CRAFT NEWS Feb18 📣| CES, Legrand, Hostabee, AI leaders, Cybersecurity, arXiv, new team member...
AI is all the rage @CES, again!
- A Legrand smart building solution powered by craft ai…
Legrand recently launched its new interoperability program “Works with Legrand” and put AI at its core to improve overall user experience.
We’re delighted to keep working with legrand to “ […] make it possible to imagine a self-learning building, which learns about users’ habits by gathering data from the various connected products located inside.”
- Hostabee make the Buzzzzzz at CES!
Hostabee presented its connected device which can be installed in any type of bee hives to monitor their behaviors and report back to the beekeepers. Powered by craft ai AI API, it enables beekeepers to remotely check the status of the hive and anticipate when something could go wrong.
AI Night takes over CES for AI
On 8 February, we were invited to the first AI Night in Paris as one of the founding members of the Microsoft AI Factory to pitch to French corporate & government leaders. The Microsoft AI Factory was well represented alongside other top AI startups like Snips & Heuritech. Great event, great talks #FranceIsAI!
craft ai selected by SystemX, Airbus & Thales to join the Start@SystemX program for Cybersecurity
During 6 months, we will work in collaboration with SystemX, Airbus & Thales and apply our explainable AI to a major Cybersecurity issue: network security. The goal is to prototype intrusions detection and processing semi automation through Network Behavior Analysis, based on craft ai white box predictive models.
Improving prediction with more balanced decision tree splits
Our first publication on arXiv about “Information gain ratio correction: Improving prediction with more balanced decision tree splits” is available HERE!
Abstract: Decision trees algorithms use a gain function to select the best split during the tree’s induction. This function is crucial to obtain trees with high predictive accuracy. Some gain functions can suffer from a bias when it compares splits of different arities. Quinlan proposed a gain ratio in C4.5’s information gain function to fix this bias. In this paper, we present an updated version of the gain ratio that performs better as it tries to fix the gain ratio’s bias for unbalanced trees and some splits with low predictive interest.
Welcome to our new team member!
Joining our craft ai team as a mad scientist, Tom will help us improve even more the craft ai core engine. After stopping by angus.ai and completing his final UTC (Technology University of Compiègne, France) semester in Singapore on deep learning and NLP, we’re happy to welcome him!