Typically, errors in Artificial Intelligence take a substantial amount of time to resolve. New research from the University of Leicester has come up with a novel method to resolve errors in Artificial Intelligence immediately.
A team of Researchers from the University of Leicester’s Department of Mathematics have published a paper in the journal Neural Networks outlining mathematical foundations for new algorithms capable of allowing Artificial Intelligence to gather error reports and rectify them instantly without disturbing existing skills - at the same time collecting corrections which could be used for updates or future versions.
This could basically offer robots with the ability to correct errors promptly, effectively ‘learn’ from their errors without damage to the knowledge already acquired, and eventually share new knowledge amongst themselves.
Along with Industrial partners from ARM, the algorithms are integrated into a system, an AI corrector, capable of enhancing performance of legacy AIs on-the-fly (the technical report is can be obtained online).
ARM is the largest provider of semiconductor IP in the world and is the architecture of choice for over 90% of the smart electronic products being built currently.
Multiple versions of Artificial Intelligence Big Data analytics systems have been deployed to date on millions of computers and gadgets across various platforms. They are functioning in non-uniform networks and interact. Industrial technological giants such as Amazon, IBM, Google, Facebook, SoftBank, ARM and many others are involved in the development of these systems. Performance of them increases, but sometimes they make mistakes like false alarms, misdetections, or wrong predictions. The mistakes are unavoidable because inherent uncertainty of Big Data.
Professor Alexander Gorban, Department of Mathematics, the University of Leicester
Gorban also stated, “It seems to be very natural that humans can learn from their mistakes immediately and do not repeat them (at least, the best of us). It is a big problem how to equip Artificial Intelligence with this ability... It is difficult to correct a large AI system on the fly, more difficult as to shoe a horse at full gallop without stopping."
“We have recently found that a solution to this issue is possible. In this work, we demonstrate that in high dimensions and even for exponentially large samples, linear classifiers in their classical Fisher's form are powerful enough to separate errors from correct responses with high probability and to provide efficient solution to the non-destructive corrector problem.” said Gorban.
There is a frantic need in an economical, fast and local correction process that does not damage significant skills of the AI systems during the course of correction.
Iterative techniques of machine learning for Big Data and huge AI systems are never economical and thus the Researchers propose that the corrector should be non-iterative with the reversible correctors necessary to reconfigure and combine local corrections.
It is often infeasible just to re-train the systems for several reasons: they are huge and re-training requires significant computational resources or long time or both; it may be impossible retrain the system locally, at the point where mistake occur; and we can fix one thing but break another leading to that important skills could vanish... The development of sustainable large intelligent systems for mining of Big Data requires creation of technology and methods for fast non-destructive, non-iterative, and reversible corrections. No such technology existed until now.
Dr Ivan Tyukin, Department of Mathematics the University of Leicester
The Researchers have discovered and demonstrated stochastic separation theorems which provide tools for rectification of the large intelligent data analytic systems.
With this method, immediate learning in Artificial Intelligence could be conceivable, providing AI with the ability to re-learn following an error after a mistake has transpired.
The study has been partly supported by Innovate UK through Knowledge Transfer Partnership grants: KTP010522 between Visual Management Systems Limited and the University of Leicester and KTP009890 between ARM/Apical Ltd and the University of Leicester.
The Knowledge Transfer Partnership (KTP) scheme assists businesses to modernize and grow. It does this by connecting them with a University and a Graduate to undertake a specific project.
Having a system like this is indispensable in large-scale deployment of AI services to customers and end-users. Customer-specific usage of AI enabled devices gives rise to customer-specific errors, which is perceived by the end-user as non-acceptable situation. Retraining core AI to deal with these errors is technically challenging and potentially risky. Newly trained AI, taught to avoid specific errors may exhibit unexpected behavior in another situation. In this scenario, the scale of a problem grows exponentially as the size of AI deployment grows, making it practically impossible.
New technology enables to remove these obstacles altogether, making AI enabled devices collaborative in error removal process. This new quality allows large AI enabled deployments to become more intelligent as their size grows. In practice it means that error-less AI powered devices become reality. Recently we filed a patent application to secure our priority in this area.
Dr Ilya Romanenko, Director of R&D, Computer Vision, Imaging and Vision Group, ARM