Artificial Intelligence

Your challenges 

The AI market is booming and many organizations see it as a driver of innovation. However, many organizations are still struggling to extract value from their projects and many questions remain: Where is the state of the art in AI? How to fully exploit the potential of AI? How do we move from PoC status to deployment at scale? Which AI sub-domains should be explored first to gain performance? 

Our value proposition 

As a pioneer in AI, Novelis helps you create value and generate ROI with Artificial Intelligence: 

• Ideation and identification of use cases  
• Data mining and PoC (Proof of concept) experimentation  
• Development, design and integration  
• Monitoring and reporting on the state of the art by our laboratory researchers   
• Training, acculturation and change management 

Our AI experts and doctors are thus able to intervene in many fields of Artificial Intelligence: 

Machine Learning to enable systems to improve their capabilities through experience. This technique is notably used in one of the software solutions, Novy POM, for the intelligent processing of invoices and purchase orders.

• Predictive analytics or data science: by combining data mining, machine learning and predictive modelling, AI makes it possible to predict future events (such as sales volumes) based on a history and by integrating both internal and external variables.  
• Deep Learning (machine learning): this learning technique is particularly effective in the context of translation or voice or image recognition. 

NLP (Natural Language Processing) or TAL (Traitement Automatique du Langage) in French to give machines the ability to understand and process human language. This is what allows conversational agents to function today.  AI, and especially NLP (Natural Language Processing) or NLU (Natural Language Understanding), allows to create conversations between man and machine. More precisely, NLU allows to analyze a sentence and to extract its meaning (the intention) and eventually its variables (the entities).   

• Automatic translation (from a language A to a language B): since the creation of our R&D Laboratory, our doctors have been working on the problem of generating a programming language from a need described in natural language. The work of our researchers on this subject has been praised by the scientific community on several occasions. For more information
• Classification: consists in predicting the class of a data from the characteristics of the data. The objective is to identify the letters of the alphabet or a number from a database of labeled characters. A concrete application of this tool is the automatic reading of mail addresses at the post office, allowing the automatic redirection of mail used in certain areas 
• Information/extraction: i.e. the extraction of information or data from an element (such as an invoice or the processing of information contained in emails).   

Speech (or speech processing) to convert audio content into text (Speech-to-Text) or to read texts aloud via a machine (Text-to-Speech). This is the technique used by Voicebot, robots capable of conversing with a person by voice in natural language. Unlike Chatbots, which are able to communicate in writing

Computer Vision and Machine Vision to give machines the ability to “see” and “understand” images or videos. This technology is used in particular for optical character recognition (OCR), which is very effective for automating financial processes (invoice reading).