Artificial intelligence promises to help, and perhaps even replace, humans to perform everyday tasks and solve problems that humans have been unable to solve, but ironically AI is facing to a major scaling issue. It’s only as good as the models and data used to train it, so ever-larger troves of data need to be searched and ingested. But annotating and manipulating this training data takes a lot of time and money, slowing down work or overall efficiency, and maybe both.
A startup called V7 Labs thinks it has made a breakthrough in how this is approached. These are training models that are efficiently built to automate the training of these models. Today, he is announcing $33 million in funding to fuel his growth after seeing strong demand for his services.
Today, V7 focuses on computer vision and helping to identify objects. He says he can learn what to do from just 100 human-annotated examples.
It currently has strong traction in the fields of medicine and science, where its platform is used to help train AI models to speed up, for example, how cancers and other problems are identified on scanners. V7 is also starting to see activity with tech and tech-savvy companies looking to apply its technology in a wide variety of other applications, including companies building engines to create images from natural language commands and images. industrial applications. It doesn’t disclose a full list of customers and those reviewing its technology, but the list has more than 300 customers and includes GE Healthcare, Paige AI and Siemens, along with other Fortune 500 companies and larger private companies.
Radical Ventures and Temasek are co-leading this round, with Air Street Capital, Amadeus Capital Partners and Partech (three previous backers) also participating, along with a number of leading people in the world of machine learning and of AI. Among them, François Chollet (the creator of Keras, the Python open source neural network library), Oriol Vinyals (principal researcher at DeepMind), Jose Valim (creator of the Elixir programming language), Ashish Vaswani (co-founder of Adept AI who had previously been at Google Brain, where he invented Transformers) and other Anonymous from OpenAI, Twitter and Amazon.
CEO Alberto Rizzoli said in an interview that this is the biggest Series A funding round in this category to date, and will be used both to hire more engineers and to grow its businesses. business operations to meet a new wave of customer interest. with a focus on the United States. He declined to comment on the valuation, but the startup has now raised around $36 million, and from what I understand the valuation is now around $200 million.
Rizzoli also declined to talk about revenue figures, but said ARR had tripled in 2022.
A number of other startups have sprung up to help improve the efficiency of AI data training and to address the broader field of AI modeling. SuperAnnotate, which raised around $18 million per PitchBook, is one of V7’s closest competitors. (V7 even explains how the two services compare.) Others include Scale AI, which initially focused on the automotive sector but has since branched out into a number of other areas and is now valued at around $7 billion; Labelbox, which works with companies like Google and others on AI labeling; and Hive, which is now valued at around $2 billion.
V7 – named in reference to AI being the “seventh” image processing area after the six areas of the human brain that form its visual cortex (V1 to V6) – and the others rely on the fact that the training model is ineffective and can be improved.
The specific USP of V7 is automation. He estimates that about 80% of an engineering team’s time is spent managing this training data: labeling, identifying when something is mislabeled, rethinking categorizations, and so on. So he built a model to automate this process.
He calls the process he pioneered “programmatic labeling”: using general-purpose AI and its own algorithms to segment and label images, Rizzoli (who co-founded the company with its CTO Simon Edwardsson ) says it only takes 100 “human-guided” sessions for its automated labeling to spring into action.
Investors are betting that shortening the time between design and application of AI models will boost business activity. “Computer vision is being deployed at scale across industries, bringing innovation and breakthroughs, and a rapidly growing $50 billion market. Our thesis for V7 is that the breadth of applications and the speed at which new products are expected to come to market require a centralized platform that connects AI models, code, and humans in a looping ecosystem. “said Pierre Socha. , partner of Amadeus Capital Partners, in a press release.
V7 describes the process as “autopilot”, but the co-pilot could be more specific: the idea is that anything flagged as unclear is sent back to humans for evaluation and review. It doesn’t so much replace those humans as it allows them to manage workloads more efficiently. (It can also work better than humans at times, so both used in tandem could be useful for checking each other’s work.) Below is an example of how image training works on a scan to detect pneumonia.
Given the many areas where AI is being applied to improve the way images are processed and used, Rizzoli said the decision to double down on medicine was initially partly to keep its feet on the ground. the startup and to focus on a market that could I’ve never built this kind of technology internally, but I would definitely like to use it.
“We decided to focus on verticals that already bring AI-based applications to market, or where a lot of visual processing work is done, but by humans,” he said. “We didn’t want to be tied to moonshots or projects that lack big R&D budgets, because that means someone is looking to solve the problem entirely themselves, and they’re doing something more specialized, and that he may want to have his own technology, not that of a third party like us.
And in addition to companies searching for “their own secret sauce,” sometimes projects never see the light of day outside of the lab, Rizzoli added. “We work more for real applications,” he said.
In another way, the startup represents a shift we’re seeing in the way information is obtained and embraced by businesses. Investors believe that the framework built by V7 could potentially change the way data is ingested by these companies in the future.
“V7 is well positioned to become the industry standard for data management in modern AI workflows,” Radical Ventures partner Parasvil Patel said in a statement. Patel joins the V7 board with this round.
“The number of problems that can now be solved with AI is vast and growing rapidly. As companies of all sizes race to seize these opportunities, they need the best data and model frameworks to deliver exceptional products. that are constantly improving and adapting to real-world needs,” added Nathan Benaich of Air Street Capital, in a statement. “This is where V7’s AI data engine shines. Regardless of industry or application, customers rely on V7 to ship robust AI-driven products faster than ever. V7 incorporates rapidly evolving industry best practices in multiplayer workflows from data to model to product.”
#snags #million #automate #training #data #computer #vision #models