Cross-lingual learning for text processing: A survey

Pikuliak, Matúš; Šimko, Marián; Bieliková, Mária

doi:10.1016/j.eswa.2020.113765

Cited by 34 publications

(13 citation statements)

References 124 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Of course, it will be interesting to verify if our approach is applicable to languages other than English. One of the possible approaches which could bring benefit in the future is to explore the utilization of transfer learning (Pikuliak, Simko, and Bielikova 2021).…”

Section: Discussionmentioning

confidence: 99%

Automatic question generation based on sentence structure analysis using machine learning approach

Blšták¹,

Rozinajová²

2021

Nat. Lang. Eng.

View full text Add to dashboard Cite

Automatic question generation is one of the most challenging tasks of Natural Language Processing. It requires “bidirectional” language processing: first, the system has to understand the input text (Natural Language Understanding), and it then has to generate questions also in the form of text (Natural Language Generation). In this article, we introduce our framework for generating the factual questions from unstructured text in the English language. It uses a combination of traditional linguistic approaches based on sentence patterns with several machine learning methods. We first obtain lexical, syntactic and semantic information from an input text, and we then construct a hierarchical set of patterns for each sentence. The set of features is extracted from the patterns, and it is then used for automated learning of new transformation rules. Our learning process is totally data-driven because the transformation rules are obtained from a set of initial sentence–question pairs. The advantages of this approach lie in a simple expansion of new transformation rules which allows us to generate various types of questions and also in the continuous improvement of the system by reinforcement learning. The framework also includes a question evaluation module which estimates the quality of generated questions. It serves as a filter for selecting the best questions and eliminating incorrect ones or duplicates. We have performed several experiments to evaluate the correctness of generated questions, and we have also compared our system with several state-of-the-art systems. Our results indicate that the quality of generated questions outperforms the state-of-the-art systems and our questions are also comparable to questions created by humans. We have also created and published an interface with all created data sets and evaluated questions, so it is possible to follow up on our work.

show abstract

Section: Discussionmentioning

confidence: 99%

Automatic question generation based on sentence structure analysis using machine learning approach

Blšták¹,

Rozinajová²

2021

Nat. Lang. Eng.

View full text Add to dashboard Cite

show abstract

“…TL methods aim at enhancing model performance based on transferring existing knowledge in the source domain to the target domain, which can reduce the dependency on the target domain dataset to some extent. Based on the above advantages, TL methods have been widely employed in medical image analysis [ 79 , 80 ], engineering [ 81 ], text processing [ 82 ], natural language processing [ 83 ] etc. Considering whether the existing dataset is labeled or not, TL can be classified into three categories, i.e., inductive, transductive and unsupervised transfer learning [ 77 ].…”

Section: Challenges and Solutionsmentioning

confidence: 99%

Belt Tear Detection for Coal Mining Conveyors

et al. 2022

View full text Add to dashboard Cite

The belt conveyor is the most commonly used conveying equipment in the coal mining industry. As the core part of the conveyor, the belt is vulnerable to various failures, such as scratches, cracks, wear and tear. Inspection and defect detection is essential for conveyor belts, both in academic research and industrial applications. In this paper, we discuss existing techniques used in industrial production and state-of-the-art theories for conveyor belt tear detection. First, the basic structure of conveyor belts is discussed and an overview of tear defect detection methods for conveyor belts is studied. Next, the causes of conveyor belt tear are classified, such as belt aging, scratches by sharp objects, abnormal load or a combination of the above reasons. Then, recent mainstream techniques and theories for conveyor belt tear detection are reviewed, and their characteristics, advantages and shortcomings are discussed. Furthermore, image dataset preparation and data imbalance problems are studied for belt defect detection. Moreover, the current challenges and opportunities for conveyor belt defect detection are discussed. Lastly, a case study was carried out to compare the detection performance of popular techniques using industrial image datasets. This paper provides professional guidelines and promising research directions for researchers and engineers based on the leading theories in machine vision and deep learning.

show abstract

“…Additional examples of downstream applications of cross-lingual learning are document classification (Holger and Xian, 2018), named entity recognition (Xie et al, 2018) and part-of-speech tagging (Cohen et al, 2011). For a thorough review on cross-lingual learning, we refer the reader to Pikuliak et al (2021).…”

Section: Related Workmentioning

confidence: 99%

Cross-Lingual Event Detection via Optimized Adversarial Training

Guzman-Nateras¹,

Nguyen²,

Nguyen³

2022

Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Langua

View full text Add to dashboard Cite

In this work, we focus on Cross-Lingual Event Detection where a model is trained on data from a source language but its performance is evaluated on data from a second, target, language. Most recent works in this area have harnessed the language-invariant qualities displayed by pre-trained Multi-lingual Language Models. Their performance, however, reveals there is room for improvement as the crosslingual setting entails particular challenges. We employ Adversarial Language Adaptation to train a Language Discriminator to discern between the source and target languages using unlabeled data. The discriminator is trained in an adversarial manner so that the encoder learns to produce refined, language-invariant representations that lead to improved performance. More importantly, we optimize the adversarial training process by only presenting the discriminator with the most informative samples. We base our intuition about what makes a sample informative on two disparate metrics: sample similarity and event presence. Thus, we propose leveraging Optimal Transport as a solution to naturally combine these two distinct information sources into the selection process. Extensive experiments on 8 different language pairs, using 4 languages from unrelated families, show the flexibility and effectiveness of our model that achieves state-of-the-art results.

show abstract

Cross-lingual learning for text processing: A survey

Cited by 34 publications

References 124 publications

Automatic question generation based on sentence structure analysis using machine learning approach

Automatic question generation based on sentence structure analysis using machine learning approach

Belt Tear Detection for Coal Mining Conveyors

Cross-Lingual Event Detection via Optimized Adversarial Training

Contact Info

Product

Resources

About