Abstract:Deep sequence recognition (DSR) models receive increasing attention due to their superior application to various applications. Most DSR models use merely the target sequences as supervision without considering other related sequences, leading to over-confidence in their predictions. The DSR models trained with label smoothing regularize labels by equally and independently smoothing each token, reallocating a small value to other tokens for mitigating overconfidence. However, they do not consider tokens/sequences correlations that may provide more effective information to regularize training and thus lead to sub-optimal performance. In this work, we find tokens/sequences with high perception and semantic correlations with the target ones contain more correlated and effective information and thus facilitate more effective regularization. To this end, we propose a Perception and Semantic aware Sequence Regularization framework, which explore perceptively and semantically correlated tokens/sequences as regularization. Specifically, we introduce a semantic context-free recognition and a language model to acquire similar sequences with high perceptive similarities and semantic correlation, respectively. Moreover, over-confidence degree varies across samples according to their difficulties. Thus, we further design an adaptive calibration intensity module to compute a difficulty score for each samples to obtain finer-grained regularization. Extensive experiments on canonical sequence recognition tasks, including scene text and speech recognition, demonstrate that our method sets novel state-of-the-art results. Code is available at <a class="link-external link-https" href="https://github.com/husterpzh/PSSR" rel="external noopener nofollow">this https URL</a>.

Regularized Structured Perceptron: A Case Study on Chinese Word Segmentation, POS Tagging and Parsing.

STRAPPER: Preference-based Reinforcement Learning via Self-training Augmentation and Peer Regularization

Wordreg: Mitigating the Gap Between Training and Inference with Worst-Case Drop Regularization

A Comparative Study on Regularization Strategies for Embedding-based Neural Networks.

Network as Regularization for Training Deep Neural Networks: Framework, Model and Performance

Regularizing Deep Convolutional Neural Networks with a Structured Decorrelation Constraint.

62 43 v 2 [ cs . L G ] 30 J an 2 01 5 Structure Regularization for Structured Prediction : Theories and Experiments

Structure Regularization for Structured Prediction.

Structure Regularization for Structured Prediction: Theories and Experiments

Measuring and Reducing Model Update Regression in Structured Prediction for NLP

Improving Chinese Word Segmentation Using Partially Annotated Sentences

LDA-Reg: Knowledge Driven Regularization using External Corpora

Leveraging Part-of-Speech Tagging Features and a Novel Regularization Strategy for Chinese Medical Named Entity Recognition

SubRegWeigh: Effective and Efficient Annotation Weighing with Subword Regularization

Performance Improvement on Traditional Chinese Task-Oriented Dialogue Systems With Reinforcement Learning and Regularized Dropout Technique

Perception and Semantic Aware Regularization for Sequential Confidence Calibration

Complex Structure Leads to Overfitting: A Structure Regularization Decoding Method for Natural Language Processing

Effects of Common Regularization Techniques on Open-Set Recognition

Comparing effectiveness of regularization methods on text classification: Simple and complex model in data shortage situation

Improved Regularization Techniques for End-to-End Speech Recognition

Class Probability Space Regularization for Semi-Supervised Semantic Segmentation