AI Solutions in Industries
MEDICAL
Clinical Graph
For patients on a clinical trial who are hospitalized, this assigns a micro function model to each patient and creates daily graphs of statistics available for that patient. More complex alerting scenarios covering combinatorics of patient vitals are easily supported. The researcher can now obtain a daily graph of vitals and other statistics rather than isolated data points. Although subtle, this can make a meaningful difference. The inherent one-shot learning makes it simple to integrate the model with a feed of instrument data so the model learns dynamically.
mQSAR
QSAR (Quantitative Structure-Activity Relationship) is a molecular analysis predictive approach used by medical researchers. It helped find the COVID vaccine quickly and has aided in cancer drug discovery and many others. mQSAR (multi-dimensional QSAR) has been problematic in AI, one reason being that the parameters are very complex. Functor models can handle complex parameters as the model learns by function changes, not the parameters. It is therefore promising that mQSAR will be feasible, and we are working with QSAR manufacturers to help make this a reality. Once mQSAR is in place we could see meaningful progress in drug discovery. There are no guarantees, but estimates indicate this is achievable.
CYBERSECURITY
BioGuard — Halt ATO
BioGuard is a behavioral biometrics SDK for mobile banking apps. It is unique in that all operations run on the device. There is no exposure of sensitive personal data via calls to cloud systems or anything external to the phone.
Fraud Detection
The annual cost of fraud globally exceeds $5 trillion. US banks are losing billions annually. We specialize in real-time fraud detection systems and have designed an auto-evolving system. Please contact us if you have needs in this area.
ENVIRONMENTAL
The Integration of Bioinformatics and Cheminformatics
Nobel Laureate Walter Gilbert had the brilliant idea that we could provide an integration platform for bio and cheminformatics that would enable personalized medical research. Lion Bioscience undertook the implementation.
The same idea carries over into biofuels and agricultural research. Rather than human DNA, we look at plant DNA and seek alternatives to fossil fuels. The platform is also easier to achieve, as there are fewer applications to integrate and data is not restricted in the manner it is in medical research.
Autonomic AI is investigating this space and seeking opportunities to motivate development.
EDGE
Functor models created by Autonomic AI deliver significant advantages for Edge AI:
-
Edge AI devices are battery dependent — smartphones, EVs, robotics and others all run on battery power. Functor models use less energy and have less impact on the critical battery that powers the device.
-
Functor models are capable of batch learning but are naturally suited to one-shot learning, meaning they can learn unit by unit from streaming data.
-
The Functor Model Architecture supports many model types. Functor Micro Models are ideal for edge AI deployments.
FINANCE
The Autonomic AI approach to AI in finance emphasizes transparency and simplifies auditing so that regulatory requirements can be satisfied — in contrast to other approaches that leave auditors and monitors in a cloud of obscurity. Specific areas of applicability include:
-
Fraud Detection
-
Insurance Actuarial
-
Stock Market — Robo Advisors / Algorithmic Trading
-
Banking
-
Automated Reporting
-
Cybersecurity