить возможные причины, препятствующие пониманию образного языка искусственными системами и обозначить возможные направления для решения указанной проблемы. Установлено, что основные барьеры на пути ИИ к человекоподобной интерпретации образного естественного языка обусловлены отсутствием физического тела, неспособностью мыслить по аналогии и делать инференции на основе здравого смысла, последние при этом могут быть охарактеризованы одновременно как результат и как когнитивный процесс при обработке и извлечении информации.

In recent years, numerous studies have pointed to the ability of artificial intelligence (AI) to generate and analyze expressions of natural language. However, the question of whether AI is capable of actually interpreting human language, rather than imitating its understanding, remains open. Metaphors, being an integral part of human language, as both a common figure of speech and the predominant cognitive mechanism of human reasoning, pose a considerable challenge to AI systems. Based on an overview of the existing studies findings in computational linguistics and related fields, the paper identifies a number of problems associated with the interpretation of non-literal expressions of language by large language models (LLM). It reveals that there is still no clear understanding of the methods for training language models to automatically recognize and interpret metaphors that would bring it closer to the level of human “interpretive competencies”. The purpose of the study is to identify possible reasons that hinder the understanding of figurative language by artificial systems and to outline possible directions for solving this problem. The study suggests that the main barriers to AI’s human-like interpretation of figurative natural language are the absence of a physical body, the inability to reason by analogy and make inferences based on common sense, the latter being both the result and the cognitive process in extracting and processing information. The author concludes that further improvement of the AI systems creative skills should be at the top of the research agenda in the coming years.

метафорический языкрассуждение на основе аналогииискусственный интеллектLLMинтерпретация метафорвоплощенное познаниеинференцияmetaphorical languageanalogical reasoningartificial intelligenceLLMmetaphor interpretationembodied cognitioninferenceChakrabarty T., Choi Y., Shwartz V., 2022. It’s Not Rocket Science: Interpreting Figurative Language in Narratives. Transactions of the Association for Computational Linguistics, vol. 10, pp. 589-606. DOI: https://doi.org/10.1162/tacl_a_00478Choi M. et al., 2021. MelBERT: Metaphor Detection via Contextualized Late Interaction Using Metaphorical Identification Theories. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pp. 1763-1773. DOI: https://doi.org/10.48550/arXiv.2104.13615Comsa I.-M., Eisenschlos J., Narayanan S., 2022. MiQA: A Benchmark for Inference on Metaphorical Questions. Proceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing (Vol. 2: Short Papers), pp. 373-381. DOI: https://doi.org/10.48550/arXiv.2210.07993Falkenhainer B., Forbus K.D., Gentner D., 1989. The Structure-Mapping Engine: Algorithm and Examples. Artificial Intelligence, vol. 41, iss. 1, pp.1-63. DOI: https://doi.org/10.1016/0004-3702(89)90077-5Fussell S., Moss M., 1998. Figurative Language in Emotional Communication. Fussell S.R., Kreuz R.J., eds. Social and Cognitive Approaches to Interpersonal Communication, pp.113-141.Gibbs R.W., 2021. Metaphorical Embodiment. Handbook of Embodied Psychology. Cham, Springer. DOI: https://doi.org/10.1007/978-3-030-78471-3_5Hofstadter D.R., Mitchell M., 1994. The Copycat Project: A Model of Mental Fluidity and Analogy-Making. Holyoak K.J., Barnden J.A., eds. Analogical Connections. Ablex Publishing, pp. 31-112.Holland O., Knight R., 2006. The Anthropomimetic Principle. Burn J., Wilson M., eds. Proceedings of the AISB06 Symposium on Biologically Inspired Robotics, pp. 1-8.Ilyin G.L., 2013. Sovremennoe myshlenie: ot logiki k analogii [Modern Thinking: From Logic to Analogy]. Obrazovatelnye tekhnologii [Educational Technologies], no. 2, pp. 21-33.Keefer L.A. et al., 2014. Embodied Metaphor and Abstract Problem Solving: Testing a Metaphoric Fit Hypothesis in the health domain. Journal of Experimental Social Psychology, vol. 55, pp. 12-20. DOI: https://doi.org/10.1016/j.jesp.2014.05.012Klebanov B.B., Leong Ch.W., Gutierrez E.D. et al., 2016. Semantic Classifications for Detection of Verb Metaphors. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Vol. 2: Short Papers). Stroudsburg, PA, Association for Computational Linguistics, pp. 101-106. DOI: https://doi.org/10.18653/v1/P16-2017Lakoff G., Johnson M., 2003. Metaphors We Live By. University of Chicago Press. 256 p.Lakoff G., Johnson M., 1999. Philosophy in the Flesh. Nova Iorque, A Member of the Persus Books Group, 640 p.Leong C.W., Klebanov B.B., Hamill C., Stemle E.W., 2020. A Report on the 2020 VUA and TOEFL MetaphorDetection Shared Task. Proceedings of the 2nd Workshop on Figurative Language Processing, pp. 18-29. DOI: 10.18653/v1/2020. figlang-1.3Liu E., Cui Ch., Zheng K., Neubig G., 2022. Testing the Ability of Language Models to Interpret Figurative Language. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, July. DOI: https://doi.org/10.48550/arXiv.2204.12632Migun Yu., 2020. Gipoteza metaforicheskoy otnositelnosti: priznat nelzya otvergnut [Metaphorical Relativity Hypothesis: Recognize Cannot Be Denied]. Psikhologicheskie issledovaniya [Psychological Studies], vol. 13, iss. 73. DOI: https://doi.org/10.54359/ps. v13i73.172Mio J.S, Katz A.N., 1996. Metaphor: Implications and Applications. Psychology Press. 292 p.Mitchell M., 2021. Abstraction and Analogy-Making in Artificial Intelligence. Annals of the New York Academy of Sciences, vol. 1501, iss. 1, pp. 79-101. DOI: https://doi.org/10.1111/nyas.14619Murry J.M., 1931. Countries of the Mind. London, W. Collins Sons & Co., Ltd. 206 p.Neidlein A., Wiesenbach Ph., Markert K., 2020. An Analysis of Language Models for Metaphor Recognition. Proceedings of the 28th International Conference on Computational Linguistics, Barcelona, Spain (Online). International Committee on Computational Linguistics, pp. 3722-3736. DOI: https;//doi.org/10.18653/v1/2020.coling-main.332Pavlus J., 2021. The Computer Scientist Training AI to Think With Analogies. Quanta Magazine, July 14. URL: https://www.quantamagazine.org/melanie-mitchell-trains-ai-to-think-with-analogies-20210714/?print=1Rasskin-Gutman D., 2009. Chess Metaphors: Artificial Intelligence and the Human Mind. The MIT Press. DOI: https://doi.org/10.7551/mitpress/7925.001.0001Ripley A., 2021. Artificial Intelligence Interprets Metaphors (Second Place). University Library Prize Finalists for First Year Seminars, iss. 23.URL: https://www.exhibit.xavier.edu/library_ prize/23Shutova E., 2011. Computational Approaches to Figurative Language. University of Cambridge. 217 p.Shutova E., 2015. Design and Evaluation of Metaphor Processing Systems. Computational Linguistics, vol. 41, iss. 4, pp. 579-623. DOI: https://doi.org/10.1162/COLI_a_00233Skrynnikova I.V., 2023. Metaforicheskiy vyzov iskusstvennomu intellektu [Metaphorical Challenge to Artificial Intelligence]. Lektorskiy V.A., ed. Soznanie, telo, intellekt i yazyk v epokhu kognitivnykh tekhnologiy [Consciousness, Body, Intelligence and Language in the Era of Cognitive Technologies]. Pyatigorsk, pp. 220-221. URL: https://mbil-conf.ru/uploads/docs/sbornik-tezisov-MBIL-2023.pdfSrivastava A. et al., 2022. Beyond the Imitation Game: Quantifying and Extrapolating the Capabilities of Language Models. Computation and Language. DOI: https://doi.org/10.48550/arXiv.2206.04615Steen G.J., Dorst A.G., Herrmann J.B. et al, 2010. A Method for Linguistic Metaphor Identification. From MIP to MIPVU. Converging Evidence in Language and Communication Research, John Benjamins. URL: http://www.benjamins.com/cgi-bin/t_bookview.cgi?bookid=CELCR%2014Veale T., Hao Y., 2008. A Fluid Knowledge Representation for Understanding and Generating Creative Metaphors. Proceedings of COLING 2008. Manchester, pp. 945-952.