Artificial Intelligence
Artificial Intelligence to meet your challenges and go further
Big data, cloud and digital transformation, the foundations of AI
Improving processes and making them more reliable, more efficient, better adapted and more personalised is at the heart of our approach, so that we can offer you ever more relevant solutions to your challenges.
Digital transformation, Big Data, open data, more powerful servers and the cloud are all steps that have facilitated the deployment of platforms and solutions, each more innovative than the last, without the need for substantial investment.
These developments have spawned a meteoric rise in new developments, with Artificial Intelligence and Data Sciences being the most emblematic of the last 10 years. Like all new technologies, they are having an impact on our businesses and will open up new opportunities as yet unknown.
Coexya, a strong expertise in data processing
And what about Coexya? With over 20 years of experience, Coexya is one of the recognized specialists in data processing, structured or not (text, 2D and 3D images, sound, video), from capture to its valorization through the design and implementation of solutions adapted to different use cases.
Artificial Intelligence, at the centre of our business solutions
From 2000 to 2012, Coexya carried out a large number of projects based on symbolic AI.
Significant achievements:
- Industrial property: measurement of similarity between images and automatic classification,
- Healthcare: automatic interpretation of medical reports for the management of emergency patients (LERUDI)
- Healthcare: automatic interpretation of medical reports for the constitution of oncology cohorts (CONSORE)
- Legislative: assistance in the consolidation of French law by analyzing the JO and searching for modifying actions,
- Legislative: anonymization of jurisprudence texts
Symbolic AI: how machines imitate human thought
Symbolic AI: Component that allows reasoning based on formal rules (expert system)
Our view: Easily explained because it is based on formal rules. However, it requires upstream work of coding by domain experts
Since 2012, connectionist AIs, driven by unexpected and impressive results, have revolutionized image analysis, altering the overall approach, and prompting our methodologies to adapt.
Our achievements include:
- Investigation: Multilingual multimedia data analysis (Text, Sound, Image, Video), Facial recognition, Translation
- Healthcare: Overhaul of the Consore platform
- Road infrastructure: Bot for defect detection on roads from images
- Industrial Property: Redesign of similarity search and classification (Learning based on enriched results from Symbolic AI)
- Digital Twin: 3D reconstruction of metro stations and automatic recognition of equipment
- Nuclear: Instant translation for nuclear waste reprocessing training
- Access control: Facial recognition and validation of identity documents
Connectionist AI: how machines learn like the human brain
Connectionist AI : combining machine learning and deep learning approaches. Rules are no longer coded, they are learned.
Our view: We see no limit to the highly versatile approach of these architectures and the ability to specialise existing AI.
The downside is that nothing is magic yet, and the majority of implementations are based on supervised learning and require a set of data annotated by experts that is not always available or is poorly annotated or even biased.
In 2022, generative AIs exploded into the public domain through OpenAI and its chatGPT platform built on LLM (Large Language Model). These LLMs were developed thanks to the famous Transformers defined in 2017 in a groundbreaking article by a Google team, “Attention is all you need.”
Coexya uses, prototypes and monitors Generative AI.
- Coexya Software Factory: Integration of Github Copilot and Starcoder AIs into development environments for pair programming, analysis, bug searching, analysis assistance, optimization, etc.
- Search: Prototyping of response generation based on private document corpus (fine-tuned LLM)
- Analytics: Prototype for assisting in the resolution of investigation cases
- Analytics: Prototype for aiding in the analysis of numerical results
Generative AI: A component that generates text, images, sound, videos, presentations based on prompts.
Our view: After understanding content through AI, the generative dimension naturally meets expectations. It’s an exocortex that needs to be tamed to become a relevant and versatile assistant in various generation and analysis tasks, with varying degrees of creativity. It’s essential to approach it with good targeting and usage guidance, or else it might be challenging to manage!
