scholarly journals Regression Loss in Transformer-based Supervised Neural Machine Translation

2021 ◽  
Vol 16 (4) ◽  
Author(s):  
Dongxing Li ◽  
Zuying Luo
Keyword(s):  
Machine Translation ◽  
Loss Function ◽  
Likelihood Estimation ◽  
Classification Problem ◽  
Cross Entropy ◽  
Slight Reduction ◽  
Neural Machine Translation ◽  
Entropy Loss ◽  
Semantic Difference ◽  
Reference Distance

Transformer-based model has achieved human-level performance in supervised neural machine translation (SNMT), much better than the models based on recurrent neural networks (RNNs) or convolutional neural network (CNN). The original Transformer-based model is trained through maximum likelihood estimation (MLE), which regards the machine translation task as a multilabel classification problem and takes the sum of the cross entropy loss of all the target tokens as the loss function. However, this model assumes that token generation is partially independent, without realizing that tokens are the components of a sequence. To solve the problem, this paper proposes a semantic regression loss for Transformer training, treating the generated sequence as a global. Upon finding that the semantic difference is proportional to candidate-reference distance, the authors considered the machine translation problem as a multi-task problem, and took the linear combination of cross entropy loss and semantic regression loss as the overall loss function. The semantic regression loss was proved to significantly enhance SNMT performance, with a slight reduction in convergence speed.

scholarly journals Approximating the Gradient of Cross-Entropy Loss Function

IEEE Access ◽  
2020 ◽  
Vol 8 ◽  
pp. 111626-111635
Author(s):  
Li Li ◽  
Milos Doroslovacki ◽  
Murray H. Loew
Keyword(s):  
Loss Function ◽  
Cross Entropy ◽  
Entropy Loss

scholarly journals MPCE: A Maximum Probability Based Cross Entropy Loss Function for Neural Network Classification

IEEE Access ◽  
2019 ◽  
Vol 7 ◽  
pp. 146331-146341 ◽  
Author(s):  
Yangfan Zhou ◽  
Xin Wang ◽  
Mingchuan Zhang ◽  
Junlong Zhu ◽  
Ruijuan Zheng ◽  
...  
Keyword(s):  
Neural Network ◽  
Loss Function ◽  
Cross Entropy ◽  
Maximum Probability ◽  
Entropy Loss ◽  
Neural Network Classification

scholarly journals Shaping the learning landscape in neural networks around wide flat minima

2019 ◽  
Vol 117 (1) ◽  
pp. 161-170 ◽  
Author(s):  
Carlo Baldassi ◽  
Fabrizio Pittorino ◽  
Riccardo Zecchina
Keyword(s):  
Neural Networks ◽  
Loss Function ◽  
Critical Points ◽  
Learning Process ◽  
Numerical Study ◽  
Cross Entropy ◽  
Stochastic Gradient Descent ◽  
Neural Network Models ◽  
Entropy Loss ◽  
Error Loss

Learning in deep neural networks takes place by minimizing a nonconvex high-dimensional loss function, typically by a stochastic gradient descent (SGD) strategy. The learning process is observed to be able to find good minimizers without getting stuck in local critical points and such minimizers are often satisfactory at avoiding overfitting. How these 2 features can be kept under control in nonlinear devices composed of millions of tunable connections is a profound and far-reaching open question. In this paper we study basic nonconvex 1- and 2-layer neural network models that learn random patterns and derive a number of basic geometrical and algorithmic features which suggest some answers. We first show that the error loss function presents few extremely wide flat minima (WFM) which coexist with narrower minima and critical points. We then show that the minimizers of the cross-entropy loss function overlap with the WFM of the error loss. We also show examples of learning devices for which WFM do not exist. From the algorithmic perspective we derive entropy-driven greedy and message-passing algorithms that focus their search on wide flat regions of minimizers. In the case of SGD and cross-entropy loss, we show that a slow reduction of the norm of the weights along the learning process also leads to WFM. We corroborate the results by a numerical study of the correlations between the volumes of the minimizers, their Hessian, and their generalization performance on real data.

scholarly journals AttentionBased Deep Feature Fusion for the Scene Classification of HighResolution Remote Sensing Images

2019 ◽  
Vol 11 (17) ◽  
pp. 1996 ◽  
Author(s):  
Zhu ◽  
Yan ◽  
Mo ◽  
Liu
Keyword(s):  
Remote Sensing ◽  
Loss Function ◽  
Feature Fusion ◽  
Cross Entropy ◽  
Scene Classification ◽  
Remote Sensing Images ◽  
Graphic Processing Units ◽  
Entropy Loss ◽  
Deep Feature

Scene classification of highresolution remote sensing images (HRRSI) is one of the most important means of landcover classification. Deep learning techniques, especially the convolutional neural network (CNN) have been widely applied to the scene classification of HRRSI due to the advancement of graphic processing units (GPU). However, they tend to extract features from the whole images rather than discriminative regions. The visual attention mechanism can force the CNN to focus on discriminative regions, but it may suffer from the influence of intraclass diversity and repeated texture. Motivated by these problems, we propose an attention-based deep feature fusion (ADFF) framework that constitutes three parts, namely attention maps generated by Gradientweighted Class Activation Mapping (GradCAM), a multiplicative fusion of deep features and the centerbased cross-entropy loss function. First of all, we propose to make attention maps generated by GradCAM as an explicit input in order to force the network to concentrate on discriminative regions. Then, deep features derived from original images and attention maps are proposed to be fused by multiplicative fusion in order to consider both improved abilities to distinguish scenes of repeated texture and the salient regions. Finally, the centerbased cross-entropy loss function that utilizes both the cross-entropy loss and center loss function is proposed to backpropagate fused features so as to reduce the effect of intraclass diversity on feature representations. The proposed ADFF architecture is tested on three benchmark datasets to show its performance in scene classification. The experiments confirm that the proposed method outperforms most competitive scene classification methods with an average overall accuracy of 94% under different training ratios.

scholarly journals Building Outline Extraction Directly Using the U2-Net Semantic Segmentation Model from High-Resolution Aerial Images and a Comparison Study

2021 ◽  
Vol 13 (16) ◽  
pp. 3187
Author(s):  
Xinchun Wei ◽  
Xing Li ◽  
Wei Liu ◽  
Lianpeng Zhang ◽  
Dayu Cheng ◽  
...  
Keyword(s):  
Edge Detection ◽  
Loss Function ◽  
Semantic Segmentation ◽  
Cross Entropy ◽  
Aerial Images ◽  
Building Extraction ◽  
Precise Position ◽  
Entropy Loss ◽  
Imbalance Problem ◽  
Outline Extraction

Deep learning techniques have greatly improved the efficiency and accuracy of building extraction using remote sensing images. However, high-quality building outline extraction results that can be applied to the field of surveying and mapping remain a significant challenge. In practice, most building extraction tasks are manually executed. Therefore, an automated procedure of a building outline with a precise position is required. In this study, we directly used the U2-net semantic segmentation model to extract the building outline. The extraction results showed that the U2-net model can provide the building outline with better accuracy and a more precise position than other models based on comparisons with semantic segmentation models (Segnet, U-Net, and FCN) and edge detection models (RCF, HED, and DexiNed) applied for two datasets (Nanjing and Wuhan University (WHU)). We also modified the binary cross-entropy loss function in the U2-net model into a multiclass cross-entropy loss function to directly generate the binary map with the building outline and background. We achieved a further refined outline of the building, thus showing that with the modified U2-net model, it is not necessary to use non-maximum suppression as a post-processing step, as in the other edge detection models, to refine the edge map. Moreover, the modified model is less affected by the sample imbalance problem. Finally, we created an image-to-image program to further validate the modified U2-net semantic segmentation model for building outline extraction.

scholarly journals Maximum Expected Likelihood Estimation for Zero-resource Neural Machine Translation

2017 ◽  
Author(s):  
Hao Zheng ◽  
Yong Cheng ◽  
Yang Liu
Keyword(s):  
Machine Translation ◽  
Likelihood Estimation ◽  
Neural Machine Translation ◽  
Parallel Corpora ◽  
Translation Model ◽  
Parallel Corpus ◽  
Target Model ◽  
High Resource ◽  
Remarkable Progress

While neural machine translation (NMT) has made remarkable progress in translating a handful of high-resource language pairs recently, parallel corpora are not always available for many zero-resource language pairs. To deal with this problem, we propose an approach to zero-resource NMT via maximum expected likelihood estimation. The basic idea is to maximize the expectation with respect to a pivot-to-source translation model for the intended source-to-target model on a pivot-target parallel corpus. To approximate the expectation, we propose two methods to connect the pivot-to-source and source-to-target models. Experiments on two zero-resource language pairs show that the proposed approach yields substantial gains over baseline methods. We also observe that when trained jointly with the source-to-target model, the pivot-to-source translation model also obtains improvements over independent training.

scholarly journals Ranking Loss: Maximizing the Success Rate in Deep Learning Side-Channel Analysis

2020 ◽  
pp. 25-55
Author(s):  
Gabriel Zaid ◽  
Lilian Bossuet ◽  
François Dassance ◽  
Amaury Habrard ◽  
Alexandre Venelli
Keyword(s):  
Deep Learning ◽  
Success Rate ◽  
Loss Function ◽  
Estimation Error ◽  
Approximation Error ◽  
Cross Entropy ◽  
Side Channel ◽  
Entropy Loss ◽  
The Cross ◽  
Ranking Loss

The side-channel community recently investigated a new approach, based on deep learning, to significantly improve profiled attacks against embedded systems. Compared to template attacks, deep learning techniques can deal with protected implementations, such as masking or desynchronization, without substantial preprocessing. However, important issues are still open. One challenging problem is to adapt the methods classically used in the machine learning field (e.g. loss function, performance metrics) to the specific side-channel context in order to obtain optimal results. We propose a new loss function derived from the learning to rank approach that helps preventing approximation and estimation errors, induced by the classical cross-entropy loss. We theoretically demonstrate that this new function, called Ranking Loss (RkL), maximizes the success rate by minimizing the ranking error of the secret key in comparison with all other hypotheses. The resulting model converges towards the optimal distinguisher when considering the mutual information between the secret and the leakage. Consequently, the approximation error is prevented. Furthermore, the estimation error, induced by the cross-entropy, is reduced by up to 23%. When the ranking loss is used, the convergence towards the best solution is up to 23% faster than a model using the cross-entropy loss function. We validate our theoretical propositions on public datasets.

scholarly journals Minimizing the Bag-of-Ngrams Difference for Non-Autoregressive Neural Machine Translation

2020 ◽  
Vol 34 (01) ◽  
pp. 198-205
Author(s):  
Chenze Shao ◽  
Jinchao Zhang ◽  
Yang Feng ◽  
Fandong Meng ◽  
Jie Zhou
Keyword(s):  
Machine Translation ◽  
Cross Entropy ◽  
Sequential Dependency ◽  
Weak Correlation ◽  
Neural Machine Translation ◽  
Translation Quality ◽  
Word Level ◽  
Translation Errors ◽  
Training Objective ◽  
Target Side

Non-Autoregressive Neural Machine Translation (NAT) achieves significant decoding speedup through generating target words independently and simultaneously. However, in the context of non-autoregressive translation, the word-level cross-entropy loss cannot model the target-side sequential dependency properly, leading to its weak correlation with the translation quality. As a result, NAT tends to generate influent translations with over-translation and under-translation errors. In this paper, we propose to train NAT to minimize the Bag-of-Ngrams (BoN) difference between the model output and the reference sentence. The bag-of-ngrams training objective is differentiable and can be efficiently calculated, which encourages NAT to capture the target-side sequential dependency and correlates well with the translation quality. We validate our approach on three translation tasks and show that our approach largely outperforms the NAT baseline by about 5.0 BLEU scores on WMT14 En↔De and about 2.5 BLEU scores on WMT16 En↔Ro.

scholarly journals Sequence-Level Training for Non-Autoregressive Neural Machine Translation

2021 ◽  
pp. 1-36
Author(s):  
Chenze Shao ◽  
Yang Feng ◽  
Jinchao Zhang ◽  
Fandong Meng ◽  
Jie Zhou
Keyword(s):  
Machine Translation ◽  
Cross Entropy ◽  
Gradient Estimation ◽  
Neural Machine Translation ◽  
Training Strategy ◽  
Word Generation ◽  
Translation Quality ◽  
Word Level ◽  
Training Objective ◽  
Level Training

Abstract In recent years, Neural Machine Translation (NMT) has achieved notable results in various translation tasks. However, the word-by-word generation manner determined by the autoregressive mechanism leads to high translation latency of the NMT and restricts its low-latency applications. Non-Autoregressive Neural Machine Translation (NAT) removes the autoregressive mechanism and achieves significant decoding speedup through generating target words independently and simultaneously. Nevertheless, NAT still takes the word-level cross-entropy loss as the training objective, which is not optimal because the output of NAT cannot be properly evaluated due to the multimodality problem. In this article, we propose using sequence-level training objectives to train NAT models, which evaluate the NAT outputs as a whole and correlates well with the real translation quality. Firstly, we propose training NAT models to optimize sequence-level evaluation metrics (e.g., BLEU) based on several novel reinforcement algorithms customized for NAT, which outperforms the conventional method by reducing the variance of gradient estimation. Secondly, we introduce a novel training objective for NAT models, which aims to minimize the Bag-of-Ngrams (BoN) difference between the model output and the reference sentence. The BoN training objective is differentiable and can be calculated efficiently without doing any approximations. Finally, we apply a three-stage training strategy to combine these two methods to train the NAT model.We validate our approach on four translation tasks (WMT14 En↔De, WMT16 En↔Ro), which shows that our approach largely outperforms NAT baselines and achieves remarkable performance on all translation tasks. The source code is available at https://github.com/ictnlp/Seq-NAT.

Sign in / Sign up

Export Citation Format

Share Document