Technology

Deduction

This is the most natural form of reasoning for humans. Deduction is the path that leads from observations to conclusions. ExpressIF^® allows its reasoning to be applied to static data (e.g., tabular data) or streaming data (e.g., sensor data). This way, as soon as the situation changes, it can automatically change its decision in no time.
Depending on the quality of the input data, ExpressIF^® can infer more or less certain conclusions in an extremely simple way.

Constraint satisfaction

This is a combinatorial reasoning that consists of finding a value for each variable described in the problem, such that the constraints that govern them are satisfied. ExpressIF^® allows for the finding of potentially degraded solutions, by allowing constraints to be partially satisfied. This makes it possible to represent, for example, regulatory constraints (which must be satisfied) and preferential constraints (which are preferred to be satisfied, but not required).

Planning

In complex collaborative situations, this engine helps find the optimal sequence of actions for each agent to achieve a situation or goal. Planning uses the description of the agents and their environment while exploiting vague knowledge about the effects of actions and the agents' capabilities. Combined with the constraint solver, ExpressIF^® allows finding an efficient compromise between planning and scheduling.

Temporal domain

ExpressIF^® allows situations to be characterized based on a time horizon defined with varying degrees of certainty. This allows, for example, the expression that one event occurred just before another. The time domain is particularly useful for system monitoring, diagnostics, and predictive maintenance.

Spatial domain

Spatial knowledge is also defined vaguely. ExpressIF^® allows you to express relationships between localized objects that can themselves be defined vaguely, such as "being close to," "being to the right of," or "being symmetrical to." This is particularly useful for crisis management, image or video analysis, or security applications. Learn more about spatial reasoning in ExpressIF^®

Spatio-temporal domain

By combining temporal and spatial aspects, ExpressIF^® allows the characterization of the movement of geolocated and mobile entities. ExpressIF^® provides spatio-temporal relationships, such as "a vehicle parks too long", "a person follows a path", "a machine crosses the construction site". This is used, for example, for scene analysis, particularly for security or for home care for elderly people.

Other domains

ExpressIF^® goes further by allowing you to express other types of relationships in a very simple way. In an identity management context, we have created comparators based on words, allowing you to express that "one name is pronounced almost like another", or that "one word is spelled almost the same as another".

Expertise

If you don't have any data yet, ExpressIF^® allows you to rely solely on your expertise and know-how. You have all its expressiveness to describe the behavior it should have. This allows you to establish, for example, a recommendation based solely on expertise while waiting to accumulate historical data and produce a model from it. If your case is complex, we assist you in building the knowledge base by interviewing your experts.

Structured data

From structured data, ExpressIF^® can generate a new knowledge base or optimize an existing one for a given purpose. The transparency of ExpressIF^®'s models allows you to understand what was "seen" during training and to correct or complete it if it doesn't suit you.

Textual documents

Does your company have knowledge in text form? CEA List has the expertise to implement a custom tool to extract the knowledge base from a text corpus.

ExpressIF^® was initially developed for security and crisis management projects. Particularly in these fields, trust in AI is paramount. ExpressIF^® offers state-of-the-art tools and those derived from our own research to understand the knowledge base, whether it was created by experts or machine learning. The concept of explanation is also an integral part of the ExpressIF^® architecture and provides support for understanding the decision made by the AI.

Interoperability

We are constantly working to improve the integrability of ExpressIF^® into your information system. We offer ExpressIF^® WebApp to manage access to technology through a web service. WebApp allows you to manage your engines, access rights, and knowledge bases. You can also directly integrate our libraries. In addition, ExpressIF^® allows you to import and export your existing knowledge bases using standard formats (PDDL for planning, IEEE for fuzzy inference systems, etc.).

Multiplatform and modular

ExpressIF^® is built on the .NET platform and can therefore run on many operating systems (Windows, Linux, macOS, iOS, Android, Raspbian, etc.). Our modular, plug-in approach allows you to scale ExpressIF^®'s footprint depending on the components you use, making it particularly suitable for AI-embedded scenarios.