Language models are usually designed by tech giants like Open AI’s GPT-3 or Deepmind’s Gopher. Recently we devoted an article to BLOOM, the largest open-source multilingual language model trained to date involving a thousand researchers. LightOn, a French start-up that has developed a language model for European companies, has just announced that its LightOn Muse application is now available in French.
Created in 2016 by Igor Carron, Laurent Daudet, Florent Krzakala and Sylvain GiganCo, LightOn has 20 employees among the best European engineers and researchers in ML. After launching its very first photonic co-processor in 2020, the start-up tackled a new challenge in the emerging field of AI: creating the first major language models for European languages, in particular French. .
The genesis of the VLM-4 project and the Muse API
Initially, training large language models was one of the possible applications for LightOn’s photonic processor. Very quickly, the team became passionate about this new generation of AI, still unknown in France, and developed and trained its own models. In 2020, it provides free access to its first French model through a simple text generation interface: PAGnol. In 2022, after a year and a half of intensive work, it has developed VLM-4, a suite of large language models in 5 European languages: English, German, Spanish, French and Italian.
Language models and European sovereignty
Most of the language models, due to their complexity and costs, are the prerogative of large companies like GAFAM, and are only available in English, Chinese and Korean. The use of these large models for other languages can only be done by going through a translation tool, which implies a drop in quality and an increase in costs, models like BLOOM or LightOn Muse are a game changer.
The essential question, raised by the relatively low development of these technologies in Europe, is that of sovereignty: these technologies are essential to the digital transformation of companies, which gives them a decisive competitive advantage over their competitors. European companies find themselves powerless in the face of the free recovery and use of their data, which will be used to improve non-European products and services.
LightOn is the first company to train large language models directly in four European languages other than English.
Democratize access to major VLM-4 language models in 5 languages and to customization features (skills)
The Muse API is aimed at all European players, whatever their size or sector of activity (marketing, media, leisure industry, technology companies, even administrations), who need to address their audience in their own language. It gives them access to the major VLM-4 language models in five languages (French, English, German, Spanish, Italian), as well as customization features (skills) allowing them to “specialize” the model for specific tasks.
The goal is to allow all these actors to easily build:
- Commercial products and services around text-related tasks: a retailer will thus be able to be helped by AI to create daily content for the Instagram account of his shop, a student better classify and synthesize the countless articles for his dissertation work… Content editors, for their part, will be able to tap on the model’s suggestions to write faster;
- Products and services for their internal use: managers will be able to get help from an AI assistant in project management, emails or internal documents will be automatically summarized so that employees only have to read the essentials, by detecting those who call for an immediate response…
A multilingual API
Muse opens large-scale models to a wide range of languages: French, English, Italian, Spanish, German, many others are planned (40 languages by the end of 2023).
Ease of use
The Muse API is designed to be flexible, easy to integrate into any system, and usable by anyone, anywhere. You just need to give instructions and examples in natural language, as if you were interacting with a human.
Save time and increase efficiency
Time-consuming tasks, which are essential for the operation of a company, such as responding to e-mails and customer reviews, writing posts to go up in online searches can be taken care of by the Muse API.
Ultra-powerful language models
The Muse API uses VLM-4, some of the most powerful large language models on the market. LightOn engineers are constantly innovating to increase the size of their models and the quality of the data with which they are trained (two essential parameters when it comes to text generation). VLM-4s thus have the ability to respond in context, to learn to perform a task from only a few examples (few-shot models) or even without any examples (zero-shot models).
Many customization possibilities
By making the models very efficient to perform specific tasks (skills), LightOn engineers can adapt them to the needs and particularities of each company.
Regarding e-marketing, for example, the Muse API provides:
- A high-performance SEO to automatically generate text around popular keywords, to improve the visibility of the company;
- Emailing campaigns and advertisements that hit the bull’s eye since they target each customer in particular;
- The creation of much higher quality content, with much less effort.
Improved customer experience
Customer satisfaction increases as they get real-time support through more efficient chatbots. At the same time, the implementation of personalized search engines facilitates the analysis of data of all types.
Analysis of customer feedback
The “Sentiment Analysis” feature delivers a reliable summary of customer feedback to simplify decision-making. The various opinions and evaluations are analyzed and classified. The customer database is structured, which allows for more efficient management.
Muse can summarize documents and emails to extract essential information to save time. In addition, by creating a custom search or classification tool, it is possible to efficiently process large amounts of data of all types.