Abstract
The amount of information being generated in different domains (e.g., in the context of genomics) and especially in complex environments requires the use of artificial intelligence techniques to facilitate data analysis and generate relevant conclusions for understanding the domain under study.
Currently, AI algorithms, some of them based on black box models such as neural networks, are being used to perform data analysis and generate conclusions. However, the results are not 100% accurate and black-box techniques do not allow practitioners to check the reasons behind the results obtained in order to improve them. This increases the distrust of users and may lead to the disuse of such techniques.
We propose the application of an Explainable AI that allows the user to identify the decisions that motivate the results, in an understandable way, and that allows the system to evolve according to their needs.
Scientific officer
Stakeholders
Applications
- Information management
- Analysis and interpretation of large amounts of data (Big Data).
- Selection and consolidation of information from different sources or data repository AND more specific applications in the genomics domain:
- Integration and analysis of clinical and genomic data.
- Understanding of the biological mechanisms that lead to disease
- Identification and development of drugs directed to specific targets.
- Information systems for genomic data management and analysis
- Identification of genomic variations likely to cause disease
- Detection of patterns for genomic analysis
Technical advantages
- The technical advantages of these applications include:
- Use of IAE technologies for data analysis and generation of results, allowing to understand the decision process through understandable explanations for the user.
- Creation of adaptive systems based on the accuracy of the results obtained. This will facilitate the evolution of the systems, adapting them much better to the user’s needs, since the user can actively participate in the process.
- Reduction of the cost of development and implementation of Information Systems in the bioinformatics domain.
Benefits it provides
Benefits: * Improved identification and processing of data used * Efficient and effective data management * Improved data exploitation * Improved quality of software applications for the bioinformatics context * Cost reduction in the study and analysis of the genomic domain * Automatic generation of genomic analysis systems.
Relevant experience
The Center for Research in Software Production Methods (PROS) has extensive experience in Modeling, Design and Development of Information Systems as well as in the development of software tools for the management and analysis of genomic information. It also has many articles and projects related to the subject.