Our mission is to create automated solutions with human-like powers of reasoning that magnify the talents and capabilities of people. Our solutions provide better information to reduce risk, identify opportunities, and aid human decision-making. Beyond Limits is the only AI company in the world with advanced technology proven in extreme environments from 150 million miles into space for scientific research to 15,000 feet underground for oil and gas exploration. Our breakthrough cognitive technology goes beyond conventional AI, blending deep learning and machine learning tools together with symbolic AIs that emulate human intuition to produce our cognitive intelligence.
The Cognitive Leap - How it Works
Beyond Limits technology is a cognitive leap beyond conventional AI to a human-like ability to perceive, understand, correlate, learn, teach, reason and solve problems faster than existing AI solutions. Core to Beyond Limits intelligent systems is the symbolic reasoner. A cognitive engine that uses the outputs from sensors and neural nets and uses its education to reason about what it sees. This is a new approach to reasoning, based on opportunistic self-discovery monitoring, bi-modal cognitive-based reasoning, and autonomous self-discovery to resolve ambiguity. The power of the system is demonstrated as it autonomously shifts through corridors of information to discover plausible facts and scenarios from interpretations of diverse data while avoiding the usual computational complexities of traditional systems through a technique called autonomic monitoring that enables serendipitous discovery during cognitive forensic analysis. Our philosophy is that the brain is composed of distinct but interacting modules that self-organize to solve problems using a complement of local learning (“training”) and innate knowledge (“education”). So, our systems are designed to be trainable and learn on the fly, autonomously.
Two Sides of the Cognitive Brain
From the Beyond Limits perspective, AI can be divided into two distinct areas: Numeric-based AI and Symbolic-based AI.
- Numeric-based systems (machine learning, neural nets, and deep learning, etc.), are trained from thousands of examples to reduce a problem to sophisticated pattern matching rather than a reasoning-based process. They are wonderful when you have lots of data to train a system. These systems, unfortunately, are “black boxes” that cannot explain how they arrive at their answers. So they don’t know what they know anymore than they know what they don’t know.
- Symbolic-based AI systems are educated from declarative knowledge (case-based reasoning, rule-based inference, etc.) much as a person would be, and deeply think about each fact that is presented to them. Symbolic systems can reason through ambiguity and missing information with human-like intuition and can provide a detailed explanation how they arrived at their answers (AKA explainability). Symbolic systems can also make changes to their knowledge base in real-time. However, because they spend so much time processing each piece of data, our approach is to combine them with numeric-based AI to reduce the problem to labels.
Both approaches are valuable AI techniques and typically both are required for complex problems that involve massive, disparate data sets and high-value assets like in energy, fintech, and healthcare.
Explainability in Cognitive AI
AI systems are built and deployed to deliver actionable information to solve problems. Trust in the system’s recommendations can only be ensured if the system can explain its reasoning. Because Beyond Limits customers use our cognitive solutions to analyze and support high value decisions, our AI systems are designed to explain their thinking. Unlike conventional “black box” approaches like machine learning, deep learning or neural networks that cannot explain their reasoning, all Beyond Limits cognitive AI solutions are capable of delivering clear explanations of reasoning and evidence in transparent audit trails, including risk, uncertainty and facts.
Conventional AI Problems
Conventional deep learning & neural networks can learn from data, but cannot easily explain their answers. Conventional AI approaches are brittle, and cannot handle new, never seen-before problems.
Our Cognitive systems are educated, and trained, leveraging deep learning insights plus expert human knowledge to solve problems even with missing or misleading information. Beyond Limits systems can explain results in terms of human-like reasoning and evidence
Download our Explainable AI whitepaper here.