Intelligent Artifacts Enables Safe And Secure Incorporation Of AI/ML Into Safety-Critical Systems

SSupported by cloud service provider DigitalOcean – Try DigitalOcean now and receive a $200 when you create a new account!
Listen to this article

Below is our recent interview with Emily Mathwich, Communications Manager at Intelligent Artifacts:

Q: Emily, what is Intelligent Artifacts?

A: Intelligent Artifacts is a leader in machine intelligence dedicated to enabling advanced applications with solutions built from our groundbreaking Artificial Intelligence / Machine Learning framework, called GAIuS™, that combines human-like Reasoning (AI/ML/R) with machine speed and repeatable precision. Solutions built from our GAIuS framework pair complexity with transparency by solving a range of difficult problems while still allowing users to understand why, how, when, and where actionable information is generated and make changes to source data without retraining the entire system. The GAIuS solution adheres to the ExCITE paradigm, a technology-agnostic evaluation methodology focused on transparency in AI solutions that gauges the level at which a solution can Explain, Compute, Interpret, Trace, and Edit (ExCITE) its outputs and internal processes. Intelligent Artifacts provides the complete intelligence layer for customers in need of modular, flexible, and scalable intelligent systems. Our AI/ML offering is designed as a trustworthy, transparent, and effective solution for mission and safety-critical applications as a pathway to achieving human-level intelligence.

Q: Any highlights on your recent announcement?

A: We are very excited to continue working through the DO-178C compliance process to become the first certifiable, reusable software component to enable AI/ML for avionics systems. Unlike other technologies working toward certifiability, the GAIuS solution provides a general-purpose AI/ML/R engine that does not change from application to application (i.e., is reusable). This allows for the certification effort to be amortized across customers or projects. Some aspects will need to change from platform to platform, but that does not require a complete re-certification of the engine each time.

The fact that we, today, have a clear path to certifiability as a reusable AI/ML/R software component speaks volumes. Neural-net-based technologies are not certifiable under existing DO-178C guidance due to their well-known deficiencies in explainability and traceability. As a result, industry trade groups are pushing to modify existing software requirements so that their neural-net-based solutions can achieve certification. However, changing the test to receive a passing grade is a poor strategy for assuring safe AI/ML systems. Behind this thought process is the misconception that neural nets are the only viable machine learning technology. This isn’t the case. Our recent announcement is proof of viable AI/ML alternatives. We expect GAIuS to become the aerospace industry standard for AI/ML applications.

Recommended: BlueThrone Raises Seed Funding To Acquire And Grow iOS And Android Apps With Proven Market-Fit

Q: Can you give us more insights into your offering?

A: Rather than engineering AI for specific applications, we have built a framework that allows intelligence to emerge, learn and evolve similarly to other biologically intelligent creatures. Each future use case benefits from what our system has previously learned from past use cases within an enterprise environment. Our system is structured in such a way that we are able to evolve our solutions within these environments using genetic algorithms. These configuration files, or genomes, which act as the DNA of our solutions, allow our system to continuously learn in real-time, reduce manual labor hours, reduce, eliminate, or discover conscious and unconscious bias, and provide optimized automated solutions.

Intelligent Artifacts’ offering enables the safe and secure incorporation of AI/ML into safety-critical systems at a level no other company is yet capable of matching or surpassing. Our symbolic, deterministic AI solution is able to present information and processes in human-understandable terms, return information in a specified amount of time for Real-Time Operating Systems (RTOS), and give stakeholders confidence that errors are unlikely and can be explained, traced to their root cause, and definitively remedied (ExCITE AI). These features render our system “regulation-ready” and unlock the possibility of more extensively autonomous systems. Additionally, our system processes data in real-time providing decision-makers with the most up-to-date, actionable information. Because we scale vertically as well as along the horizontal these capabilities are not limited by use-case and can be applied to logistics management, sensor fusion, ISR, predictive maintenance, machine vision, and more.

Q: What can we expect from your company in the next 6 months? What are your plans?

A: We are demonstrating our solution for Avionics at Aerospace Tech Week Americas in early November. Our team has constructed a sensor fusion rig with an embedded GAIuS agent running on an RTOS which will demonstrate object detection/recognition, face and emotion recognition, path planning and navigation as a means to gain more information, and a live stream of our reasoning engine at work to highlight our ExCITE features.

Within the next six months, we intend to complete the process for DO-178C compliance which will culminate in a DO-178C Cert-kit; the complete documentation of our certifiable AI/ML solution validated as a component for Avionics systems looking to be certified by the FAA for reuse across various Aerospace applications and platforms. From there, we will continue to work closely with the Aerospace, Defense, and Intelligence Communities to incorporate our AI/ML/R to meet a spectrum of needs across agencies and organizations.

Recommended: An Interview With Trevor Hewes, The Founder Of TP2 Products

Q: What is the best thing about your company that people might not know about?

A: The team at Intelligent Artifacts shares a collective vision for the future of AI/ML/R and we are acutely aware of our responsibility to keep transparency at the forefront of innovation. We believe that transparency is a powerful tool in illuminating AI shortcomings, like conscious and unconscious bias, and a lack of explainability, as well as highlighting the capacity for good that comes with ExCITE AI/ML. This is why we built a reusable AI/ML/R solution, defined here as a system consisting of repeatable functional units that is capable of scaling vertically and along the horizontal without much re-tooling if any. Everything we create, every use case we explore, and every new customer relationship under consideration is filtered through our ultimate goal of evolving our system to achieve human-level intelligence.