Neural Machine Translation Models with Attention-Based Dropout Layer

In bilingual translation,attention-based Neural Machine Translation(NMT)models are used to achieve synchrony between input and output sequences and the notion of alignment.NMT model has obtained state-of-the-art performance for several language pairs.However,there has been little work exploring useful architectures for Urdu-to-English machine translation.We conducted extensive Urdu-to-English translation experiments using Long short-term memory(LSTM)/Bidirectional recurrent neural networks(Bi-RNN)/Statistical recurrent unit(SRU)/Gated recurrent unit(GRU)/Convolutional neural network(CNN)and Transformer.Experimental results show that Bi-RNN and LSTM with attention mechanism trained iteratively,with a scalable data set,make precise predictions on unseen data.The trained models yielded competitive results by achieving 62.6%and 61%accuracy and 49.67 and 47.14 BLEU scores,respectively.From a qualitative perspective,the translation of the test sets was examined manually,and it was observed that trained models tend to produce repetitive output more frequently.The attention score produced by Bi-RNN and LSTM produced clear alignment,while GRU showed incorrect translation for words,poor alignment and lack of a clear structure.Therefore,we considered refining the attention-based models by defining an additional attention-based dropout layer.Attention dropout fixes alignment errors and minimizes translation errors at the word level.After empirical demonstration and comparison with their counterparts,we found improvement in the quality of the resulting translation system and a decrease in the perplexity and over-translation score.The ability of the proposed model was evaluated using Arabic-English and Persian-English datasets as well.We empirically concluded that adding an attention-based dropout layer helps improve GRU,SRU,and Transformer translation and is considerably more efficient in translation quality and speed.

基于权重微博链的改进LDA微博主题模型

社交网络尤其是微博中含有大量的短文本。短文本不同于传统的文本,其携带的语义特征信息密度低,很难对其进行准确的主题挖掘。针对这一问题,提出根据微博发布时间与原创、转发、评论微博等社交行为信息分配权重,使用背景知识丰富语义特征的微博链结构与基于此改进的LDA主题模型。实验结果表明,相比于标准的LDA模型,本模型的perplexity值更低,即具有较低的预测不确定度。

走进庄子的人生世界——读王充闾《逍遥游·庄子传》

王充闾新著《逍遥游·庄子传》是一部成功的作品,也是传记领域的新收获。作者通过以对《庄子》的解读整合还原庄子、以心灵感悟庄子、以疑难深化庄子、以诗性语言表述庄子等几条途径,整理出庄子一生完整的行迹,把握住了庄子的人格特点,凸显出了庄子的精神内涵,塑造出了一位有血有肉的智者形象。

An Empirical Study of Good-Turing Smoothing for Language Models on Different Size Corpora of Chinese

Data sparseness has been an inherited issue of statistical language models and smoothing method is usually used to resolve the zero count problems. In this paper, we studied empirically and analyzed the well-known smoothing methods of Good-Turing and advanced Good-Turing for language models on large sizes Chinese corpus. In the paper, ten models are generated sequentially on various size of corpus, from 30 M to 300 M Chinese words of CGW corpus. In our experiments, the smoothing methods;Good-Turing and Advanced Good-Turing smoothing are evaluated on inside testing and outside testing. Based on experiments results, we analyzed further the trends of perplexity of smoothing methods, which are useful for employing the effective smoothing methods to alleviate the issue of data sparseness on various sizes of language models. Finally, some helpful observations are described in detail.

A KIND OF HYBRID LANGUAGE MODEL EMPLOYED IN POST PROCESSING OF CHINESE SPEECH RECOGNITION

A kind of Chinese combined language model, that takes into account POS(part of speech) information in a trigram-based statistical language model, is presented in the paper. The theoretical analysis and experiments all show that the model not only is lower than trigram model in PP(perplexity), but also is superior to trigram model in dependence on testing text domain.

Analysis of the Concept and Connotation of the New Economy and Reconstruction of Its Development Path:A Study Based on Chengdu’s Practices

In order to promote the transformation and development of China’s economy and cope with the realistic challenges posed by the global economic recession and slowing domestic economic growth,the Fourth Session of the 12 th National People’s Congress adopted a strategic resolution on developing the "new economy" and fostering new drivers of growth.Chengdu,capital of Sichuan province,is one of the cities in China which firstly begin to carry out the layout of the new economy development.While achieving results,it also faces many troubles due to its insufficient cognition of the connotation,definition,and path of the "new economy." The paper is based on a theoretical study in combination with the new economic development of Chengdu’s five new paths,to analyze the concept and connotation of the "new economy." It also analyzes from the perspective of higher dimensional reconstruction and new strategic path of economic development,in order to add some insights to the theorectical study and to promote better strategic value of the new economy.

Petrology of Garnet Amphibolites from the Hualong Group: Implications for Metamorphic Evolution of the Qilian Orogen, NW China

The Qilian Orogen marks the junction of the North China, South China and Tarim cratons. The mechanism of continental growth during the formation of the orogen remains unclear. Based on detailed fieldwork, we present a systematic study of petrography, mineral chemistry and phase equilibria of garnet amphibolites from the Hualong Group, which represents the Precambrian basement in the southern accretionary belt of the Qilian Orogen. The garnet amphibolites mainly consist of amphibole, plagioclase, garnet and quartz, with minor pyroxene, biotite and ilmenite. A peak stage of upper amphibolite facies to low-temperature granulite facies metamorphism and retrograde metamorphism in the amphibolite facies affected the samples. Garnet has a homogeneous composition of Alm66-71Grs14-17Prp9_12Sps3-s, amphibole is ferro-hornblende, biotite belongs to the ferro-biotite species and pyroxene is dominated by orthopyroxene with few clinopyroxene. Pseudosection modeling of the garnet amphibolite samples indicates clockwise P-T paths. The samples witness peak metamorphism at conditions of -4.9-6.3 kbar and -755-820 ℃ in the upper amphibolite facies to low- temperature granulite facies, and retrograde cooling and decompression at conditions of-2.5-3.1 kbar and -325-545 ℃. It is inferred that peak metamorphism with high temperature and low pressure occurred at ca. 450 Ma during northward subduction of the South Qilian oceanic crust beneath the central Qilian Block. When continental collision occurred between the central Qilian and the Qaidam blocks, the Hualong Block was aecreted onto the South Qilian accretionary complex and experienced amphibolite facies retrograde metamorphism at ca. 440 Ma.

Semantic Model for Voice Controlled Telephone Dialing and Inquiry Systems

A new scheme is presented to detect a large number ofKeywordsin voice controlled switchboard tasks. The new scheme is based on two stages. In the first stage, N best syllable candidates with their corresponding acoustic scores are generated by an acoustic recognizer. In the second stage, a semantic model based parser is applied to determine the optimum keywords by searching through the lattice of N best candidates. The experimental results show that when the spoken input deviates from the predefined syntactic constraints, the parser can also demonstrate high performance. For comparison purposes, the most common way to incorporate the syntactic knowledge of the task directly into the acoustic recognizer in the form of a finite state network is also investigated. Furthermore, to address the sparse data problems, out of domain data in the form of newspaper text are used to obtain a more robust combined semantic model. The experiments show that the combined semantic model can improve the keywords detection rate from 90.07% to 92.91% when 80 ungrammatical sentences which do not conform to the task grammar are used as testing material.

Mathematical beauty and Palladian architecture:Measuring and comparing visual complexity and diversity

Palladio’s design principles,including proportion and harmony,have often been associated with mathematical definitions of beauty.However,the geometric and semantic properties of his fac¸ades have rarely been analysed in a holistic manner.There is little evidence of how mathematical beauty may be embodied in his architecture.This research investigates complexity(fractal dimension or D)and diversity(perplexity or PP)as aesthetic indices,aiming to capture distinct characteristics of Palladian architecture.The D and PP values of 22 Palladian villa fac¸ades are measured and analysed,before being compared with those of three Renaissance facades by Sebastiano Serlio.The combination of D and PP captures the geometric and semantic aesthetic qualities of architectural compositions.Importantly,the developed scatter plot of D and PP results supports the identification of four distinct aesthetic types of Palladian fac¸ade designs.The novel combination of D and PP measures contributes to a better understanding of one definition of the mathematical beauty of architecture,wherein“the whole is other than the sum of the parts”in compositional terms(a famous Gestalt principle).Specifically,this research provides new mathematical insights into the visual character of Palladian architecture and compares two important measures thereof.