With the exponential growth of multimedia content, visual sentiment classification has emerged as a significant research area. However, it poses unique challenges due to the complexity and subjective nature of the visual information. This can be attributed to the significant presence of semantically ambiguous images within the current benchmark datasets, which enhances the performance of sentiment analysis but ignores the differences between various annotators. Moreover, most current methods concentrate on improving local emotional representations that focus on object extraction procedures rather than utilizing robust features that can effectively indicate the relevance of objects within an image through color information. Motivated by these observations, this paper addresses the need for efficient algorithms for labeling and classifying sentiment from visual images by introducing a novel hybrid model, which combines content‐based image retrieval (CBIR) and a multi‐input convolutional neural network (CNN). The CBIR model extracts color features from all dataset images, creating a numerical representation for each. It compares a query image to dataset images’ features to find similar features. This process continues until the images are grouped according to color similarity, which allows accurate sentimental categories based on similar features and feelings. Then, a multi‐input CNN model is utilized to extract and efficiently incorporate high‐level contextual visual information. This model comprises 70 layers, with six branches, each containing 11 layers. It seeks to facilitate the fusion of complementary information by incorporating multiple input categories that differ according to the color features extracted by the CBIR technique. This feature enables the model to understand the target and generate more precise predictions fully. The proposed model demonstrates significant improvements over existing algorithms, as evidenced by evaluations of six benchmark datasets of varying sizes. Also, it outperforms the state of the art in sentiment classification accuracy, getting 87.88%, 84.62%, 84.1%, 83.7%, 80.7%, and 91.2% accuracy for the EmotionROI, ArtPhoto, Twitter I, Twitter II, Abstract, and FI datasets, respectively. Furthermore, the model is evaluated on two newly collected large datasets, which confirm its scalability and robustness in handling large‐scale sentiment classification tasks, and thus achieves a significant accuracy of 85.21% and 83.72% with the BGETTY and Twitter datasets, respectively. This paper contributes to the advancement of visual sentiment classification by offering a comprehensive solution for analyzing sentiment from images and laying the foundation for further research.

ABSTRACT Named entity recognition (NER) is a fundamental process in NLP and a requirement for most processes. This article aims to identify the named entities in the context of social networks. For this purpose, the idea of segmenting text into suitable and unsuitable expressions for the named entities has been used. So the contribution of this article is to process informal text in the Persian language by the Beam search algorithm to detect named entities. Due to the reproductive nature of language, new words and names are always produced, and available NER systems are inefficient in detecting new entities. The other contribution of this article is to make it possible to recognize the emerging named entity by applying dynamic external knowledge. According to a sense of the lack of datasets in low‐resource languages, N‐Gram and Wikipedia anchor datasets have been prepared for Persian and deployed as external knowledge. Also, a corpus of named entities in Persian from the telegram dataset has been generated. Three native experts have done labeling of this corpus. Evaluation of these three experts and the proposed method shows that the result of the proposed method is acceptable compared to the result of a human‐to‐human also to other methods.

During the development of social media, there has been a transformation in social communication. Despite their positive applications in social interactions and news spread, it also provides an ideal platform for spreading rumors. Rumors can endanger the security of society in normal or critical situations. Therefore, it is important to detect and verify the rumors in the early stage of their spreading. Many research works have focused on social attributes in the social network to solve the problem of rumor detection and verification, while less attention has been paid to content features. The social and structural features of rumors develop over time and are not available in the early stage of rumor. Therefore, this study presented a content-based model to verify the Persian rumors on Twitter and Telegram early. The proposed model demonstrates the important role of content in spreading rumors and generates a better-integrated representation for each source rumor document by fusing its semantic, pragmatic, and syntactic information. First, contextual word embeddings of the source rumor are generated by a hybrid model based on ParsBERT and parallel CapsNets. Then, pragmatic and syntactic features of the rumor are extracted and concatenated with embeddings to capture the rich information for rumor verification. Experimental results on real-world datasets demonstrated that the proposed model significantly outperforms the state-of-the-art models in the early rumor verification task. Also, it can enhance the performance of the classifier from 2% to 11% on Twitter and from 5% to 23% on Telegram. These results validate the model's effectiveness when limited content information is available.

Online social media such as Twitter are growing so rapidly. Recently, Twitter has become one of the popular microblogging services on the Internet. It lets millions of users to communicate and interact by sending short messages of up to 140 characters. The massive amount of information over the web from Twitter requires an automatic tool that can determine the topics that people are talking about. The Topic Detection task is concentrated on discovering the main topics automatically. In this article at first, we explore different approaches to detect topics of tweets. Then, we will classify these topic detection approaches to four classes of categories, including with word embedding or without word embedding, specified or unspecified, offline (RED) or online (NED), and supervised or unsupervised. Finally, we will discuss the studied approaches in detail.

Abstract Nowadays, it is necessary that users have access to information in a concise form without losing any critical information. Document summarization is an automatic process of generating a short form from a document. In itemset‐based document summarization, the weights of all terms are considered the same. In this paper, a new approach is proposed for multidocument summarization based on weighted patterns and term association measures. In the present study, the weights of the terms are not equal in the context and are computed based on weighted frequent itemset mining. Indeed, the proposed method enriches frequent itemset mining by weighting the terms in the corpus. In addition, the relationships among the terms in the corpus have been considered using term association measures. Also, the statistical features such as sentence length and sentence position have been modified and matched to generate a summary based on the greedy method. Based on the results of the DUC 2002 and DUC 2004 datasets obtained by the ROUGE toolkit, the proposed approach can outperform the state‐of‐the‐art approaches significantly.

Deception plays a leading role in many intelligent systems. In order to model the uncertainties and make artificial deception near the one existing in the real world, fuzzy theory seems a reasonable tool. This study is the pioneering one to incorporate fuzzy logic concepts into the phenomenon of deceptive robotics. In this study, a hide-and-seek process is considered in a maze imaginable in any form between two robots, one of which is basically trying to deceive the other. The fuzzy definition of behavioral strategies based on past experience for both the deceiver and the competitor robot makes them act like human beings in conflict with each other. Combining the fuzzy reasoning with ant-inspired metaheuristics is another aspect of novelty in this study: Fulfilling the deception, the deceiver is supposed to produce two deceptive signals (track and pheromone) using a fuzzy inference system in order to arrange the environment as desired. After the deceiver decides where to go, the robot under deception is to decide which path to choose based on a utility function calculated within a hierarchical fuzzy inference system whose direct inputs are the value of deception signals and also his behavioral strategy. Since the ant-inspired deception signal varies by passing time, not always will everything look like as the deceiver designs. In addition, in order to test the experimental results along with the simulations, special robots with demanded features are designed and manufactured who benefit from a vision-based feedback to move in a real maze. The results of a series of extensive experiments give an evidence of the effectiveness of the proposed deception algorithm in terms of a sufficiently high deception success percentage. Furthermore, it is demonstrated that even the rare cases of deception failure speak for the human reasoning abilities of the robots, which is another support to effectiveness of the proposed algorithm. Moreover, due to the nature of the problem, here, we will face a noncooperative game with incomplete information between rational players in which the belief of the robot under deception is manipulated by the deceiver through the arrangement of the environment by deception signals.

Graph partitioning is a NP-hard problem with multiple conflicting objectives. The graph partitioning should minimize the inter-partition relationship while maximizing the intra-partition relationship. Furthermore, the partition load should be evenly distributed over the respective partitions. Therefore this is a multi-objective optimization problem (MOO). There are two approaches to MOO using genetic algorithms: weighted cost functions and finding the Pareto front. We have combined Pareto front method and cellular learning automata to exploit the potentiality of both in the hybridized algorithm. The proposed methods of this paper used to improve the performance addition to hybridization, are using an optimized method in generating reinforcement signal vector and considering the solutions of each non-dominated set as neighbours. These improvements make the search more efficient and increase the probability of finding more optimal solutions, also changing neighbour set at each generation, prevent the neighbours from getting stuck in the neighbourhood local optima. Finally, a simulation research is carried out to investigate the effectiveness of the proposed hybrid algorithm. The simulation results confirm the effectiveness of the proposed method.