Pontimax Technologies Announces Its Innovative Dynamic Meta-Inferencing (DM-I) AI Technology

Pontimax Technologies LLC announced its Dynamic Meta-Inferencing (DM-I) Technology, an innovative approach to predicate reasoning-based Artificial Intelligence (AI). Combining Machine Learning (ML) inferencing with Spatiotemporal Contextual Inferencing (SCI), it provides a new capability for user-specified, knowledge-based, generative predicate dynamic reasoning.

Ray Keating, Managing Member of Pontimax Technologies, described the core of DM-I as providing “a capability for “Inferencing Goal Events” to be defined by the user in terms of their constituent “Predicate Fact Pattern” (PFP). Based on those PFP meta-specifications, the Fact Pattern is dynamically generated, invoked and evaluated using the DM-I Agent’s Inferencing Engine. Importantly, these Inferencing Goal Predicate Fact Patterns can be composed of ML or SCI Predicate Fact Patterns.”

Read: ChatGPT Is Much More Than A Data Prep Assistant

Mr. Keating provided two examples to illustrate how DM-I works in stating “Let’s take a scenario where you’re conducting a military ground mission and there are motion sensors placed along the ingress route and a tasked overhead video sensor platform. An “Anti-Ambush” Inferencing Goal Event could be specified with two PFPs: a Motion Occurrences Pattern-of-Interest and an Object Recognition Machine Learning PFP. Based on the Location and Time Contexts specified, the DM-I Agent will initiate the specified predicates, starting with checking for qualifying motion data. The DM-I Agent then applies its built-in Spatiotemporal pattern recognition capability to determine if the Motion Occurrences Predicate Fact Pattern has occurred. If so, it sends user-specified Consequent Action message to the overhead Video Sensor platform as to the location of inferenced motion pattern and then initiates the “chained” Object Recognition PFP. If the PFP result qualifies per a user-specified threshold, the “Anti-Ambush” Goal Event is inferenced and its user-specified Consequent Actions (CA) are launched; besides recording the Event inferencing data details, one of those CAs would be to send a Threat Alert message to the Mission Lead.”

Latest Insights: Leverage Artificial Intelligence Tools to Reduce Fraud and Friction for Customers

He also described a Network Cyber Security scenario illustrating how the “DM-I technology can effectively address “Data Firehose” Flood-of-Data environments through use of a low-resource IP Originating Address Occurrences Fact Pattern, followed by a resource-intensive Machine Learning-based Suspicious IP Packet Classification Fact Pattern. The former would use the Spatiotemporal Occurrences Fact Pattern to identify a suspicious set of originating IP address “locations” then initiate the “chained” ML Classification Fact Pattern. These Use Cases illustrate how Dynamic Meta-Inferencing provides domain experts with the tools and harness to apply their knowledge to “real world” scenarios while also demonstrating the ability of Pontimax’s DM-I technology to deliver Sensor-to-Shooter. or, more broadly, true “Data-to-Decision” capability.”

AiThority: The 3 Building Blocks to Make AI Accessible

[To share your insights with us, please write to sghosh@martechseries.com]