paper-with-me

홈 › Papers

More Parameters? No Thanks!

2021-07-20 · Findings (ACL) 2021 8 · Zeeshan Khan, Kartheek Akella, Vinay P. Namboodiri, C V Jawahar

This work studies the long-standing problems of model capacity and negative interference in multilingual neural machine translation MNMT. We use network pruning techniques and observe that pruning 50-70% of the parameters from a trained MNMT model results only in a 0.29-1.98 drop in the BLEU score. Suggesting that there exist large redundancies even in MNMT models. These observations motivate us to use the redundant parameters and counter the interference problem efficiently. We propose a novel adaptation strategy, where we iteratively prune and retrain the redundant parameters of an MNMT to improve bilingual representations while retaining the multilinguality. Negative interference severely affects high resource languages, and our method alleviates it without any additional adapter modules. Hence, we call it parameter-free adaptation strategy, paving way for the efficient adaptation of MNMT. We demonstrate the effectiveness of our method on a 9 language MNMT trained on TED talks, and report an average improvement of +1.36 BLEU on high resource pairs. Code will be released here.

📄 PDF Abstract BibTeX arXiv:2107.09622

Code (1)

zeecoder606/PF-Adaptation 공식 구현 pytorch

Tasks

Learning Language specific modelsMachine TranslationTranslation

Methods 이 논문이 사용한 방법론

Pruning 설명 없음

Similar Papers 제목 키워드 기반

Legendre Decomposition for Tensors

2018-02-13 · NeurIPS 2018 12 · Mahito Sugiyama, Hiroyuki Nakahara, Koji Tsuda

We present a novel nonnegative tensor decomposition method, called Legendre decomposition, which factorizes an input tensor into a multiplicative combination of parameters. Thanks to the well-developed theory of informat…

Tensor Decomposition

Practical Black Box Hamiltonian Learning

2022-06-30 · Andi Gu, Lukasz Cincio, Patrick J. Coles

We study the problem of learning the parameters for the Hamiltonian of a quantum many-body system, given limited access to the system. In this work, we build upon recent approaches to Hamiltonian learning via derivative …

Genetic Algorithms for the Optimization of Diffusion Parameters in Content-Based Image Retrieval

2019-08-19 · Federico Magliani, Laura Sani, Stefano Cagnoni, Andrea Prati

Several computer vision and artificial intelligence projects are nowadays exploiting the manifold data distribution using, e.g., the diffusion process. This approach has produced dramatic improvements on the final perfor…

Content-Based Image RetrievalImage RetrievalRetrieval

Distributed Training and Optimization Of Neural Networks

2020-12-03 · Jean-Roch Vlimant, Junqi Yin

Deep learning models are yielding increasingly better performances thanks to multiple factors. To be successful, model may have large number of parameters or complex architectures and be trained on large dataset. This le…

AdaIN-Switchable CycleGAN for Efficient Unsupervised Low-Dose CT Denoising

2020-08-13 · Jawook Gu, Jong Chul Ye

Recently, deep learning approaches have been extensively studied for low-dose CT denoising thanks to its superior performance despite the fast computational time. In particular, cycleGAN has been demonstrated as a powerf…

Denoising