An international research team with participants from several universities including the FAU has proposed a standarised registry for Artificial Intelligence work in biomedicine to enhance the reproducibility of results & create trust within the use of AI algorithms in biomedical research & in future, in everyday clinical practice. The scientists presented their proposal in journal Nature Methods.
In the last decades, new technologies have made it possible to develop a good sort of systems which might generate huge amounts of biomedical data, for instance in cancer research. At same time, completely new possibilities have developed for examining & evaluating this data using Artificial Intelligence methods. AI algorithms in medical care units, e.g., can predict circulatory-failure at an early-stage based-on large amounts of data from several monitoring-systems by processing tons of complex information from different sources at same time, which is way beyond human capabilities.
This great potential of AI systems results in an unmanageable number of biomedical AI applications. Unfortunately, the corresponding reports & publications don’t always adhere to best practices or provide only incomplete information about the algorithms used or the origin of the info . This makes assessment & comprehensive comparisons of AI models difficult. the choices of AIs aren’t always comprehensible to humans & results are seldomly fully reproducible. This situation is un-tenable, especially in clinical research, where trust in AI models & transparent research reports are crucial to increase acceptance of AI-algorithms and to develop improved AI methods for basic biomedical research.
To address this problem, International research team including the researchers from the FAU has proposed the AIMe registry for Artificial Intelligence in biomedical research, a community-driven registry that permits users of latest biomedical AI to make easily accessible, searchable & citable reports which might be studied & reviewed by scientific community.
The freely accessible registry is out there online & consists of a user-friendly web service that guides users through the AIMe standard & enables them to get complete and standardized reports on the AI models used. Unique AIMe identifier is automatically created, which ensures that the report remains persistent & may be specified in publications. Hence, authors do not have to deal with the time-consuming description of all facets of the AI utilized in articles for scientific journals & easily refer to report within AIMe registry.
Since the registry is de-signed as web platform maintained by the scientific community, every user can ask questions on existing reports, make comments or suggest improvements. This feedback from the community also will be included within the annual update of the AIMe standard, and interested researchers can join the AIMe committee to become more involved within the further standardization of biomedical AI.
“The AIMe registry not only makes it possible to simply report Artificial Intelligence methods in citable form, but also contains a database which enables searching out for relevant existing AI systems. This prevents researchers from reinventing an already existing approach and makes it easier for them to assess whether a potentially useful AI method has been evaluated in sufficient depth,” reports Prof. Dr. David B. Blumenthal from the Biomedical Network lab at Department AI in Biomedical Engineering of FAU.
The findings were published in the Journal Nature Methods.
