About this Research Topic
Artificial Intelligence (AI) has been increasingly adopted to a wide range of biological areas such as systems biology, pharmacogenomics, network pharmacology, chemical property prediction, synthesis planning, molecular design and generation, protein-ligand interaction, drug target identification and interaction network, drug-related hybrid knowledge graphs, big data analysis, pandemic diseases analysis and image recognition, etc. The practice of AI in biological areas faces not only technical challenges of the integration among biological resources which is usually multi-modal, intertwined, and not guaranteed to be consistent, but also the explainability obsesses of deep learning, etc. Moreover, alongside the COVID emergency, more and more attention is focused on balancing social welfare, cultural moralities, and the biological practices involving privacy-preserving data collection and legal information usage, under rapid iterations of international political and technical negotiations, towards a responsive AI-enabled AI governance implementing justice, transparency and fairness.
From experimental and empirical criteria, a lot of AI practices on biological resources are successful in terms of speeding up, improving the success rates, and lowering the cost of scientific discovery and solution. However, most AI-related practices cannot provide a theoretical explanation for improved accuracy, enhanced robustness, with proven trustworthy legal infrastructure overtaking stakeholders’ right to know, right to participate, and right of supervision of related data, information, and knowledge and thereafter ensure the feasible reproducibility on changed data samples. With the accelerated integration of various sensor facilities, especially Internet of Things (IoT) and instrumental equipment at various degrees of details and depths, integrated processing of hybrid biological resources covering incomplete data, uncertain information, and multiple categories of knowledge originating in clinical, genetic, genomic, proteomic sources have been increasing both necessary and challenging, towards achieving improved accuracy, precision, and adaptability. Various transformation and integration efforts have been devoted to processing capabilities on retrieving information from multi-modal data, abstraction on information to attain knowledge hypotheses, and crossing modals transformations and optimization among data, information, and knowledge. Transformations among data, information, and knowledge open possibilities to comply with the incompleteness of data samples, insufficiency of information and vulnerability of invalid knowledge towards achieving more precise robust reproducibility and less repeated operations of data collection and information synthesis, and more comprehensive knowledge reproducibility through multiple sources reasoning and abstraction. These trends further increase the challenge of providing explainable and responsive AI solutions.
Prevailing challenges towards the solution arise across conceptual foundations, technical preparation, and legislation framework, especially involving cognitive semantic understanding, sharing, and utilization. As the Knowledge Graphs are increasingly recognized as an important approach to solving problems related to semantic understanding beyond question and answering systems integrating both subjective purposes and objective formalism, various innovations on Knowledge Graphs have been proposed, especially and most recently embedding technologies with Machine Learning. A foreseeable AI landscape with explainable, interactive, trustworthy, and responsive technical interactions is increasingly elusive based on Data, Information, Knowledge, and Wisdom (DIKW) hierarchical architectures which well pares with the organizing capability of Knowledge Graph technologies in terms of the 5W ( What, Where, When, How and Why), taking the forms of Data Graph, Information Graph, Knowledge Graph, and Wisdom Graph. This Research Topic aims to address experimental and theoretical results towards explainable, trustworthy, and responsive intelligent processing of biological resources crosscutting or integrating data, information, knowledge, and wisdom.
- Latest machine learning algorithms with applications on graphical biomedicine and bioinformatics
- Graphical content analysis techniques for biomedicine and bioinformatics
- Explainable machine learning in biomedicine and bioinformatics
-Trustworthy machine learning in biomedicine and bioinformatics
-Responsive AI strategies in biomedicine and bioinformatics
- Knowledge creation and management in biomedicine and bioinformatics
- Ontology modeling for biological information retrieve
- Multi-modal biological data collections and annotation
- Knowledge-based signal/image/video/text feature extraction
- Knowledge creation and reasoning for biological processing
- DIKW architecture-based biological resource processing and service provision
- Ontological approach and formal modeling, and verification methods of graphical biological resources
- Multi-objective graphical optimization with an application on biomedicine
-Legislation on collection, processing and usage of biological privacy information
From experimental and empirical criteria, a lot of AI practices on biological resources are successful in terms of speeding up, improving the success rates, and lowering the cost of scientific discovery and solution. However, most AI-related practices cannot provide a theoretical explanation for improved accuracy, enhanced robustness, with proven trustworthy legal infrastructure overtaking stakeholders’ right to know, right to participate, and right of supervision of related data, information, and knowledge and thereafter ensure the feasible reproducibility on changed data samples. With the accelerated integration of various sensor facilities, especially Internet of Things (IoT) and instrumental equipment at various degrees of details and depths, integrated processing of hybrid biological resources covering incomplete data, uncertain information, and multiple categories of knowledge originating in clinical, genetic, genomic, proteomic sources have been increasing both necessary and challenging, towards achieving improved accuracy, precision, and adaptability. Various transformation and integration efforts have been devoted to processing capabilities on retrieving information from multi-modal data, abstraction on information to attain knowledge hypotheses, and crossing modals transformations and optimization among data, information, and knowledge. Transformations among data, information, and knowledge open possibilities to comply with the incompleteness of data samples, insufficiency of information and vulnerability of invalid knowledge towards achieving more precise robust reproducibility and less repeated operations of data collection and information synthesis, and more comprehensive knowledge reproducibility through multiple sources reasoning and abstraction. These trends further increase the challenge of providing explainable and responsive AI solutions.
Prevailing challenges towards the solution arise across conceptual foundations, technical preparation, and legislation framework, especially involving cognitive semantic understanding, sharing, and utilization. As the Knowledge Graphs are increasingly recognized as an important approach to solving problems related to semantic understanding beyond question and answering systems integrating both subjective purposes and objective formalism, various innovations on Knowledge Graphs have been proposed, especially and most recently embedding technologies with Machine Learning. A foreseeable AI landscape with explainable, interactive, trustworthy, and responsive technical interactions is increasingly elusive based on Data, Information, Knowledge, and Wisdom (DIKW) hierarchical architectures which well pares with the organizing capability of Knowledge Graph technologies in terms of the 5W ( What, Where, When, How and Why), taking the forms of Data Graph, Information Graph, Knowledge Graph, and Wisdom Graph. This Research Topic aims to address experimental and theoretical results towards explainable, trustworthy, and responsive intelligent processing of biological resources crosscutting or integrating data, information, knowledge, and wisdom.
- Latest machine learning algorithms with applications on graphical biomedicine and bioinformatics
- Graphical content analysis techniques for biomedicine and bioinformatics
- Explainable machine learning in biomedicine and bioinformatics
-Trustworthy machine learning in biomedicine and bioinformatics
-Responsive AI strategies in biomedicine and bioinformatics
- Knowledge creation and management in biomedicine and bioinformatics
- Ontology modeling for biological information retrieve
- Multi-modal biological data collections and annotation
- Knowledge-based signal/image/video/text feature extraction
- Knowledge creation and reasoning for biological processing
- DIKW architecture-based biological resource processing and service provision
- Ontological approach and formal modeling, and verification methods of graphical biological resources
- Multi-objective graphical optimization with an application on biomedicine
-Legislation on collection, processing and usage of biological privacy information
Keywords: Strong Intelligence, Explain-ability, Responsive AI, Trustworthy AI, Biological Resources, DIKW, Knowledge Graph
Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.
