{"url":"/method/simadapter","slug":"simadapter","name":"SimAdapter","full_name":"SimAdapter","full_name_withheld":false,"description_markdown":"**SimAdapter** is a module for explicitly learning knowledge from adapters. SimAdapter aims to learn the similarities between the source and target languages during fine-tuning using the adapters, and the similarity is based on an [attention mechanism](https://paperswithcode.com/methods/category/attention-mechanisms-1). \r\n\r\nThe detailed composition of the SimAdapter is shown in the Figure. By taking the language-agnostic representations from the backbone model as the query, and the language-specific outputs from multiple adapter as the keys and values, the final output for SimAdapter over attention are computed as (For notation simplicity, we omit the layer index $l$ below):\r\n\r\n$$\r\n\\operatorname{SimAdapter}\\left(\\mathbf{z}, \\mathbf{a}\\_{\\left\\(S\\_{1}, S\\_{2}, \\ldots, S\\_{N}\\right\\)}\\right)=\\sum_{i=1}^{N} \\operatorname{Attn}\\left(\\mathbf{z}, \\mathbf{a}\\_{S\\_{i}}\\right) \\cdot\\left(\\mathbf{a}\\_{S\\_{i}} \\mathbf{W}\\_{V}\\right)\r\n$$\r\n\r\nwhere SimAdapter $(\\cdot)$ and $\\operatorname{Attn}(\\cdot)$ denotes the SimAdapter and attention operations, respectively. Specifically, the attention operation is computed as:\r\n\r\n$$\r\n\\operatorname{Attn}(\\mathbf{z}, \\mathbf{a})=\\operatorname{Softmax}\\left(\\frac{\\left(\\mathbf{z} \\mathbf{W}\\_{Q}\\right)\\left(\\mathbf{a} \\mathbf{W}\\_{K}\\right)^{\\top}}{\\tau}\\right)\r\n$$\r\n\r\nwhere $\\tau$ is the temperature coefficient, $\\mathbf{W}\\_{Q}, \\mathbf{W}\\_{K}, \\mathbf{W}\\_{V}$ are attention matrices. Note that while $\\mathbf{W}\\_{Q}, \\mathbf{W}\\_{K}$ are initialized randomly, $\\mathbf{W}\\_{V}$ is initialized with a diagonal of ones and the rest of the matrix with small weights $(1 e-6)$ to retain the adapter representations. Furthermore, a regularization term is introduced to avoid drastic feature changes:\r\n\r\n$$\r\n\\mathcal{L}\\_{\\mathrm{reg}}=\\sum\\_{i, j}\\left(\\left(\\mathbf{I}\\_{V}\\right)\\_{i, j}-\\left(\\mathbf{W}\\_{V}\\right)_{i, j}\\right)^{2}\r\n$$\r\n\r\nwhere $\\mathbf{I}\\_{V}$ is the identity matrix with the same size as $\\mathbf{W}\\_{V}$","description_state":"present","introduced_year":null,"introduced_by":{"title":"Exploiting Adapters for Cross-lingual Low-resource Speech Recognition","paper":"/paper/exploiting-adapters-for-cross-lingual-low","first_author":"Wenxin Hou","n_authors":7,"url_abs":null,"archive_paper_url":"https://paperswithcode.com/paper/exploiting-adapters-for-cross-lingual-low"},"source":{"url":"https://arxiv.org/abs/2105.11905v2","title":"Exploiting Adapters for Cross-lingual Low-resource Speech Recognition","url_on_a_paper_host":true},"code_snippet_url":null,"code_snippet_url_on_a_code_host":false,"categories":[{"area":"General","area_id":"general","collection":"Attention Modules","url":"/methods/category/attention-modules","pwc_aliases":[]}],"n_papers_tagged":1,"archive_num_papers":1,"papers_newest_first":[{"paper":"/paper/exploiting-adapters-for-cross-lingual-low","title":"Exploiting Adapters for Cross-lingual Low-resource Speech Recognition","date":"2021-05-18","arxiv_id":"2105.11905","n_code_links":2,"syntology":null}],"papers_shown":1,"tasks":[{"task":"/task/cross-lingual-asr","name":"Cross-Lingual ASR","papers":1},{"task":"/task/general-knowledge","name":"General Knowledge","papers":1},{"task":"/task/meta-learning","name":"Meta-Learning","papers":1},{"task":"/task/speech-recognition","name":"Speech Recognition","papers":1},{"task":"/task/speech-recognition-1","name":"speech-recognition","papers":1}],"tasks_shown":5,"n_tasks":5,"usage_by_year":[{"year":"2021","papers":1}],"row_source":"methods_table","archive":{"source":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","archive_url":"https://paperswithcode.com/method/simadapter"},"syntology_read_at":"2026-09-24T18:15:14+00:00"}