scholarly journals Attentive Convolution: Equipping CNNs with RNN-style Attention Mechanisms

2018 ◽  
Vol 6 ◽  
pp. 687-702 ◽  
Author(s):  
Wenpeng Yin ◽  
Hinrich Schütze
Keyword(s):  
Neural Networks ◽  
Sentiment Analysis ◽  
Recurrent Neural Networks ◽  
Representation Learning ◽  
Local Context ◽  
Fixed Size ◽  
Convolution Operation ◽  
Input Text ◽  
Multiple Context ◽  
Textual Entailment

In NLP, convolutional neural networks (CNNs) have benefited less than recurrent neural networks (RNNs) from attention mechanisms. We hypothesize that this is because the attention in CNNs has been mainly implemented as attentive pooling (i.e., it is applied to pooling) rather than as attentive convolution (i.e., it is integrated into convolution). Convolution is the differentiator of CNNs in that it can powerfully model the higher-level representation of a word by taking into account its local fixed-size context in the input text t x. In this work, we propose an attentive convolution network, ATTCONV. It extends the context scope of the convolution operation, deriving higher-level features for a word not only from local context, but also from information extracted from nonlocal context by the attention mechanism commonly used in RNNs. This nonlocal context can come (i) from parts of the input text t x that are distant or (ii) from extra (i.e., external) contexts t y. Experiments on sentence modeling with zero-context (sentiment analysis), single-context (textual entailment) and multiple-context (claim verification) demonstrate the effectiveness of ATTCONV in sentence representation learning with the incorporation of context. In particular, attentive convolution outperforms attentive pooling and is a strong competitor to popular attentive RNNs. 1

scholarly journals Analys is and Creation of Free Sentiment Analysis Programs

2019 ◽  
Vol 25 (1) ◽  
pp. 83-105
Author(s):  
Josip Mihaljević
Keyword(s):  
Social Networks ◽  
Sentiment Analysis ◽  
Recurrent Neural Networks ◽  
Web Search ◽  
The Internet ◽  
Pilot Program ◽  
Online Programs ◽  
Input Text ◽  
Second Category ◽  
Pie Chart

This paper analyzes free online programs for sentiment analysis which can, on the bases of their algorithm, give a positive, negative or neutral opinion of a text. At the beginning of the paper sentiment analysis programs and techniques they use such as Naive Bayes and Recurrent Neural Networks are presented. The programs are divided into two categories for analysis. The fi rst category consists of sentiment analysis programs which analyze texts written or copied inside the user interface. The second category consists of programs for analyzing opinions posted on social networks, blogs, and other media sites. Programs from both categories were chosen for this research on the bases of positive reviews on computer science portals and their popularity on web search engin es such as Google and Bing. The accuracy of the programs from the fi rst category was checked by inserting the same sentence from movie reviews and comparing the results. Their additional options have also been analyzed. For the second category of programs, it was determined which social networks, blogs, and other social media they cover on the internet. The purpose of this analysis was to check the overall quality and options that free sentiment analysis programs provide. An example of how to create one’s own custom sentiment analyzer by using the available Python code and libraries found online is also given. Two simple programs were created using Python. The fi rst program belongs to the fi rst category of programs for analyzing an input text. This program serves as a pilot program for Croatian which gives only the basic analysis of sentences. The second program collects recent tweets from Twitter containing certain words and creates a pie chart based on the analysis of the results.

scholarly journals A Unified Model for Opinion Target Extraction and Target Sentiment Prediction

2019 ◽  
Vol 33 ◽  
pp. 6714-6721 ◽  
Author(s):  
Xin Li ◽  
Lidong Bing ◽  
Piji Li ◽  
Wai Lam
Keyword(s):  
Neural Networks ◽  
Sentiment Analysis ◽  
Recurrent Neural Networks ◽  
Unified Model ◽  
Primary Target ◽  
Target Extraction ◽  
Final Output ◽  
Benchmark Datasets ◽  
Task Dependency ◽  
Current Word

Target-based sentiment analysis involves opinion target extraction and target sentiment classification. However, most of the existing works usually studied one of these two sub-tasks alone, which hinders their practical use. This paper aims to solve the complete task of target-based sentiment analysis in an end-to-end fashion, and presents a novel unified model which applies a unified tagging scheme. Our framework involves two stacked recurrent neural networks: The upper one predicts the unified tags to produce the final output results of the primary target-based sentiment analysis; The lower one performs an auxiliary target boundary prediction aiming at guiding the upper network to improve the performance of the primary task. To explore the inter-task dependency, we propose to explicitly model the constrained transitions from target boundaries to target sentiment polarities. We also propose to maintain the sentiment consistency within an opinion target via a gate mechanism which models the relation between the features for the current word and the previous word. We conduct extensive experiments on three benchmark datasets and our framework achieves consistently superior results.

Sentiment Analysis Using Gated Recurrent Neural Networks

2020 ◽  
Vol 1 (2) ◽  
Author(s):  
Sharat Sachin ◽  
Abha Tripathi ◽  
Navya Mahajan ◽  
Shivani Aggarwal ◽  
Preeti Nagrath
Keyword(s):  
Neural Networks ◽  
Sentiment Analysis ◽  
Recurrent Neural Networks

Live Chat Analysis Using Machine Learning

2021 ◽  
Vol 8 (1) ◽  
pp. 39-44
Author(s):  
S. Kavibharathi ◽  
S. Lakshmi Priyankaa ◽  
M.S. Kaviya ◽  
Dr.S. Vasanthi
Keyword(s):  
Neural Networks ◽  
Natural Language Processing ◽  
Natural Language ◽  
Sentiment Analysis ◽  
Language Processing ◽  
Recurrent Neural Networks ◽  
Learning Model ◽  
Live Chat ◽  
Wide Range ◽  
Depth Analysis

The World Wide Web such as social networking sites and blog comments forum has huge user comments emotion data from different social events and product brand and arguments in the form of political views. Generate a heap. Reflects the user's mood on the network, the reader, has a huge impact on product suppliers and politicians. The challenge for the credibility of the analysis is the lack of sufficient tag data in the Natural Language Processing (NLP) field. Positive and negative classify content based on user feedback, live chat, whether the user is used as the base for a wide range of tasks related to the text content of a meaningful assessment. Data collection, and function number for all variants. A recurrent neural network is very good text classification. Analyzing unstructured form from social media data, reasonable structure, and analyzes attach great importance to note for this emotion. Emotional rewiring can use natural language processing sentiment analysis to predict. In the method by the Recurrent Neural Networks (RNNs) of the proposed prediction chat live chat into sentiment analysis. Sentiment analysis and in-depth learning technology have been integrated into the solution to this problem, with their deep learning model automatic learning function is active. Using a Recurrent Neural Networks (RNNs) reputation analysis to solve various problems and language problems of text analysis and visualization product retrospective sentiment classifier cross-depth analysis of the learning model implementation.

Sign in / Sign up

Export Citation Format

Share Document