ACL Findings2026
When Meaning Travels: A Granular Lens on Hybrid-MoE's Role in Idiomatic Understanding for Language Models
Sarmistha Das, Vaibhav Vishal, Shreyas Guha, Amaan Ali, Kitsuchart Pasupa, Sriparna Saha
Abstract
In the contemporary epoch of multilingual education, learning idioms provides a fascinating gateway towards creativity, cultural values, historical context, and diverse perspectives inherent to various linguistic traditions. This paper showcases the navigation of retaining figurative and cultural semantics in low-resource Southeast Asian languages such as Hindi, Bengali, and Thai, where culturally rich idioms pose significant obstacles for computational modeling and cross-linguistic transfer due to their deep metaphorical complexity. To tackle such complexity, we present Varnika (वर्ि ण का), a reconstructed multimodal idiom corpus comprising 3,533 multilingual idioms, enriched with seven idiomatic tones aligned with both textual and visual representations. Additionally, to infer informative idiomatic understanding, we introduce a Hybrid Mixture-of-Experts (HybridMoE) framework that embeds multiple idiomatic expert opinions while mitigating expert sparsity by integrating outputs from both selected and unselected experts through controlled hybridization, further augmented with Idiomatic Property Signals via masked multimodal embeddings. To analyze the performance across multiple dimensions, we propose the IDIO-TONE and Idiomatic Validation Score, a three-stage evaluation pipeline measuring (i) literal translation fidelity, (ii) visualsemantic alignment, and (iii) idiomatic meaning retention. Empirical evaluations highlight that HybridMoE achieves 5-6% performance gains across advanced vision language models, demonstrating improved representation of figurative language and culturally embedded meaning in multilingual multimodal settings 1 .
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
Builds on9
- An Image is Worth 16x16 Words: Transformers for Image Recognition at ScaleAlexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn et al.ICLR 2021 · 21,477 citations
- BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language ModelsJunnan Li, Dongxu Li, Silvio Savarese, Steven C. H. HoiICML 2023 · 7,873 citations
- FLUTE: Figurative Language Understanding through Textual ExplanationsTuhin Chakrabarty, Arkadiy Saakyan, Debanjan Ghosh, Smaranda MuresanEMNLP 2022 · 35 citations
- Neural Simile Recognition with Cyclic Multitask Learning and Local AttentionJiali Zeng, Linfeng Song, Jinsong Su, Jun Xie et al.AAAI 2020 · 26 citations
- TituLLMs: A Family of Bangla LLMs with Comprehensive BenchmarkingShahriar Kabir Nahin, Rabindra Nath Nandi, Sagor Sarker, Quazi Sarwar Muhtaseem et al.ACL Findings 2025 · 20 citations
Related papers
- MePe: Rethinking Multimodal Chinese Idiom Reading Comprehension from a Metaphorical PerspectiveTongguan Wang, Junkai Li, Feiyue Xue, Hui Liu et al.WWW 2026
- SEA-Vision: A Multilingual Benchmark for Comprehensive Document and Scene Text Understanding in Southeast AsiaPengfei Yue, Xingran Zhao, Juntao Chen, Peng Hou et al.CVPR 2026 · 2 citations
- Graph-Assisted Culturally Adaptable Idiomatic Translation for Indic languagesPratik Rakesh Singh, Kritarth Prasad, Mohammadi Zaki, Pankaj WasnikACL Findings 2025 · 1 citation
- Express What You See: Can Multimodal LLMs Decode Visual Ciphers with Intuitive Semiosis Comprehension?Jiayi Kuang, Yinghui Li, Chen Wang, Haohao Luo et al.ACL Findings 2025
- Multilingual Idioms in Sentences and Conversations Across High-, Medium-, and Low-Resource LanguagesSaeed Almheiri, Bilal Elbouardi, Salsabila Zahirah Pranida, Irina Nikishina et al.ACL 2026
