Company Phone:
Company Email:
Company Address:
1040 Thornridge Ct, Brookfield, WI 53045
Company Purpose:
Compsim's mission is to be "the leading provider" of technology that will allow its partners/customers to incorporate human-like judgment and reasoning into devices and software applications. Compsim's objective is to have its technology deployed globally for the benefit of mankind.

Compsim, a privately owned R&D company since 1999, and a 100% woman-owned small business, is a "Technology Provider." Compsim can team with organizations that have operational expertise and platform expertise so those organizations can package their expertise for delivery in devices or software applications.

Compsim licenses KEEL (“Knowledge Enhanced Electronic Logic”) Technology to companies that want to obtain a competitive edge. KEEL allows Subject Matter Experts to model and deploy their reasoning (decision-making process) in applications and in devices without extensive support from mathematicians and software engineers. When deployed, systems can dynamically adapt to the environment (make or suggest decisions and/or actions).

Compsim's patent portfolio covers:
• A dynamic graphical language (KEEL Toolkit)
• A model for accumulating supporting and objecting arguments in order to make a decision, or take an action
• An architecture for small footprint engine(s) (KEEL Engine) that processes sensors or other inputs according to the design of a system created in the KEEL Toolkit
• A model for implementing a KEEL Engine as an analog circuit

Helena Keeley, CEO; BSE (Electrical Engineering); 23 years of development engineering and management experience at GE and Rockwell.

Tom Keeley, President and Technical Visionary; BSE (Electrical Engineering); 35 years of development engineering, engineering management, and corporate strategy development experience at GE and Rockwell.



Competitive Advantage:
• Since ANN is a pattern matching approach for cognitive solutions, it could present a problem when it is important to validate the decisions or actions.
• KEEL logic can be easily visualized and it is 100% explainable.
• With pattern matching there is no “reasoning” taking place other than to suggest that there is some degree of a matched pattern.
• A KEEL system evaluates information based on “policies” to make decisions and is not based on patterns.
• ANN solutions do not react well to “surprise”.
• KEEL solutions are based on human reason, not trained images, and can therefore “decide” on the spot how to respond.
• ANN systems are difficult to train.
• Using the KEEL Toolkit, a reasoning model can be defined graphically. It “thinks” as the person defines it, so creating and debugging the process is easy.

Fuzzy Logic
• Based on the concept of linguistic uncertainty.
• It uses a geometric Fuzzification / Defuzzification process that might be considered an “art” rather than a science.
• Ask a Fuzzy Logic Designer “Why” they choose a particular geometric fuzzification scheme???
• KEEL uses visible supporting and blocking information and visible relationships between data items to define outputs or actions –there is no fuzzification / defuzzification step.
• Fuzzy Logic is difficult to explain or audit in human terms.
• KEEL is 100% explainable.
• Look at the wires and the importance of information and “see” why decisions or actions are made.

Bayesian Belief Nets
• Based on probability.
• If one has “good data” then probability based answers can provide very useful information.
• Works well on well-understood problems.
• KEEL defines functional relationships between data items. When probabilities are known, they can be used to drive KEEL engines.
• BBN may be difficult to explain or audit in human terms.
• Garbage in = Garbage out
• Difficult to collect good statistics on complex systems
• Dynamic systems are difficult to evaluate and gather “good” statistics
• There is a temptation to blindly believe the statistics.
• KEEL is 100% explainable
• Look at the wires and the importance of information and “see” why decisions or actions are made.

Rule Based Systems
• Text Based Forward and Reverse Chaining
• May not be conducive to real time control applications
• Conventional Textual Rule-based systems can be used where the rules are fixed and the impact of each rule is stable
• The world is rarely simple! KEEL can be used to define systems which are complex and involve the dynamic changing of information value.
• KEEL can define non-linear systems

Differential Equations
• Complex mathematical formulas may require special expertise to develop.
• KEEL Technology allows complex models to be created without complex mathematics.
• Systems utilizing complex mathematics may require multiple levels of knowledge transfer (domain expert, to mathematician, to software engineer / systems engineer) offering multiple opportunities to corrupt understanding
• A domain expert using KEEL may be able to give a working / tested KEEL Engine to the system integrator without the intermediaries.

Hard-Coded Logic
• The brute force method of creating solutions to complex problems is to use conventional IF | THEN | ELSE logic that has the potential of creating large monolithic, unmanageable chunks of code that is prone to human error and hard to debug.
• KEEL Engines are created automatically from the design.
• The graphical language allows the designer to “see” the system think.

Machine Learning/Deep Learning
• The major focus of today’s Machine Learning systems (such as IBM Watson) is to generate knowledge from reading, and attempt to understand unstructured textual documents. These tend to extract features to drive Artificial Neural Net based systems.
• KEEL-based systems are human-driven systems
• KEEL can be used to audit the Machine-Learning systems
• KEEL-based systems DO NOT LEARN, instead- humans can tell them how they want them to ADAPT.
KEEL-based systems are under the control of humans.
• Machine Learning systems “could” provide inputs to KEEL Engines (IF that is what the humans-in-charge want to do).
Favorite Quote:
Hope for the best, be prepared for the worst, expect everything, and accept nothing as final. – Tom Keeley
Summary of Best Use Cases:

ANEMIA – (Link) This was an actual project defined by a Pathologist. People still use our website and input their yearly physical blood test results. (Documentation is located below the demo) Medical Demo #1
NATO- MCDC #1 – (Link) This applies to when to shoot and when not to shoot, and with what ordnance.  Military Demo #10
NATO #2 – (Link) “Berlin 2015.”  See movies #16-19
Demo project involving policies of the Royal Air Force (Canada) – (Link) This is regarding Mammals, specifically when they do NOT want to use their sonar when mammals are in the area in which they are flying. See movies #5 
NATO related demos(LinkNATO wrote and published a book on Autonomy - Tom Keeley wrote chapter 9. “Auditable Policies for Autonomous Systems (Decisional Forensics).”


Artificial Intelligence
Big Data Analytics
Cloud Computing
Cognitive Computing
Emotional Intelligence
Intelligent Agents (Primary)
Intelligent Decision Support
Internet of Things
Leadership at the Intersection of Business and Technology
Reasoning and Decision Automation
Thought Leader
Vision-based Sensing and Image Recognition