Convolutional neural network (CNN) is easily affected by backdoor injections, whose models perform normally on clean samples but produce specific outputs on poisoned ones. Most of the existing studies have focused on the effect of trigger feature changes of poisoned samples on model generalization in spatial domain. We focus on the mechanism of CNN memorize poisoned samples in frequency domain, and find that CNN generate generalization to poisoned samples by memorizing the frequency domain distribution of trigger changes. We also explore the influence of trigger perturbations in different frequency domain components on the generalization of poisoned models from visible and invisible backdoor attacks, and prove that high-frequency components are more susceptible to perturbations than low-frequency components. Based on the above fundings, we propose a universal invisible strategy for visible triggers, which can achieve trigger invisibility while maintaining raw attack performance. We also design a novel frequency domain backdoor attack method based on low-frequency semantic information, which can achieve 100\% attack accuracy on multiple models and multiple datasets, and can bypass multiple defenses.

Due to the scarcity of high-quality bilingual sentence pairs, some deep-learning-based machine translation algorithms cannot achieve better performance in low-resource machine translation. On this basis, we are committed to integrating the ideas of machine learning algorithm improvement and data augmentation, propose a novel multiloop incremental bootstrapping framework, and design the corresponding semi-supervised learning algorithm. This framework is a meta-frame independent of specific machine translation algorithms. This algorithm makes full use of bilingual seed data of appropriate scale and super-large-scale monolingual data to expand bilingual sentence pair data incrementally, and trains machine translation models step by step to improve the translation quality. The experimental results of neural machine translation on multiple language pairs prove that our proposed framework can make use of continuous monolingual data to raise itself. Its effectiveness is not only reflected in the easy implementation of state-of-the-art low-resource machine translation, but also in the practical option to quickly establish precise domain machine translation systems.

Vietnamese tokenization is a challenging basic issue, and the corresponding algorithms can be used in many applications of natural language processing. In this paper, we investigate the Vietnamese tokenization problem and propose a supervised ensemble learning (SEL) framework as well as a SEL-based tokenization (SELT) algorithm. Supported by the data structure of syllable-syllable frequency index, the SELT algorithm combines multiple weak tokenizers to form a strong tokenizer. Within the SEL framework, we also investigate the efficient construction problem of a weak tokenizer. We suggest two prediction methods to select a suitable dictionary, and efficiently implement two weak tokenizers by the simple dictionary-based tokenization algorithm. The experimental results show that the SELT algorithm integrating our weak tokenizers can achieve state-of-the-art performance in the Vietnamese tokenization task.

Word segmentation is a challenging issue, and the corresponding algorithms can be used in many applications of natural language processing. This paper addresses the problem of Vietnamese word segmentation, proposes a probabilistic ensemble learning (PEL) framework, and designs a novel PEL-based word segmentation (PELWS) algorithm. Supported by the data structure of syllable-syllable frequency index, the PELWS algorithm combines multiple weak segmenters to form a strong segmenter within the PEL framework. The experimental results show that the PELWS algorithm can achieve the state-of-the-art performance in the Vietnamese word segmentation task.

Through the investigation of email document structure, this paper proposes a multi-field learning (MFL) framework, which breaks the multi-field document Text Classification (TC) problem into several sub-document TC problems, and makes the final category prediction by weighted linear combination of several sub-document TC results. Many previous statistical TC algorithms can be easily rebuilt within the MFL framework via turning binary result to spamminess score, which is a real number and reflects the likelihood that the classified email is spam. The experimental results in the TREC spam track show that the performances of many TC algorithms can be improved within the MFL framework.