Table data is a wide range of terms that generally contain a structured data that conforms to a specific line and a column. SQL database, spreadsheet, .csv file, etc.
Although artificial intelligence applied to untracted data has been greatly advanced, these large language models are ambiguous in design. They are constructed to generate coherent output by manipulating input tokens without always following the fixed structure. Also, the best LLM is expensive to access through the API or to execute it in a cloud infrastructure.
Nevertheless, many companies are already in order to concentrate all important data and all important data and some data scientists that can use this data to improve the company's strategy. I already have a data strategy.
French startup neuralk-AI is an artificial intelligence company and is working on AI models focusing on table data. The company announced $ 4 million funding this week.
“The data with the true value of a company is a long -specific data, consisting of a table, and uses data scientists in these companies to create all machine learning algorithms.” Supreme scientist Alexander Pasquio told TechCrunch.
Neuralk-AI believes that there is a chance to reconsider the AI model development, but is particularly focused on structured data. At first, since these companies are considering data catalogs, customer databases, shopping cart trends, etc., we plan to provide a model as an API for data scientists working for commercial companies.
“Today, LLM is the best way to answer questions based on searching, natural users, and non -structured documentation, but there are some restrictions when you return to classical machine learning. This is actually based on the data on the classic table, “said Pasquiou.
Neural-AI can automate complex data workflows by eliminating and enrichments by wise duplication and concentration. However, you can use the company's model to detect fraud, optimize product recommendations, and generate sales predictions that can be used for inventory management and product prices.
Fly Ventures also participated in Steamai and led the company's $ 4 million rounds. Several business angels have also invested in the startups, including Hug Face's Thomas Wolf, Alan's Charles Golintin, Miracle's Philip Colott, and Nagi Lethaifa.
The team is still actively working on the model. We plan to test with major French retailers and emerging companies in commercial transactions, such as E.LECLERC, Auchan, Mirakl, Lucky Cart.
“We will release a public benchmark that can rank the model in three or four months with the first version of the model and the state -of -the -art model in this field,” said Pasquiou. “And in September, this idea will be the best table model for everything related to expression learning.”