Edge-ML benefits explained to...
Decision Makers, Business and Marketing Managers
A generic tool, useful for all your projects
Edge-ML is a generic tool which makes all your Machine Learning projects more efficient: banking scoring, marketing targeting, text / e-mail classification, online ad targeting, sentiment analysis, etc. (see Use Cases)
Two possible uses of Edge-ML ...
- Data Scientists use Edge-ML with regular Machine Learning techniques in order to speed up and make secure their projects (eg: very robust base-line model which can be used for the "Go/No go" of the projects, detection of useless variables, calibrating any models, recoding data and detecting drifts between train and deploy sets, automated feature engineering ...).
- In automatic mode, Edge-ML carries out Machine Learning projects from A to Z, from the data preparation to the training of a classifier, and its release into the production environ. The models are produced in record time and reach an excellent trade off between Performance and Robustness.
Cost / Benefit ratio
The cost of a Machine Learning project with Edge-ML ranges from a hundred euros for a small project to a few thousand euros for a big project. But what benefits you draw?
- Go / No go projects: At the beginning of a Machine Learning project, Edge ML quickly provides a baseline model which allows you to evaluate the input dataset: is it informative enough to meet the needs of the project? Otherwise, it is wiser not to start the project because the results may be insufficient.
- Reduction in the project's duration: A 2-week project carried out by using regular Machine Learning techniques will only takes a few hours with Edge-ML. Whether in automatic mode or used with regular Machine Learning techniques, Edge-ML hugely increases the productivity of your teams.
- Reduction of used hardware resources: Edge ML is very hardware-efficient because it is based on a disruptive mathematical approach. In practice, tens of millions of rows can be processed by using a standard server. Edge-ML saves you heavy investment in computing infrastructure (Hadoop clusters, HPC ...).
- Robust and almost optimal models: Edge ML automatically provides high-performance models (which are rarely mistaken) and very reliable (stable performance when the model is applied on new data). It is always possible to obtain more accurate models than those of Edge-ML (which is a generic tool) by looking for the most suitable Machine Learning algorithm for each project. But this maximum accuracy is very costly to obtain regarding human efforts and required hardware resources. Moreover, the return on investment of such a model is uncertain due to the lack of robustness of regular approaches.
- Interpretable models: Since Edge-ML models are easily interpretable, they can be efficiently exploited for marketing purposes or to improve the understanding of business phenomena. Interpretability is also a regulatory requirement when the models are applied on personal data.