paper-with-me

홈 › Papers

SEARNN: Training RNNs with Global-Local Losses

2017-06-14 · ICLR 2018 1 · Rémi Leblond, Jean-Baptiste Alayrac, Anton Osokin, Simon Lacoste-Julien

We propose SEARNN, a novel training algorithm for recurrent neural networks (RNNs) inspired by the "learning to search" (L2S) approach to structured prediction. RNNs have been widely successful in structured prediction applications such as machine translation or parsing, and are commonly trained using maximum likelihood estimation (MLE). Unfortunately, this training loss is not always an appropriate surrogate for the test error: by only maximizing the ground truth probability, it fails to exploit the wealth of information offered by structured losses. Further, it introduces discrepancies between training and predicting (such as exposure bias) that may hurt test performance. Instead, SEARNN leverages test-alike search space exploration to introduce global-local losses that are closer to the test error. We first demonstrate improved performance over MLE on two different tasks: OCR and spelling correction. Then, we propose a subsampling strategy to enable SEARNN to scale to large vocabulary sizes. This allows us to validate the benefits of our approach on a machine translation task.

📄 PDF Abstract BibTeX arXiv:1706.04499

Code (1)

RemiLeblond/SeaRNN-open 공식 구현 pytorch

Tasks

Machine TranslationOptical Character Recognition (OCR)Spelling CorrectionStructured PredictionTranslation

Similar Papers 제목 키워드 기반

Beyond MLE: Investigating SEARNN for Low-Resourced Neural Machine Translation

2024-05-20 · Chris Emezue

Structured prediction tasks, like machine translation, involve learning functions that map structured inputs to structured outputs. Recurrent Neural Networks (RNNs) have historically been a popular choice for such tasks,…

Machine TranslationStructured PredictionTranslation

Decoupling Hierarchical Recurrent Neural Networks With Locally Computable Losses

2019-10-11 · Asier Mujika, Felix Weissenberger, Angelika Steger

Learning long-term dependencies is a key long-standing challenge of recurrent neural networks (RNNs). Hierarchical recurrent neural networks (HRNNs) have been considered a promising approach as long-term dependencies are…

Combining Global Sparse Gradients with Local Gradients in Distributed Neural Network Training

2019-11-01 · IJCNLP 2019 11 · Alham Fikri Aji, Kenneth Heafield, Nikolay Bogoychev

One way to reduce network traffic in multi-node data-parallel stochastic gradient descent is to only exchange the largest gradients. However, doing so damages the gradient and degrades the model{'}s performance. Transfor…

Machine TranslationTranslation

Deep Loss Convexification for Learning Iterative Models

2024-11-16 · Ziming Zhang, Yuping Shao, Yiqing Zhang, Fangzhou Lin 외

Iterative methods such as iterative closest point (ICP) for point cloud registration often suffer from bad local optimality (e.g. saddle points), due to the nature of nonconvex optimization. To address this fundamental c…

Point Cloud Registration

Multi-level Contextual RNNs with Attention Model for Scene Labeling

2016-07-08 · Heng Fan, Xue Mei, Danil Prokhorov, Haibin Ling

Context in image is crucial for scene labeling while existing methods only exploit local context generated from a small surrounding area of an image patch or a pixel, by contrast long-range and global contextual informat…

Scene Labeling