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 Knowledge Graph Representation & Reasoning
Submission Date: 2020-08-31

Recent years have witnessed the release of many open-source and enterprise-driven knowledge graphs with a dramatic increase of applications of knowledge representation and reasoning in fields such as natural language processing, computer vision, and bioinformatics. With those large-scale knowledge graphs, recent research tends to incorporate human knowledge and imitate human’s ability of relational reasoning. Factual knowledge stored in knowledge bases or knowledge graphs can be utilized as a source for logical reasoning and, hence, be integrated to improve real-world applications.

Emerging embedding-based methods for knowledge graph representation have shown their
ability to capture relational facts and model different scenarios with heterogenous information. By combining symbolic reasoning methods or Bayesian models, deep representation learning techniques on knowledge graphs attempt to handle complex reasoning with relational path and symbolic logic and capture the uncertainty with probabilistic inference. Furthermore, efficient representation learning and reasoning can be one of the paths towards the emulation of high-level cognition and human-level intelligence. Knowledge graphs can also be seen as a means to tackle the problem of explainability in AI. These trends naturally facilitate relevant downstream applications which inject structural knowledge into wide-applied neural architectures such as attention-based transformers and graph neural networks.

This special issue focuses on emerging techniques and trendy applications of knowledge graph representation learning and reasoning in fields such as natural language processing, computer vision, bioinformatics, and more.

The topics of this special issues include but not limited to:

Representation learning on knowledge graphs
Representation learning on text data
Logical rule mining and symbolic reasoning
Knowledge graph completion and link prediction
Relation extraction
Community embeddings
Knowledge representation and reasoning over large-scale knowledge graphs
Hybrid methods with symbolic and non-symbolic representation and reasoning
Automatic knowledge graph construction
Domain specific knowledge graphs, e.g., medical knowledge graphs
Knowledge dynamics of temporal knowledge graphs
Time-evolving knowledge representation learning
Question answering and dialogue systems with knowledge graphs
Knowledge-injected sentiment analysis
Commonsense knowledge representation and reasoning
Knowledge graphs for neural machine translation
Knowledge-aware recommendation systems
Knowledge graphs for digital health, e.g., mental healthcare and medical diagnosis
Few-shot relational learning on knowledge graphs
Federated learning with multi-source knowledge graphs in the decentralized setting
Graph representation learning for structured data
Explainable artificial intelligence with knowledge-aware models

Special Issue on Neural networks-based reinforcement learning control of autonomous systems (NRLC-AS)
Submission Date: 2020-09-30

Neural networks-based reinforcement learning control (NRLC) of autonomous systems is an active field due to its theoretical challenges and crucial applications. Note that there exist numerous difficulties in enhancing the intelligence and reliability of autonomous systems since autonomous and reliable techniques of guidance, navigation and control functionals are extremely involved in face of sophisticated and hazardous environments. In this context, high-intelligence reliable control technologies, especially based on neural networks tools, of autonomous systems are persistently pursued in trajectory tracking, path following, waypoints guidance, cooperative formation, etc. In addition, massive nonlinearities, sensor fault diagnosis, actuator failures tolerance, environment abnormalities and civil requirements have led to strong demands for the NRLC technologies in autonomous systems. Reinforcement learning, inspired by learning mechanisms observed in mammals, is concerned with how agent and actor ought to take actions to optimize a cost of its long-term interactions with the environment, and is gradually becoming the focus of learning control for autonomous systems. The autonomous systems inevitably suffer from actuator faults, component failures, insecurity factors, complex uncertainties, such that neural networks induced intelligence in autonomous control, fault tolerant control, network communication and signal progressing becomes dramatically significant. To be specific, by combining with neural networks and reinforcement learning, advances in the NRLC technologies of autonomous systems are exclusively pursued in this special issue.

This special issue will feature the mostly recent developments and the state-of-the-art of NRLC techniques for autonomous systems including ground, marine, and mobile vehicles, etc. The target audience includes both academic researchers and industrial practitioners. It aims to provide a platform for sharing recent results and team experience in intelligent learning control of autonomous systems. Topics to be covered in this special issue include, but are not limited to, the following

Neural networks-based reinforcement learning control of multiple autonomous systems;
Neural networks optimization-based reliable control of autonomous systems under multiple operating conditions;
Neural networks-based health monitoring and supervisory reliable control of autonomous systems;
Neural networks-based fault diagnosis and prognostics of autonomous systems
Neural networks-based intelligence application in resilient autonomous systems;
Neural networks learning-based location and navigation of autonomous systems;
Neural networks learning-based decision making of autonomous systems;
Neural networks learning for perception and recognition of autonomous systems;
Neural networks-based resilience control of autonomous systems;
Neural networks-based learning control application studies such as ground/mobile vehicles/marine vehicles.

Special Issue on Edge Intelligence: Neurocomputing Meets Edge Computing
Submission Date: 2020-09-30

Recent years have witnessed the proliferation of mobile computing and Internet-of-Things (IoT), where billions of mobile and IoT devices are connected to the Internet, generating zillions Bytes of data at the network edge. Driving by this trend and the development of 5G, edge computing, an emerging computing paradigm, has received a tremendous amount of interest. By pushing data storage, computing, and controls closer to the network edge, edge computing has been widely recognized as a promising solution to meet the requirements of low latency, high scalability and energy efficiency. In the meanwhile, with the development of neural networks, Artificial Intelligence (AI) has been applied to a variety of disciplines and proved highly successful in a vast class of intelligent applications cross many domains.

Recently, edge intelligence, aiming to facilitate the deployment of neural networks on edge computing, has received significant attention. However there are many challenges existing for a novel design of edge computing architecture to AI applications, and their co-optimization. For instance, conventional neural networks techniques usually entail powerful computing facilities (e.g., cloud computing platforms), while the entities at the edge may have only limited resources for computations and communications. This suggests that AI algorithms should be revisited for edge computing to AI models into the edge device for efficient processing. On the other hand, the adapted deployments of neural networks at the edge empower the efficient learning systems that can provide the “smartification” across different layers, e.g., from network communications to applications, and also involve collaborations across edge to cloud. Finally, designing algorithms for small-scale edge devices in a learning ambience is all the challenging as there are several conflicting issues to account for. These include, memory management, power management, and compute capability of a node, etc.

In this special issue, we solicit original work on ML/AI, specifically catered to deep neural networks on/for edge computing, and efficient learning systems on edge computing, addressing specific challenges in this field. The list of possible topics includes, but not limited to:

Neuromorphic computing challenges on Edge devices
Intelligent Edge Computing Devices for neurocomputing applications
Spiking Netural Networks on Edge devices - low-power and memory bandwidth challenges Conventional Neural Networks algorithms on edge computing
Neurocomputing Algorithms on edge devices for Wearables;
Edge/Fog-infused Cloud architectures for ML/AI applications
Efficient Artificial intelligence algorithms on edge computing
Few-shot learning on edge devices for ML/AI applications
Resource and data management for edge intelligence
AI/ML for small-scale low-power edge devices
Distributed and cooperative learning with edge devices on Cloud
Applications of edge intelligence & neurocomputing
5G-enabled services for edge intelligence & neurocomputing
System architectures of edge based neurocomputing
Architecture & application of Edge AI for IoT
Security & privacy for edge computing
Attack mitigation in edge computing

Special Issue on Real-time Dynamic Network Learning for Location Inference Modelling and Computing
Submission Date: 2020-10-15

User location information contributes to in-depth social network data analytics. Discovering physical locations of users from their online media messages helps us to bridge the online and offline worlds. This also supports many real-life applications like emergency reporting, disaster management, location-based recommendation, location-based advertisement, region-specific topic summarization, and disease outbreak monitoring. For instance, the social distance has played a key role to reduce the Covid19 outbreak. However, location information is not always available because most users may not clearly annotate their locations in user profiles. Recent research trends intend to incorporate multiple types of data including text data, linked data, sensor data, as well as auxiliary insightful feature data. These data generate the linked and dynamic network data, which can be utilized together to learn and infer the user locations in different applications.

However, the existing techniques like recurrent neural network and generative adversarial network are still expensive to train the network models. It is more challenging to handle the dynamics of the networks for particular tasks, particularly when the data distribution and the types of data are not even. Furthermore, the diverse location inference tasks in real applications make the issue being more complex, e.g., next-visit location, event-based location, shopping location, indoor location, web location, etc. As such, novel multi-model dynamic network learning techniques expect to be investigated.

This special issue focuses on emerging techniques and trendy applications of real-time dynamic network learning in fields such as neural network, dynamic network, spatial feature pattern recognition, and active learning.

Topics of Interests

The topics of this special issues include but not limited to:

Dynamic network learning
Spatial feature pattern extraction and learning
Location inference modelling
Ensemble learning for location prediction
User location profiling
Indoor location inference and learning
Location concept reasoning and learning
Spatial modelling and reasoning
Spatial information integration
Benchmarking study and novel applications of location inference