Snowflake has laid out plans to assist in democratizing accessibility to machine learning (ML) sources by removing any obstacles for users who are not experts.
In its annual user conference, Snowflake Summit, the database company unveiled a series of announcements that aim to ease the adoption in machine learning.
Most notable among them are improved assistance of Python (the language that the majority of ML software is written) and a brand new marketplace for apps that allows partners to make money off their models.
“Our aim is making it as simple as is possible for customers to use the latest ML models without the need to construct from scratch because it takes a lot of experience,” said Tal Shaked who is the head of ML for Snowflake.
“Through initiatives like Snowflake Marketplace, we want to provide customers with a means to run these models on their data both in a safe manner.”
Access for all
While the concept of machine learning has been around for a long-standing concept, only in the past few years have advancements in storage, compute software, and other technologies made it possible for widespread use.
Yet the bulk of knowledge and innovation is shared heavily in the hands of a small number of businesses, including Google or Meta.
The goal of Snowflake is to give access to opportunities on the cutting-edge of machine learning by implementing an ecosystem-driven partnership strategy.
Shaked the researcher, who was working on several machine learning initiatives at Google prior to joining Snowflake and Snowflake, said that customers get access to fundamental resources, and on top of which they will be able to make minor adjustments to their particular use scenarios.
For instance, a highly developed neural technology for processing language (NLP) model created by organizations like OpenAI could serve as the basis for a general-purpose fast food eater seeking to create an ordering system powered by ML the expert suggested.
In this case the customer has no part in the process of training or adjustments to the model but is still able to reap all the benefits from the technology.
“There’s a lot of innovation taking place in the field of ML and we’d like to incorporate this into Snowflake by way of connections,” said the founder of Snowflake TechRadar Pro.
“It’s about figuring out how we can work with these services so that our customers can take care of the fine-tuning without having to pay a whole bunch or PhD experts.”
This was also echoed earlier during this day’s proceedings by Benoit Dageville, co-founder of Snowflake who talked about its importance to exchange knowledge throughout the partner and customer ecosystem.
“Democratizing ML is an important element of what we’re doing. We’re evolving into an ML platform, however, not only the place where you build it and use it to serve yourself. The real revolution is in sharing expertise.”
“It’s not just the Meta’s and Google’s of this world who use this technology, as we’re making it simple to distribute.”