Approach

.01 Why are we unique?

Our Artificial Intelligence technology revolutionizes your processes and its implementation takes less than 8 weeks.
We propose a unique technology based on 3 pillars:

DIGITAL TWIN

The digital twin is the dynamic and digital model of your physical network, visible in our interface. The data feeding the digital double come from the sensors installed on your network.

AI MODULES

We set up our own fine tuned models, based on Machine / Deep / Reinforcement Learning technologies.

HYBRIDIZATION

Datasets are hardly ever perfect. We mitigate this by integrating theoretical models that our tool will aggregate with external databases to build a complete and reliable dataset.

DIGITAL TWIN

The digital twin is the dynamic and digital model of your physical network. The data feeding the digital double come from the sensors installed on your network.

AI MODULES

We set up our own fine tuned models, based on Machine / Deep / Reinforcement Learning technologies.

HYBRIDIZATION

Datasets are hardly ever perfect. We mitigate this by integrating theoretical models that we aggregate with external databases to build a complete and reliable dataset.

The complexity of a network consists in the links it's made of. An event impacting a point in the network will propagate upstream and downstream. Not integrating this complexity natively would not show this type of phenomenon. That's why our technology relies on databases in the form of relational graphs. These very specific databases allow us to create digital twins and to natively integrate the complexity of a network.

A

I

Our tool also integrates different modules of Artificial Intelligence that can predict events, identify anomalies or propose optimization possibilities. Finally, to overcome the lack of certain data, and to reinforce the robustness of data-driven models, we have the possibility to integrate theoretical models in our algorithm. This is called hybridization of data models by theoretical models.

  • Auto Machine Learning

Auto Machine Learning

This module allows you to create very easily thousands of models for a single curve.
  • Deep Learning
  • Deep Learning

Deep Learning

Ability to understand and identify a complex phenomenon.
  • Reinforcement Learning

Reinforcement Learning

Ability to propose action based on business rules.

DCbrain continues to innovate: in addition to our own R&D team, we are working with major academic partners: Telecom ParisTech and Laboratoire David among others.

.02 User onboarding.

To ensure the success of a “data driven” project, we build our project on 2 key elements:

  • A thorough data audit to quickly identify which datasets can be used in the short / medium / long term.
  • The user onboarding from the beginning of the project, to ensure a good grip on the tool.

An audit of your data is essential to evaluate the value of these data. We start all our projects with a thorough audit of your raw datasets in order to standardize the data and make it exploitable by our technology. Datasets often contain injection data, network descriptions, consumption data etc.

At DCbrain, we facilitate the integration phase by using the Agile method, while focusing on integrating our users at all stages of our implementation process. This allows us to adapt to changing needs and offer solutions in a common effort. We rely on a continuous adjustment of the functional needs identified over the entire duration of the project. Our ambition is to shorten and facilitate the deployment phase.