Neurocomputing publishes articles describing recent fundamental contributions in the field of neurocomputing. Neurocomputing theory, practice and applications are the essential topics being covered.

Neurocomputing welcomes theoretical contributions aimed at winning further understanding of neural networks and learning systems, including, but not restricted to, architectures, learning methods, analysis of network dynamics, theories of learning, self-organization, biological neural network modelling, sensorimotor transformations and interdisciplinary topics with artificial intelligence, artificial life, cognitive science, computational learning theory, fuzzy logic, genetic algorithms, information theory, machine learning, neurobiology and pattern recognition.

Neurocomputing covers practical aspects with contributions on advances in hardware and software development environments for neurocomputing, including, but not restricted to, simulation software environments, emulation hardware architectures, models of concurrent computation, neurocomputers, and neurochips (digital, analog, optical, and biodevices).

Neurocomputing reports on applications in different fields, including, but not restricted to, signal processing, speech processing, image processing, computer vision, control, robotics, optimization, scheduling, resource allocation and financial forecasting.

Neurocomputing publishes reviews of literature about neurocomputing and affine fields.

Neurocomputing reports on meetings, including, but not restricted to, conferences, workshops and seminars.

Neurocomputing reports on functionality/availability of software, on comparative assessments, and on discussions of neurocomputing software issues.

Now also including: Neurocomputing Letters - for the rapid publication of special short communications.

Special Issue on Distributed Machine Learning, Optimization and Applications
Submission Date: 2021-03-31

Recent advances in machine learning, information processing, multi-agent control, computational intelligence and networking have resulted in increasingly big data and distributed spatial data storage, which lead to new demands for machine learning to design more complex models and learning algorithms. In order to run algorithms with big data, the distributed machine leaning models and optimization algorithms are often required in engineering applications. This has inspired a lot of efficient learning algorithms and systems in artificial intelligence for data parallelism and model parallelism, including reinforcement learning, federated learning, deep learning, multi-task learning, and multi-agent systems. In addition, there is a surge of research activities devoted to the research of theories and applications of data driven and robust learning models on machine learning. With the development of distributed machine learning and optimization, the learning approaches have demonstrated remarkable performance across a range of applications, such as face and speech recognition, computer vision, image classification, computer game, multi-robot control, real-time resource allocation, etc.

This special issue aims to present the latest theoretical and technical advancements in the broad area of distributed machine learning and optimization for data, model and system analysis. The list of possible topics includes, but is not limited to:

 - Distributed algorithms in machine learning
 - Multi-agent reinforcement learning
 - Multi-task optimization and learning
 - Federated optimization and learning
 - Distributed optimization and real-time computing
 - Applications of distributed machine leaning and optimization

Special Issue on Learning to Combat Online Hostile Posts in Regional Languages during Emergency Situations
Submission Date: 2021-04-01

Overview: 

The increasing accessibility of the Internet has dramatically changed the way we consume information. The ease of social media consumption not only encourages individuals to freely express their opinion (freedom of speech), but also provides content polluters with ecosystems to spread hostile posts (hate speech, fake news, cyberbullying, etc.). Such hostile activities are expected to increase greatly during emergencies such the COVID-19 pandemic and in politicized events, like highly contested elections across the world. At the local level, most hostile posts are written in regional languages, and can therefore easily evade online surveillance engines, the majority of which are trained on the posts written in resource-rich languages such as English and Chinese. In much of Asia, Africa, and South America, where low-resource regional languages are used for day-to-day communication, there is no benchmark datasets and AI tools for identifying and mitigating the effects of hostile posts. This problem is faced even in developed countries, such as Italy and Spain, where local languages (like Sicilian and Catalan) are used widely for everyday communication. 

The special issue will encourage researchers working on multilingual social media analytics, text processing, learning technologies, and multimodal computation to think beyond the conventional ways (e.g., manual investigation, lexicon/dictionary/thesaurus-based matching, crowd-sourcing, etc.) of combating online hostile posts. The special issue will emphasize the following four major points and the solutions that connect these four points: (i) Regional Language: The offensive posts under inspection should be written in low-resource regional languages (e.g., Tamil, Urdu, Bangali, Polish, Czech, Lithuanian, etc.); however, solutions underpinning the barrier of resource-rich languages are also encouraged. (ii) Emergency Situation: The proposed solutions should be able to tackle misinformation during emergency situations where, due to the lack of enough historical data, learning models need to adopt additional intelligence (exogenous signals) to handle emerging and novel posts. (iii) Early Detection: Since the effect of misinformation during emergency situations is highly detrimental for society (e.g., health-related mis-advice during a pandemic may take human life), the solutions should be able to detect such hostile posts as early as possible after their appearance on social media. (iv) Machine Learning Solution: The advent of recent machine learning and deep learning technologies have encouraged researchers to develop automated approaches to combat hostile posts on social media. While statistical learning approaches are useful when the size of the datasets is small, recent neural approaches have shown tremendous improvement in terms of accuracy. However, the latter need a significant amount of data for training, which may not be feasible with resource-constrained languages. The challenge is to bring the best out of both into a single model in solving emerging crises related to abundant use of abusive texts on social media.

Topics of Interests 

We invite the submission of high-quality manuscripts reporting relevant research in the area of collecting, managing, mining, and understanding hostile posts on social media. The proposed solutions should build learning technologies (machine learning, deep learning) to solve the problem of fewer resources in regional languages to combat online hostile posts. Topics of interest include, but are not limited to: 

    Learning to detect hostile posts in regional languages 
    Modeling the evolution of online hostile posts ● Analyzing and modeling user behavior for hostile post propagation 
    Developing real-world tools for combating hostile posts 
    Behavioral study of the spreaders of hostile posts 
    Predicting the virality and severity of hostile posts—pre- and post-facto study 
    Hate speech normalization—alerting users about the hatefulness of the content during posting 
    Information extraction, ontology design, and knowledge graph for combating hostile posts 
    Early detection technologies of hostile posts on social media 
    Designing lightweight machine learning tools with less data for hostile post detection 
    Code-mixed and code-switched hostile post analysis 
    Open benchmark (tools, datasets) and dashboard related to regional hostile posts 
    Specific case studies and surveys related to hostile posts 
    Claim detection and verification related to misinformation 
    Fact-check worthiness of misinformation 
    Utilizing multimodality (audio, video, text, etc.) in combating misinformation 
    Impact/role of memes in spreading hostility on social media 
    Cross-region language analysis for hostile posts 
    Utilizing cross-platform learning and exogenous signals to combat misinformation 
    Computational social science analysis for hostile posts 
    Graph mining for fake news spreading and evolution
     
Papers Considered for the Special Issue

We plan to consider about 10 high-quality papers to be published in the special issue. We will invite authors of the top 4-5 papers accepted in CONSTRAINT-2021 workshop, collocated with AAAI’21 to submit an extended version to the special issue. The remaining papers will be selected through regular calls. All papers will go through a peer review process by at least three reviewers. 

Paper Submission Instruction: Follow the “Guideline for Authors” in https://www.journals.elsevier.com/neurocomputing

Important Dates (Tentative) 

Paper submission due: April 1, 2021 
Initial review Feedback: June 1, 2021 
Revision Due: July 1, 2021 
Final review decision: August 1, 2021

Guest Editors 

Tanmoy Chakraborty, IIIT-Delhi, India
Kai Shu, Illinois Institute of Technology, USA
H. Russell Bernard, Arizona State University, USA
Huan Liu, Arizona State University, USA

Special Issue on Graph-Powered Learning for Social Networks
Submission Date: 2021-04-15

Over the last decade, we have witnessed how social networks have evolved from being an entertaining extra to an integrated part of nearly every aspect of peoples’ daily lives. Social networks have profoundly changed how we interact with the world around us, including the ways to access news and information, the strategies to run business, the policy guidelines to prevent virus pandemic, the response to deal with disasters, the communication channels to improve healthcare and public health, etc.

At the same time, widespread usage of social networks has introduced various security and privacy challenges.

The arrival of smart mobile devices and the booming of mobile social applications in the recent years have only accelerated this trend. The shipment of social media users in January 2020 was about 3.80 billion, with an increase rate of 7 percent per year. Social networks naturally generate an unprecedented volume of graph data continuously, which pave a road for designing high quality services and applications such as recommendation systems, event detection, scam detection, rumor blocking, and privacy leakage detection, taking advantage of powerful machine learning techniques and tools.

The existing graph-powered learning methods cannot effectively capture and process sequential, topological, geometric, or other relational characteristics of graphical data, which is one of the major barriers to the widespread adoption of social network-based applications. Furthermore, these continuously evolving networks pose several challenges like growing user population, heterogeneity of user activities, explosion of generated data, and increasing concern of privacy leakage. Thus, there is an unprecedented need for more advanced graph-powered learning methods to be scalable for large-scale networks, feasible for utilizing multimodal data, flexible to model complex patterns, and capable of protecting user privacy. The goal of the special issue is to solicit high-quality, high-impact, original papers aiming at investigating emerging techniques and trending applications under the social network scenario using sophisticated graph- power learning methods. We are interested in submissions covering different types of models for sustainable services and applications in social networks.

Topics of Interest

Topics of particular interest include, but are not limited to:

Graph representation learning in social networks
Graph-based methods for analyzing social networks
Graph-based methods on anomaly detection for social networks
Graph classification, clustering, link prediction for social networks
Graph-based methods on recommendation for social networks
Knowledge graph modeling and management for social networks
Graph-based methods on privacy enhancing and anonymization techniques in social networks
Acceleration for graph-based methods in social networks
Parallel and distributed algorithms for graph data in social networks
Visualization of graph data in social networks
Graphical data based multimodal data fusion for social networks
Large scale graphical data processing for social networks
Graph-based methods for real/industry applications and systems for social networks