top of page

What We Do Best

We help commercial and government customers to scale and expand operations by connecting their existing capabilities and systems with emerging technologies, such as:

  • Information models that provide software with meaningful context.

  • Software agents capable of reasoning about information in a context model.

  • Decision-support and data analysis capabilities with embedded intelligence.

  • Machine Learning algorithms leveraging advanced AI methodologies.

  • Reusable, distributed, semantic web services in an adaptable architecture.

  • Systems that take advantage of Cloud capabilities but are Cloud-Provider agnostic.

What is a Virtual Model of Context?

For a computer to be an intelligent assistant to the human user it must have some level of understanding of the context of the information that it is processing. This can be achieved by incorporating in the software a virtual information model that defines not only the properties of the objects that play a role in the application domain (e.g., trucks, trailers, drivers, roads, loading docks, and so on, in the transportation domain) but even more importantly the relationships between those objects. There are technical strategies available for constructing and subsequently leveraging such information models as the context-enabled foundation for computer-based reasoning. CHOBU has a great deal of experience in constructing such virtual information models and developing software agents that are capable of reasoning about the data flowing through the system in the context provided by the virtual information model. The result is a powerful human-computer partnership with a surprising level of artificial software intelligence. 


What are Intelligent Software Agents?

Software agents are coded modules with built-in reasoning capabilities. A simple example of a reasoning capability is an if … then statement. For example, in the transportation domain: if the locational coordinates received from the truck-mounted GPS device have not changed for the past 12 minutes, then the truck has been stationary. The combination of a sequence of such if … then statements with the context provided by a virtual information model leads to a powerful automated decision-assistance capability. CHOBU’s software solutions typically incorporate multiple agents, each focused on a particular aspect of the application domain. For example, CHOBU’s iCADS+software system for building design includes Daylight, Electric Light, Noise Control, Room Acoustics, and Thermal agents, each with expertise in its focus area. Operating in background the agents continuously monitor the evolving design solution, collaborating with each other while alerting the human designer to conflicts and assisting in the resolution of such conflicts. 


What is the Nature of Software Intelligence?

Computer intelligence is not the same as human intelligence. Whereas human intelligence is not only logical but often to an even greater extent influenced by emotions and intuition, computer software is entirely logical. Therefore, the level of intelligence that can be embedded in software depends almost entirely on the accuracy of step-by-step reasoning sequences and the richness of the context provided by the virtual information model that describes the application domain. While the formulation and translation of a step-by-step reasoning sequence into software code is a relatively easy task, the construction of the virtual information model that can be processed by the computer into a form that provides the software system with a useful understanding of the context of the application domain is a more complex undertaking. The CHOBU team has deep expertise in this field and has built many such models with considerable success during the past 20 years. 

Embedded Intelligence

An algorithm is no more than a mathematical equation that can be as simple as a=b+c  or much more complex involving calculus and/or statistical analysis. The approach in Machine Learning is to have algorithms sift through very large volumes of data and find hidden relationships and emerging patterns that suggest trends. For example, most large retail chains apply Machine Learning algorithms to their past and current sales data to discover sales patterns based on store layout alternatives. The corpus of data is typically divided into two parts. One part is used to find patterns and the other part is used to validate that the apparent pattern actually exists and is not just a coincident. There are essentially five kinds of algorithms in Machine Learning: connectionist algorithms in neural networks; genetic algorithms that are based on Darwin’s theory of natural selection; Bayesian algorithms that start with the probability that a certain hypothesis is correct; analogy algorithms that look for an analogous example; and, symbolic algorithms that represent objects as symbols that can be manipulated logically.


What are Machine Learning Algorithms?


What is an Adaptable Architecture?

With the rate of change of business processes accelerating in recent years it is important that the software systems that support those processes are readily adapted to those changes. If not, then the continued efficiency and competitiveness of the business entity is at risk. An adaptable or flexible architecture ideally consists of small self-contained modules (often referred to as microservices) that are fine-grained in scope, independent of each other, comply with standardized interface protocols, and can be replaced without jeopardizing the operation of the entire system. Such architectures tend to be more fault-tolerant and scalable than their more traditional counterparts and are steeped in service-oriented design principles. 

Adaptable Architecture

Cloud, yes, but why Agnostic?

Taking advantage of the on-demand scalability, fail-over and back-up capabilities, data analysis tools combined with massive computing power, superior data management and storage facilities, built-in security, and ease of user access, cloud-based computing has become a commonsense decision and is now the preferred choice in most cases. The transition to remote work fueled by the COVID pandemic and increasing use of video conferencing are additional forces that are favoring the Cloud computing environment. However, to avoid unnecessary vendor lock-in it is important to ensure that the software is designed and constructed in compliance with industry standards, with little to no dependency on proprietary cloud infrastructure so that application components can be readily migrated across Clouds (commonly referred to as a multi-cloud strategy).

Cloud-Provider Agnostic
bottom of page