Context-Aware Chinese Microblog Sentiment Classification with Bidirectional LSTM

.
Recently, with the fast development of the microblog, analyzing the sentiment orientations of the tweets has become a hot research topic for both academic and industrial communities. Most of the existing methods treat each microblog as an independent training instance. However, the sentiments embedded in tweets are usually ambiguous and context-aware. Even a non-sentiment word might convey a clear emotional tendency in the microblog conversations. In this paper, we regard the microblog conversation as sequence, and leverage bidirectional Long Short-Term Memory (BLSTM) models to incorporate preceding tweets for context-aware sentiment classification. Our proposed method could not only alleviate the sparsity problem in the feature space, but also capture the long distance sentiment dependency in the microblog conversations. Extensive experiments on a benchmark dataset show that the bidirectional LSTM models with context information could outperform other strong baseline algorithms.
.
https://www.researchgate.net/publication/308188542_Context-Aware_Chinese_Microblog_Sentiment_Classification_with_Bidirectional_LSTM
.


Read More

Chinese Microblog Sentiment Analysis Based on Sentiment Features

.
As the microblog has increasingly become an information platform for netizens to share their ideas, the study on the sentiment analysis of microblog has got scholars’ wide attention both at home and abroad. The primary goal of this research is to improve the accuracy of microblog sentiment polarity classification. With a view to the characteristics of microblog, a new method of semantically related feature extraction is proposed. Firstly, the Chinese word features are selected by text presentation in VSM and computing the weight by TF*IDF. Secondly, the proposed eight microblog semantic features are extracted, including sentence sentiment judgment based on emotional dictionary. Finally, three kinds of machine learning methods are used to classify the Chinese microblog under the feature vector combining the two methods. The experimental results indicate that the proposed feature extraction method outperforms the state-of-the-art approaches, and for this feature extraction algorithm, the classification performance is best when using the Naïve Bayes algorithm.
.
https://www.researchgate.net/publication/308327580_Chinese_Microblog_Sentiment_Analysis_Based_on_Sentiment_Features
.


Read More

Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations

.
We introduce a method to train Quantized Neural Networks (QNNs) --- neural networks with extremely low precision (e.g., 1-bit) weights and activations, at run-time. At train-time the quantized weights and activations are used for computing the parameter gradients. During the forward pass, QNNs drastically reduce memory size and accesses, and replace most arithmetic operations with bit-wise operations. As a result, power consumption is expected to be drastically reduced. We trained QNNs over the MNIST, CIFAR-10, SVHN and ImageNet datasets. The resulting QNNs achieve prediction accuracy comparable to their 32-bit counterparts. For example, our quantized version of AlexNet with 1-bit weights and 2-bit activations achieves $51\%$ top-1 accuracy. Moreover, we quantize the parameter gradients to 6-bits as well which enables gradients computation using only bit-wise operation. Quantized recurrent neural networks were tested over the Penn Treebank dataset, and achieved comparable accuracy as their 32-bit counterparts using only 4-bits. Last but not least, we programmed a binary matrix multiplication GPU kernel with which it is possible to run our MNIST QNN 7 times faster than with an unoptimized GPU kernel, without suffering any loss in classification accuracy. The QNN code is available online.
.
https://www.researchgate.net/publication/308457764_Quantized_Neural_Networks_Training_Neural_Networks_with_Low_Precision_Weights_and_Activations.
Download
. Read More

Sentiment Target Extraction Based on CRFs with Multi-features for Chinese Microblog

.
Sentiment target extraction on Chinese microblog has attracted increasing research attention. Most previous work relies on syntax, such as automatic parse trees, which are subject to noise for informal text such as microblog. In this paper, we propose a modified CRFs model for Chinese microblog sentiment target extraction. This model see the sentiment target extraction as a sequence-labeling problem, incorporating the contextual information, syntactic rules and opinion lexicon into the model with multi-features. The major contribution of this method is that it can be applied to the texts in which the targets are not mentioned in the sequence. Experimental results on benchmark datasets show that our method can consistently outperform the state-of-the-art methods.
.
https://www.researchgate.net/publication/308499279_Sentiment_Target_Extraction_Based_on_CRFs_with_Multi-features_for_Chinese_Microblog
.


Read More

An approach to sentiment analysis of short Chinese texts based on SVMs

.
... Experimental results have shown that the Naive Bayes classifier performs the best. Approach to analyze the sentiment of short Chinese texts is presented in [9]. By using word2vec tool, sentiment dictionaries from NTU and HowNet are extended. ...
.
https://www.researchgate.net/publication/308868603_An_approach_to_sentiment_analysis_of_short_Chinese_texts_based_on_SVMs
.

Read More

A Dynamic Conditional Random Field Based Framework for Sentence-Level Sentiment Analysis of Chinese Microblog

.

.
https://www.researchgate.net/publication/319051638_A_Dynamic_Conditional_Random_Field_Based_Framework_for_Sentence-Level_Sentiment_Analysis_of_Chinese_Microblog
. Read More

Speech Recognition With Deep Recurrent Neural Networks

.
Recurrent neural networks (RNNs) are a powerful model for sequential data. End-to-end training methods such as Connectionist Temporal Classification make it possible to train RNNs for sequence labelling problems where the input-output alignment is unknown. The combination of these methods with the Long Short-term Memory RNN architecture has proved particularly fruitful, delivering state-of-the-art results in cursive handwriting recognition. However RNN performance in speech recognition has so far been disappointing, with better results returned by deep feedforward networks. This paper investigates $backslash$emphdeep recurrent neural networks, which combine the multiple levels of representation that have proved so effective in deep networks with the flexible use of long range context that empowers RNNs. When trained end-to-end with suitable regularisation, we find that deep Long Short-term Memory RNNs achieve a test set error of 17.7% on the TIMIT phoneme recognition benchmark, which to our knowledge is the best recorded score.
.
https://www.researchgate.net/publication/319770184_Speech_Recognition_With_Deep_Recurrent_Neural_Networks
. Read More

Sentiment analysis of Chinese micro-blog text based on extended sentiment dictionary

.
Micro-blog texts contain complex and abundant sentiments which reflect user's standpoints or opinions on a given topic. However, the existing classification method of sentiments cannot facilitate micro-blog topic monitoring. To solve this problem, this paper presents a sentiment analysis method for Chinese micro-blog text based on the sentiment dictionary to support network regulators' work better. First, the sentiment dictionary can be extended by extraction and construction of degree adverb dictionary, network word dictionary, negative word dictionary and other related dictionaries. Second, the sentiment value of a micro-blog text can be obtained through the calculation of the weight. Finally, micro-blog texts on a topic can be classified as positive, negative and neutral. Experimental results show the effectiveness of the proposed method.
.
https://www.researchgate.net/publication/320682973_Sentiment_analysis_of_Chinese_micro-blog_text_based_on_extended_sentiment_dictionary
.

Read More

Alternating Multi-bit Quantization for Recurrent Neural Networks

.
Recurrent neural networks have achieved excellent performance in many applications. However, on portable devices with limited resources, the models are often too large to deploy. For applications on the server with large scale concurrent requests, the latency during inference can also be very critical for costly computing resources. In this work, we address these problems by quantizing the network, both weights and activations, into multiple binary codes {-1,+1}. We formulate the quantization as an optimization problem. Under the key observation that once the quantization coefficients are fixed the binary codes can be derived efficiently by binary search tree, alternating minimization is then applied. We test the quantization for two well-known RNNs, i.e., long short term memory (LSTM) and gated recurrent unit (GRU), on the language models. Compared with the full-precision counter part, by 2-bit quantization we can achieve ~16x memory saving and ~6x real inference acceleration on CPUs, with only a reasonable loss in the accuracy. By 3-bit quantization, we can achieve almost no loss in the accuracy or even surpass the original model, with ~10.5x memory saving and ~3x real inference acceleration. Both results beat the exiting quantization works with large margins. We extend our alternating quantization to image classification tasks. In both RNNs and feedforward neural networks, the method also achieves excellent performance.
.
https://www.researchgate.net/publication/322886129_Alternating_Multi-bit_Quantization_for_Recurrent_Neural_Networks
. Read More

Research on sentiment analysis of microblogging based on LSA and TF-IDF

.
... The anonymity of Weibo makes peo- ple being willing to express their real sentiments. Many existing techniques of sentiment analysis are based on sen- timent lexicons and traditional feature engineering [1]- [8]. Most of these methods need resort to external resource or manually preprocess features of words. ...
.
https://www.researchgate.net/publication/324462809_Research_on_sentiment_analysis_of_microblogging_based_on_LSA_and_TF-IDF
. Read More