Nibiru – GAIuS 1.0 Release
February 19, 2012GAIuS open-source Beta version has been available for over a year now. The tremendous positive feedback from it and many feature and function requests have been undertaken and developed into the upcoming GAIuS 1.o codenamed “Nibiru” release. (Read more about this breakthrough release here.)
Why is this version codenamed Nibiru? Popular pseudo-scientific fringe-culture has declared that the world is coming to an end on December 21, 2012 when the mythological planet “Nibiru” will catastrophically collide with the Earth. As the neo-folklore goes, this apocalypse is due to coincide with the end of the Mayan calendar. Never one to give up on an “end of the world” yarn, this sensational premise has been adopted by Hollywood into the screenplay, “2012″, further embedding the myth into the zeitgeist.
Interestingly enough, the term “nibiru” translated from Akkadian means “crossing” or “point of transition”. At Intelligent Artifacts, we believe this new version of GAIuS is a crossing or point of transition. We also are very (overly!) aware of critics of our technology believing that it will end with a robot apocalypse. So, the name “Nibiru” fits perfectly with our ideas about changing the world…and satisfies our twisted sense of humor!
When is the GAIuS Nibiru release due? Well, December 21, 2012 of course!
In the meantime, scientists, researchers and developers can request an early pre-release version of Nibiru by asking for an invitation at our GAIuS Google Group.
Posted in News | Comments OffDiagnosing Heart Disease with GAIuS
December 30, 2011According to the CDC, “Heart disease is the leading cause of death in the United States”.
In response, medical cardiology imaging has undergone rapid technological advances during the last several years, producing increasingly complex information, including three-dimensional time-varying echo cardiogram and MRI data. These advances and data explosion place ever greater burden on the physician for correctly reading and interpreting the information.
In practice, the majority of a doctor’s examinations are limited to either certifying the absence of evident pathology, or certifying evidence of a clear pathology that requires a specialist’s further reading. The medical doctor’s expertize is effectively utilized in only a low percentages of patient cases. Examinations are costly, have long waiting lists, and are time consuming to patients and care givers alike. The medical sciences have long recognized the pressing demand for an intelligent and interpretative machine capability for performing rapid and automated preliminary screenings of exam data.
Medical imaging tools like the Doppler Echocardiography were quantum leaps to help to identify symptoms of heart disease. Echo cardiogram typically focused on imaging the patient heart and cardio related tissues. Echocardiogram advances provided for imaging the velocity of blood flow through the chambers of the heart. The image below demonstrates blood flow using color to represent the speed and direction of blood fluid dynamics. In healthy hearts, blood flows smoothly. In diseased hearts, the flow is turbulent.
Intelligent Artifacts is pleased to be working with Dr. Gianni Pedrizzetti of Advanced Medical Imaging Development, creating break-through imaging analysis for diagnostics and fraud detection.
Dr. Pedrizzetti is an Italian engineer who currently works as a professor in fluid mechanics at the University of Trieste. He is one of the first scientists to apply fluid mechanics to improve cardiovascular sciences and introduced optical flow techniques for the benefit of obtaining dynamic information from medical images, with the goal of improving heart diseases diagnostics. Pedrizzetti has published over 60 articles on peer-reviewed international journals, and is co-inventor of several patented solutions in the field of medical imaging and medical data processing. He is an author of a monograph which relates basic theoretical principles in bioengineering and fluid mechanics to the cardiovascular problems and medical solutions.
Dr. Pedrizzetti challenged Intelligent Artifacts to build a GAIuS agent to automate echo cardiogram fluid dynamics analysis. A sampling of 17 patients’ ECG videos were provided for the tests. The goal was to determine if a GAIuS agent’s diagnosis matched a trained doctor’s diagnosis of the same data.
Was GAIuS as good as a trained professional? Preliminary results are more than promising. The agent’s diagnosis of each image frame was consistent with a trained cardiologist’s diagnosis. Has Pedrizzetti changed his opinion of AI? “GAIuS seems to be the smart way of solving these problems. It makes sense.”, Pedrizzetti said after seeing the results.
These results have prompted deeper research in using GAIuS for medical diagnosis as part of a doctor’s arsenal. A GAIuS agent can assist lesser cardiology trained physicians in triage, particularly in remote rural areas of underdeveloped countries where handheld echo cardiographs are the most advanced heart diagnostics that a patient may ever see.
GAIuS assisted diagnostics can also help prevent misdiagnoses or uncover fraud. Fraud is a major medical concern currently receiving greater attention by the medical community.
In July 2010, a radiologist was found guilty of health care fraud. The U.S. Department of Justice has released this report stating,
Dr. Rajashakher Reddy Signed Tens of Thousands of Radiology Reports for over a Dozen Hospitals, Where Neither He Nor Any Other Radiologist Reviewed the Films
According to evidence presented, the radiologist had submitted diagnosis on thousands of medical reports without reviewing the X-rays, mammograms, or CT scans between May 2007 through January 2008. Often, his untrained employees would sign reports on his behalf using his electronic password.
With a GAIuS intelligent agent able to reviewevery report, such fraud can be discovered using context-based pattern matching techniques. More importantly, lives can be saved.
How difficult was it to create this echo cardiogram agent using GAIuS?
The GAIuS application layer consists of functions that read the echo cardiogram video file and extract image frames. Each image frame is converted into a matrix that is fed into the intelligent GAIuS layer.
The GAIuS intelligence layer uses only two Manipulatives (a fundamental GAIuS concept) from the standard GAIuS library. These GAIuS Manipulatives are connected in sequence to a single GAIuS Primative or node (Primatives, the 2nd GAIuS fundamental concept). GAIuS architecture is a graph collection of nodes (Primatives) and edges (Manipulatives).
After consulting with Dr. Pedrizzetti, GAIuS development time for actually creating the echo cardiogram application was less than one hour. This enormous efficiency was possible because of the availability of tested and proven GAIuS components.
What can you create in that time using GAIuS? We’d welcome the opportunity to discuss it with you.
Posted in News | Comments Off