Overview

This leading international journal promotes and stimulates research in the field of artificial intelligence (AI). Covering a wide range of issues - from the tools and languages of AI to its philosophical implications - Computational Intelligence provides a vigorous forum for the publication of both experimental and theoretical research, as well as surveys and impact studies. The journal is designed to meet the needs of a wide range of AI workers in academic and industrial research.

Please see the Aims and Scope to learn about the focal topics in Computational Intelligence.

Aims and Scope

FOCAL TOPICS OF COMPUTATIONAL INTELLIGENCE

Discovery science and knowledge mining. Discovery science (also known as discovery-based science) is a scientific methodology which emphasizes analysis of large volumes of experimental data or text data with the goal of finding new patterns or correlations, leading to hypothesis formation and other scientific methodologies. Tools of interest include: Data Mining: looking for associations or relationships in operational or transactional data; Text Mining and Information Extraction: looking for concepts and their associations or relationships in natural language text; Structured, semi-structured and unstructured text mining; Text Summarization: extracting terms and phrases from large text document collections that summarize their content; Web mining: Web structure, content and usage mining; and, Ontology Learning from Text and Data bases.

Web intelligence and semantic web.  Web intelligence is concerned with the application of AI to the next generation of web systems, services and resources. These include better search/retrieval algorithms, client side systems (e.g. more effective agents) and server side systems (e.g. effective ways to present material on web pages and throughout web sites, including adaptive websites and personalized interfaces).

The semantic web is an extension to the World Wide Web, in which web content is expressed in a form that is accessible to programs (software agents), following the vision of the web as universal medium for data, information and knowledge exchange.

Agents and multiagent systems.  Agents as a computational abstraction have replaced 'objects' in software and have provided the necessary ingredients to move to societies of interacting intelligent entities, based on concepts like agent societies, market economies, e-commerce models and game theory. Such abstractions are dispersed throughout the scientific world, depending largely on applications. Multiagent systems (MAS) are systems in which many autonomous intelligent agents interact with each other. Agents can be either cooperative, pursuing a common goal, or selfish, going after their own interests. Architectures, interaction protocols and languages must be developed for multiagent systems. Topics of interest include: Autonomy-oriented computing; Agent systems methodology and language; Agent-based simulation and modeling; Agent-based applications; Agent-based negotiation and autonomous auction; Advanced Software Engineering supports for Multiagent systems; Trust in Agent Society; and Distributed problem solving.

Machine learning in knowledge-based systems.  Knowledge-based systems aim to make expertise available for decision making, and information sharing, when and where needed. The next generation of such systems needs to tap into large domain-specific knowledge, which combine machine learning and structured background knowledge representation, such as ontology, and causal representations and constraint reasoning. Information sharing is concerned with creating collaborative knowledge environments for sharing and disseminating information. Learning is based on real-world data. Key challenges involve the decomposition of practical problems into multiple learnable components, the interaction between the components, and the application of suitable learning algorithms, often in the absence of adequate amounts of labeled training data. Topics of interest include the application of machine learning methods to new practical problems introducing novel algorithms, system frameworks of learnable components or evaluation techniques.

Key application areas of AI. We aim to make the journal the focus of key application areas, where AI is making a significant impact, but lack a coherent publication venue. These include: Business Intelligence, i.e. data mining to support business decision makers; Social Network mining, e.g. modelling aggregate properties and dynamics of social networks, classifying vertices and edges of social networks, identifying clusters of users; Critical Infrastructure Protection, e.g. intrusion/anomaly detection & response, learning knowledge bases of system administration, log file mining); Entertainment and Game Development, i.e. building game engines using AI techniques; Software Engineering, including program understanding, software repositories and reverse engineering; Business, Finance, Commerce and Economics: learning aggregate behaviours (e.g. stock market trends) or modeling individual and group demographics (e.g. web mining); and Knowledge-based and Personalized User Interfaces, to make interaction clearer to the user and more efficient, with better support for the users' goals, and efficient presentation of complex information.

Please note that submissions that are straightforward applications to Machine Learning or other AI techniques to new tasks or new domains will be rejected without review unless they bring novelty in other aspects, such as significance and analysis of the results, explanations of why some methods work better than others in these domains, or other relevant insights. 

Special Issue on AI-Generated Content Empowered Scalable Big Models for Industrial IoT Networks
Submission Date: 2025-02-28

With the rapid development of edge computing and wireless communication, the field of artificial intelligence (AI) has witnessed remarkable advancements, particularly in the realm of large-scale models and its application into industrial Internet of Things (IoT) networks, such as intelligent transportation and monitoring, smart cities, and smart home. One of the most promising developments in this domain is the integration of big models with AI-generated content (AIGC) techniques, due to the unprecedented ability to automate the creation of various content, such as text, images, and videos. With the ability to produce efficiently large amounts of high-quality content, AIGC can save time and resources that would otherwise be spent on manual content creation. In practical industrial IoT networks, the big models should be implemented in a scalable way adaptively to the computing and communication capability of the nodes in the system, and the AIGC-empowered scalable big models can significantly help accomplish the tasks in the networks, from the data collect, data transmission, data analysis, and automatic decision making.

However, several critical challenges still exist on the AIGC-empowered scalable big models for industrial IoT networks. One critical challenge is the scalable model training and intelligent inference for big models and AIGC in the industrial IoT networks, where the computing and communication capability of the nodes in the system should be fully taken into account. One more challenge is how to intelligently devise the system communication protocols such as transmission and reception schemes, by fully exploiting the inherent characteristics of big models and AIGC. Another critical challenge is the advanced resource management for the AIGC-empowered scalable big models in the industrial IoT networks, where some distributed and secure optimization framework should be implemented to support the AIGC in the industrial IoT networks.

Topics for this call for papers include but are not restricted to:

    Scalable model training and inference in industrial IoT networks
    Enhancing security and privacy in industrial IoT networks using AIGC-powered big models
    Advanced resource allocation in industrial IoT networks with scalable big models
    Real-time anomaly detection in industrial IoT Networks using AIGC-driven big models
    Intelligent edge computing assisted by AIGC
    Semantic communication for AIGC-empowered big models
    Efficient data processing and analysis in AIGC-empowered industrial IoT networks
    Scalable big models for energy management and optimization in industrial IoT networks
    Wireless communication and mobile edge computing
    Applications and test beds of industrial IoT networks with scalable big models

Guest Editors:

Prof. Jiawen Kang
Guangdong University of Technology
China

Assoc. Prof. Huakun Huang
Guangzhou University
China

Prof. Xutao Li
Shantou University
China

Prof. Dusit Niyato
Nanyang Technological University
Singapore

Assis. Prof. Yao Sun
University of Glasgow
United Kingdom

Keywords: AIGC; big models; intelligent model inference; advanced edge computing; advanced wireless communication; intelligent resource allocation and management; security and privacy issues; prototype and hardware design; industrial IoT networks

Special Issue on Feature Engineering and Deep Interpretable Model in Psychiatric Disorders
Submission Date: 2025-03-25

Psychiatric disorders such as schizophrenia, paranoid disorders, schizotypal personality disorder, brief psychotic disorders, major depressive disorder with psychotic features, and bipolar disorders have a significant impact on individuals, families, and society due to their unclear etiology, strong latency, and substantial emotional fluctuations. These disorders also impose enormous pressure on the public healthcare system. By implementing innovative technologies through biomedical engineering, achieving early and accurate diagnosis can greatly improve the effectiveness of healthcare, enhance quality of life, and assist in the development of safe and effective medical treatment plans.

Detection of psychiatric disorders involves various brain imaging techniques such as electroencephalography (EEG), event-related potentials (ERP), structural magnetic resonance imaging (sMRI), functional magnetic resonance imaging (fMRI), diffusion tensor imaging (DTI), computed tomography (CT), positron emission tomography (PET), magnetoencephalography (MEG), among others. With advancements in neuroscience, multimodal medical data is also becoming a research focus. Due to the diverse etiological factors of psychiatric disorders, enhancing the interpretability of features from multimodal neuroimaging data through feature engineering can effectively explore biomarkers in the neurological system, contributing to optimizing features and understanding the pathogenesis mechanisms. The multidimensional, multiscale, and multitype implementation of deep recognition models for multimodal data in neurological systems is challenging. Deep models need to exhibit robustness and generalization to different modalities, while also enhancing interpretability and controllability. Furthermore, research on deep interpretable models for subtype precise classification based on psychiatric disorders provides more accurate diagnostic aids for clinical diagnosis. We aim to collect information on the research of feature engineering and deep interpretable models in multimodal neural data for the recognition of psychiatric disorders or subtype precise classification, to provide objective diagnostic and therapeutic support for psychiatric disorders.

Topics for this call for papers include but are not restricted to:

    Interpretability of important features related to outcome in the recognition of psychiatric disorders;
    Biological markers of psychiatric disorders exploration;
    Data augmentation and feature enhancement of neuro data for small sample psychiatric disorders;
    Multimodal feature learning, pathological feature inference, and mining for psychiatric disorders using feature engineering;
    Psychiatric disorder recognition using multimodal deep interpretable models;
    Subtype precise classification of psychiatric disorders using multimodal deep interpretable models;
    Deep recognition or subtype precise classification of psychiatric disorders with interpretable using self-supervised techniques;
    Deep recognition or subtype precise classification of psychiatric disorders with interpretable using self-learning;
    Deep recognition subtype precise classification of psychiatric disorders with interpretable using reinforcement learning;
    Personalized diagnosis and treatment for patients through incremental learning in the recognition or subtype precise classification of psychiatric disorders.

Guest Editors:

Xiaofeng Li (Lead)
Heilongjiang International University
China

Gang Liu
Harbin Engineering University
China

Shenjun Zhong
Monash University
Australia

Kristen Moore
CSIRO's Data61
Australia

Keywords: Psychiatric disorders; Deep interpretable model; Feature engineering; Biomedical engineering; Brain imaging techniques; Multimodal neuroimaging data; Deep recognition models; Subtype precise classification.