How the deployment of an explainable AI solution improves energy performance management at Dalkia

13/09/2018

Applications

Tous les articles

Sommaire

Due to rising energy cost, scarcity of resources and the need to limit environmental impacts, Dalkia has decided to put the optimization of energy performance at the heart of its strategy. To meet this challenge, Dalkia created the Desc (Dalkia Energy Savings Center) back in 2013, a control center for energy performance management of its customers’ sites to ensure better comfort and lower energy consumption.

The Desc mission is to collect and manage in real time the data transmitted by a digital network of more than 600 000 meters and sensors. This data is then analyzed to improve energy performance of more than 30,000 buildings. Fully committed to helping its customers consume less and better, Dalkia has deployed Artificial Intelligence in order to improve the Desc efficiency and help energy managers in their daily work.

The Desc, in the AI age

Dalkia is now collecting more than 29 million data points per day and this number is exploding! To Handle this growth, energy managers need solutions to support them in their data analysis and performance management processes. Anticipating this challenge, Dalkia started partnering with craft ai in 2016 to combine human intelligence with cognitive automation. Dalkia has implemented an Artificial Intelligence as a digital assistant for energy managers that is fully integrated within their usual workflow. It provides recommendations to semi-automate energy performance supervision process, enabling efficiency gain without specific training. As energy managers provide feedbacks through their usual dashboards and new data is pushed, the assistant keeps learning to automatically stay up to date and provide improved recommendations.

Far from the "black box" conundrum

Thanks to craft ai’s explainable AI approach, the assistant is definitely not a mysterious black box for the energy managers. They are always provided with the reasons behind each recommendation as well as the assistant’s corresponding level of confidence. Using their usual interface, energy managers can easily indicate whether a recommendation provided by the assistant is relevant or not, and send corrective elements that will be processed as new data to ultimately generate improved recommendations. The assistant’s explainable AI played a key role in its deployment at scale as it fosters trust and control.

A trustworthy assistant...

The assistant has been designed to be transparent. It works in the background, implementing the expertise it has automatically learned from energy managers, enabling them to focus on higher added value tasks. The energy managers also have access to the reasons of the recommendations provided by the assistant, allowing them to maintain control over the final decision. Today, the assistant detects energy performance anomalies for all sites managed by Dalkia and maintains a TOP 10 of the sites to be treated as a priority. As more new data is integrated, the assistant will be able to cover the full spectrum of the supervision process, from data analysis to energy performance actions suggestions.

...that learns on his own to become even better!

The craft ai underlying automated machine learning technology works at the individual level. This makes it possible to generate a predictive model for every site and automatically update them as new data is pushed into craft ai cloud. This architecture makes the Desc assistant a self-learning application whose recommendations accuracy keeps improving. To learn more about the Dalkia energy managers digital assistant and how AI contributes to improved energy performance, come meet craft ai & Dalkia at the FEDENE congress in September the 20th! Learn more

If you have any question, get in touch at contact@craft.ai.

À télécharger

Vous pourriez également apprécier

IA de confiance
14/09/2022

Garder l'humain dans la boucle avec l'XAI

Quelle place l’explicabilité (XAI) occupe-t-elle aujourd’hui en Machine Learning et en Data Science ? Le challenge de ces dix dernières années en data science, a plutôt été de trouver la bonne “recette algorithmique” pour créer des modèles de ML toujours plus puissants, plus complexes et donc de moins en moins compréhensibles.

Lire l'article

R&D
12/07/2022

L'industrialisation de l'IA & le concept de MLOps

Le MLOps apparaît comme une nécessité pour pallier les difficultés lors du passage à l’échelle de l’IA au sein des entreprises : la reproductibilité, le versionning, l'intégration continue... C’était l’objet de l’une des conférence sur l'industrialisation de l'intelligence artificielle dans le cadre de l'Enjeu Day Industrie & Services 2022. Vous n’aviez pas pu y assister ? Retrouvez le replay.

Lire l'article

IA de confiance
10/05/2022

A guide of the most promising XAI libraries

Using Machine Learning to solve a problem is good, understanding how it does is better. Indeed, many AI-based systems are characterized by their obscure nature. When seeking an explanation, looking to understand how a given result was produced, exposing the system’s architecture or its parameters alone is rarely enough. Explaining that a CNN recognized a dog by detailing the Neural Network weights is, to say the least, obscure. Even for models deemed as glass-box such as decision trees, a proper explanation is never obvious.

Lire l'article