Multi-Sense Embeddings for Language Models and Knowledge Distillation

• URL: http://arxiv.org/abs/2504.06036v1
• Date: Tue, 08 Apr 2025 13:36:36 GMT
• Title: Multi-Sense Embeddings for Language Models and Knowledge Distillation
• Authors: Qitong Wang, Mohammed J. Zaki, Georgios Kollias, Vasileios Kalantzis,
• Abstract summary: Transformer-based large language models (LLMs) rely on contextual embeddings which generate different representations for the same token depending on its surrounding context.<n>We propose multi-sense embeddings as a drop-in replacement for each token in order to capture the range of their uses in a language.
• Score: 17.559171180573664
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Transformer-based large language models (LLMs) rely on contextual embeddings which generate different (continuous) representations for the same token depending on its surrounding context. Nonetheless, words and tokens typically have a limited number of senses (or meanings). We propose multi-sense embeddings as a drop-in replacement for each token in order to capture the range of their uses in a language. To construct a sense embedding dictionary, we apply a clustering algorithm to embeddings generated by an LLM and consider the cluster centers as representative sense embeddings. In addition, we propose a novel knowledge distillation method that leverages the sense dictionary to learn a smaller student model that mimics the senses from the much larger base LLM model, offering significant space and inference time savings, while maintaining competitive performance. Via thorough experiments on various benchmarks, we showcase the effectiveness of our sense embeddings and knowledge distillation approach. We share our code at https://github.com/Qitong-Wang/SenseDict

Related papers

• Tokenization is Sensitive to Language Variation [14.568179478275255]
  Tokenizers split texts into smaller units and might behave differently for less common linguistic forms.<n>This might affect downstream LLM performance differently on two types of tasks.<n>We investigate how key algorithmic design choices impact downstream models' performances.
  arXiv  Detail & Related papers  (2025-02-21T09:58:54Z)
• Large Concept Models: Language Modeling in a Sentence Representation Space [62.73366944266477]
  We present an attempt at an architecture which operates on an explicit higher-level semantic representation, which we name a concept.<n> Concepts are language- and modality-agnostic and represent a higher level idea or action in a flow.<n>We show that our model exhibits impressive zero-shot generalization performance to many languages.
  arXiv  Detail & Related papers  (2024-12-11T23:36:20Z)
• Word Sense Induction with Knowledge Distillation from BERT [6.88247391730482]
  This paper proposes a method to distill multiple word senses from a pre-trained language model (BERT) by using attention over the senses of a word in a context. Experiments on the contextual word similarity and sense induction tasks show that this method is superior to or competitive with state-of-the-art multi-sense embeddings.
  arXiv  Detail & Related papers  (2023-04-20T21:05:35Z)
• Beyond Contrastive Learning: A Variational Generative Model for Multilingual Retrieval [109.62363167257664]
  We propose a generative model for learning multilingual text embeddings. Our model operates on parallel data in $N$ languages. We evaluate this method on a suite of tasks including semantic similarity, bitext mining, and cross-lingual question retrieval.
  arXiv  Detail & Related papers  (2022-12-21T02:41:40Z)
• Learning an Artificial Language for Knowledge-Sharing in Multilingual Translation [15.32063273544696]
  We discretize the latent space of multilingual models by assigning encoder states to entries in a codebook. We validate our approach on large-scale experiments with realistic data volumes and domains. We also use the learned artificial language to analyze model behavior, and discover that using a similar bridge language increases knowledge-sharing among the remaining languages.
  arXiv  Detail & Related papers  (2022-11-02T17:14:42Z)
• Multilingual Word Sense Disambiguation with Unified Sense Representation [55.3061179361177]
  We propose building knowledge and supervised-based Multilingual Word Sense Disambiguation (MWSD) systems. We build unified sense representations for multiple languages and address the annotation scarcity problem for MWSD by transferring annotations from rich-sourced languages to poorer ones. Evaluations of SemEval-13 and SemEval-15 datasets demonstrate the effectiveness of our methodology.
  arXiv  Detail & Related papers  (2022-10-14T01:24:03Z)
• Integrating Language Guidance into Vision-based Deep Metric Learning [78.18860829585182]
  We propose to learn metric spaces which encode semantic similarities as embedding space. These spaces should be transferable to classes beyond those seen during training. This causes learned embedding spaces to encode incomplete semantic context and misrepresent the semantic relation between classes.
  arXiv  Detail & Related papers  (2022-03-16T11:06:50Z)
• More Than Words: Collocation Tokenization for Latent Dirichlet Allocation Models [71.42030830910227]
  We propose a new metric for measuring the clustering quality in settings where the models differ. We show that topics trained with merged tokens result in topic keys that are clearer, more coherent, and more effective at distinguishing topics than those unmerged models.
  arXiv  Detail & Related papers  (2021-08-24T14:08:19Z)
• Learning Contextualised Cross-lingual Word Embeddings and Alignments for Extremely Low-Resource Languages Using Parallel Corpora [63.5286019659504]
  We propose a new approach for learning contextualised cross-lingual word embeddings based on a small parallel corpus. Our method obtains word embeddings via an LSTM encoder-decoder model that simultaneously translates and reconstructs an input sentence.
  arXiv  Detail & Related papers  (2020-10-27T22:24:01Z)
• DeCLUTR: Deep Contrastive Learning for Unsupervised Textual Representations [4.36561468436181]
  We present DeCLUTR: Deep Contrastive Learning for Unsupervised Textual Representations. Our approach closes the performance gap between unsupervised and supervised pretraining for universal sentence encoders. Our code and pretrained models are publicly available and can be easily adapted to new domains or used to embed unseen text.
  arXiv  Detail & Related papers  (2020-06-05T20:00:28Z)

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.