Advancing science through mining libraries, ontologies, and communities.

Area of Science:

• Life Science
• Bioinformatics
• Computational Biology

Background:

• Life scientists face information overload, making it difficult to track all relevant research.
• Text-mining tools are emerging to help identify topics and extract statements from scientific literature with improved accuracy.
• Researchers organize extracted statements into ontologies, which are systems of reality claims.

Purpose of the Study:

• To review existing work on modeling scientific reasoning.
• To suggest methods for computational reasoning across ontologies to assess statement certainty.
• To explore incorporating social dependencies and cultural biases into reasoning models using digitized scientific network data.

Main Methods:

• Reviewing computational models of scientific reasoning.
• Proposing computational reasoning approaches across ontologies and textual statements.
• Analyzing digitized data on scientific authorship, institutions, and resources to account for social and cultural factors.

Main Results:

• Computational reasoning can identify internal inconsistencies within ontologies in bioscience.
• The approach has the potential to assess the certainty of scientific statements and the process of their validation.
• Digitized network data allows for the consideration of social dependencies and cultural biases in reasoning models.

Conclusions:

• Computational reasoning can help manage the complexity and inconsistencies in text-mined scientific information.
• Future applications include identifying underexplored research areas, reviving forgotten findings, and synthesizing novel hypotheses.
• Integrating reasoning models with social and cultural context can enhance the reliability and utility of scientific knowledge discovery.