ChOS Project 9
Medi-linker
This is a large two part systems biology project. Systems biology is
basically computer aided simulation of living things. Projects first
part attempts to create an AI utility which would read collections of
Medline abstracts. AI-utility then attempts to identify patterns
within sentences such as "X causes Y in patients with Z". The most
challenging AI- task is making utility to learn new patters for saying
the same "patients suffering from Z have often been shown to have
diagnosis Y, which is commonly caused by X". If the AI-utility is able
to learn new ways for saying medical 'truths', the sentence patterns
can be used in harvesting many different medical 'truths'. The target
of Medi-Linker- projects first part is to convert found 'truths' to a
logical uniform format which can be easily sorted and searched by
computers. Medi-Linker-projects second part is a utility, which
visualizes 'models' connections found in part-1. Part-2 utility may
also be equipped with simple AI which tries to find/suggest common
biomedical reasons for observed symptoms. Medi-Linker project is
useful in multiple ways; processing current medical data archives is
too massive task for humans small brain, project has potential to
quickly finding correct references, project may significantly help
identifying reasons behind various illnesses and the project might
also be able to suggest other beneficial use for medicines already in
use.