SANTM: Efficient Self-attention-driven Network for Text Matching

Self-attention mechanisms have recently been embraced for a broad range of text-matching applications. Self-attention model takes only one sentence as an input with no extra information, i.e., one can utilize the final hidden state or pooling. However, text-matching problems can be interpreted either in symmetrical or asymmetrical scopes. For instance, paraphrase detection is an asymmetrical task, while textual entailment classification and question-answer matching are considered asymmetrical tasks. In this article, we leverage attractive properties of self-attention mechanism and proposes an attention-based network that incorporates three key components for inter-sequence attention: global pointwise features, preceding attentive features, and contextual features while updating the rest of the components. Our model follows evaluation on two benchmark datasets cover tasks of textual entailment and question-answer matching. The proposed efficient Self-attention-driven Network for Text Matching outperforms the state of the art on the Stanford Natural Language Inference and WikiQA datasets with much fewer parameters.

Pembangkit Pertanyaan Otomatis pada Materi Pelajaran Ilmu Pengetahuan Alam Berbahasa Indonesia di Tingkat Sekolah Dasar Berdasarkan Revisi Taksonomi Bloom

Pembuatan pertanyaan pada suatu ujian merupakan proses yang kompleks, dikarenakan proses ini membutuhkan pengetahuan dan waktu yang lama dalam penyusunannya. Penyusunan pertanyaan dapat dilakukan dengan lebih mudah, cepat, dan terstruktur dengan adanya sistem Automatic Question Generator (AQG). Aplikasi ini memanfaatkan Metode Text Matching untuk menemukan kata kunci pada suatu paragraf, dimana kata kunci ini akan diidentifikasi menggunakan Metode Expected Answer Type (EAT). Metode EAT membantu untuk mengidentifikai jenis jawaban pada suatu paragraf sehingga dapat diketahui jenis pertanyaan yang akan di generate. Jenis pertanyaan yang digunakan yaitu 5W + 1H yang terdiri dari Siapa, Dimana, Kapan, Mengapa, Apa, Bagaimana, dan Berapa Banyak. Metode selanjutnya adalah Metode Template Based yang berperan dalam menyusun kalimat pertanyaan berdasarkan template yang sudah didaftarakan sebelumnya. Pertanyaan yang dihasilkan menggunakan konsep Revisi Taksonomi Bloom, dimana pertanyaan ini terdiri dari kategori (1) mengingat (2) memahami (3) mengaplikasikan (4) menganalisis (5) mengevaluasi dan (6) mencipta. Hasil uji coba dari 14 materi pembelajaran, aplikasi dapat meghasilkan 826 pertanyaan dengan tingkat rata-rata akurasi sebesar 89%.

Pseudo-siamese networks with lexicon for Chinese short text matching

Short text matching is one of the fundamental technologies in natural language processing. In previous studies, most of the text matching networks are initially designed for English text. The common approach to applying them to Chinese is segmenting each sentence into words, and then taking these words as input. However, this method often results in word segmentation errors. Chinese short text matching faces the challenges of constructing effective features and understanding the semantic relationship between two sentences. In this work, we propose a novel lexicon-based pseudo-siamese model (CL2 N), which can fully mine the information expressed in Chinese text. Instead of utilizing a character-sequence or a single word-sequence, CL2 N augments the text representation with multi-granularity information in characters and lexicons. Additionally, it integrates sentence-level features through single-sentence features as well as interactive features. Experimental studies on two Chinese text matching datasets show that our model has better performance than the state-of-the-art short text matching models, and the proposed method can solve the error propagation problem of Chinese word segmentation. Particularly, the incorporation of single-sentence features and interactive features allows the network to capture the contextual semantics and co-attentive lexical information, which contributes to our best result.

Research and Implementation of the Text Matching Algorithm in the Field of Housing Law and Policy Based on Deep Learning

Machine learning enables machines to learn rules from a large amount of data input from the outside world through algorithms, so as to identify and judge. It is the main task of the government to further emphasize the importance of improving the housing security mechanism, expand the proportion of affordable housing, increase financial investment, improve the construction quality of affordable housing, and ensure fair distribution. It can be seen that the legal system of housing security is essentially a system to solve the social problems brought by housing marketization, and it is an important part of the whole national housing system. More and more attention has been paid to solving the housing difficulties of low- and middle-income people and establishing a housing security legal system suitable for China’s national conditions and development stage. Aiming at the deep learning problem, a text matching algorithm suitable for the field of housing law and policy is proposed. Classifier based on matching algorithm is a promising classification technology. The research on the legal system of housing security is in the exploratory stage, involving various theoretical and practical research studies. Compare the improved depth learning algorithm with the general algorithm, so as to clearly understand the advantages and disadvantages of the improved depth learning algorithm and depth learning algorithm. This paper introduces the practical application of the deep learning model and fast learning algorithm in detail. Creatively put forward to transform it into an independent public law basis or into an independent savings system.

Context-Aware Text Matching Algorithm for Korean Peninsula Language Knowledge Base Based on Density Clustering

The majority of the traditional methods deal with text matching at the word level which remains uncertain as the text semantic features are ignored. This also leads to the problems of low recall and high space utilization of text matching while the comprehensiveness of matching results is poor. The resultant method, thus, cannot process long text and short text simultaneously. The current study proposes a text matching algorithm for Korean Peninsula language knowledge base based on density clustering. Using the deep multiview semantic document representation model, the semantic vector of the text to be matched is captured for semantic dependency which is utilized to extract the text semantic features. As per the feature extraction outcomes, the text similarity is calculated by subtree matching method, and a semantic classification model based on SWEM and pseudo-twin network is designed for semantic text classification. Finally, the text matching of Korean Peninsula language knowledge base is carried out by applying density clustering algorithm. Experimental results show that the proposed method has high matching recall rate with low space requirements and can effectively match long and short texts concurrently.