Machine Learning for Logical Processes

useAIble's RLM machine learning system uses a non-statistical approach to machine learning. While stats are great for tracking patterns over time, the RLM is designed instead to be used for business logic. Now you can use machine learning to solve problems where humans or processes introduce problems that can't be described by statistics.

Why is the RLM Better?

pie graph

useAIble’s RLM non-statistical approach to machine learning (ML) creates a paradigm for solving logical and process-driven problems not solved by traditional methods.

line graph

useAIble’s RLM performs tasks with ML in near real-time and with much lower CPU/GPU requirements.


The RLM is fast and easy to set up. No special scientists or expensive experimental teams are required.


useAIble’s RLM has the ability to adjust to changing dynamics in unprecedented ways.


useAIble’s RLM can explain every decision it makes for simple and clear diagnostics.



vs. Traditional Machine Learning

Test conducted with Walmart for product shelf placement with complex constraints.


vs. Traditional Code

Test conducted with an $8 billion third party logistics company to compare with traditional code-based solutions in a packing problem.


How Does the RLM Work?

pie graph

Traditional neural networks use statistical models based upon what are called “weights” (statistical aggregations). The RLM uses a completely different model. SmartNeurons™ use every decision ever made by the RLM to determine the best next decision. This results in the ability to solve problems that could never be solved with statistical aggregations.

line graph

Traditional neural networks (TNN)use random convergence to allow neurons to converge on a solution. This process requires an exponentially increasing amount of time and hardware, since for each learning cycle all of the neurons in the TNN must be recalculated. When problem complexity increases these calculations skyrocket. The RLM’s patent-pending SmartNeurons™ use a pre-assignment system allocate an individual SmartNeuron™ to each specific part of a problem. This means that during each learning cycle only one SmartNeuron™ must calculated instead of every neuron. In small networks the difference is minor, yet in large networks SmartNeurons™ eliminate billions of required calculations.


TNN systems are unable to explain how or why they select make a decision. This might be acceptable for a search engine; however, a jet engine requires a higher level of certainty. The RLM tracks and explains every decision. useAIble provides a package of decision diagnostics that are simply not available in TNN.


The Code

Open Source Limited Use

Machine Learning Head to Head

White Papers


- Patents, Awards, Press, Etc.-

useAIble's ground breaking technology has won awards
and provoked speaking invitations at higher education institutions.