In recent years, fascinating progresses have been made in utilizing artificial intelligence to solve a broad range of problems. AI systems today can match and even outperform human performance in certain challenging tasks, including visual cognition. Many recent AI advances have been largely inspired by Neuroscience research into biological brain, guided by architectural and algorithmic constrains from biological neural networks. However, artificial neural networks remain to be “black boxes”, where the internal representations and computations of network components are poorly understood.
In this Research Topic, we advocate the potential for cognitive neuroscience to further benefit AI. Specifically, research techniques and approaches available in cognitive neuroscience, including single-unit recording, neuroimaging, cognitive, and lesion techniques, can serve as a repertoire of tools for unveiling the black boxes of AI, illuminating the computations and representations inside AI networks. Further, research findings from cognitive neuroscience can provide inspirations to develop next generation of AI, by build-in a priori architectures, algorithms, and knowledge. The purpose of this Research Topic is to bring together research efforts from AI and cognitive neuroscience, seeking to integrate AI and cognitive neuroscience toward a new field of cognitive neurointelligence.
Topics of interest include, but are not limited to:
- Information processing levels in the DNN, e.g., sensory, perceptual, and semantic levels
- Functional organizations in the DNN, e.g., modularity and population coding
- Manipulation of input statistics and network structure components to study abnormal development of AI
- Research of brain-like AI algorithms based on biological systems
- Brain benchmark: evaluation of AI systems based on cognitive neural predictability
- Developing brain-like AI systems for dynamic open environment
In recent years, fascinating progresses have been made in utilizing artificial intelligence to solve a broad range of problems. AI systems today can match and even outperform human performance in certain challenging tasks, including visual cognition. Many recent AI advances have been largely inspired by Neuroscience research into biological brain, guided by architectural and algorithmic constrains from biological neural networks. However, artificial neural networks remain to be “black boxes”, where the internal representations and computations of network components are poorly understood.
In this Research Topic, we advocate the potential for cognitive neuroscience to further benefit AI. Specifically, research techniques and approaches available in cognitive neuroscience, including single-unit recording, neuroimaging, cognitive, and lesion techniques, can serve as a repertoire of tools for unveiling the black boxes of AI, illuminating the computations and representations inside AI networks. Further, research findings from cognitive neuroscience can provide inspirations to develop next generation of AI, by build-in a priori architectures, algorithms, and knowledge. The purpose of this Research Topic is to bring together research efforts from AI and cognitive neuroscience, seeking to integrate AI and cognitive neuroscience toward a new field of cognitive neurointelligence.
Topics of interest include, but are not limited to:
- Information processing levels in the DNN, e.g., sensory, perceptual, and semantic levels
- Functional organizations in the DNN, e.g., modularity and population coding
- Manipulation of input statistics and network structure components to study abnormal development of AI
- Research of brain-like AI algorithms based on biological systems
- Brain benchmark: evaluation of AI systems based on cognitive neural predictability
- Developing brain-like AI systems for dynamic open environment