AI

V7 snaps up $33M to automate training data for computer vision AI models

Comment

cell scan
Image Credits: V7 Labs (opens in a new window)

Artificial intelligence promises to help humans carry out everyday tasks faster and quickly solve problems that they have been too big for humans to tackle. But ironically, the building that AI can take a long time because of the data crunching needed to train the models.

That’s given rise to a wave of startups aiming to speed up that process.

In the latest development, V7 Labs, which has built tech to automate notations and other categorizing of data needed for AI training models, has raised $33 million in funding after seeing strong demand for its services.

V7’s focus today is on computer vision and automatically identifying and categorizing objects and other data to speed up how AI models are trained. V7 says it needs just 100 human-annotated examples to learn what it needs to do.

It currently has strong traction in the fields of medicine and science, where its platform is being used to help train AI models to identify, for example, how cancers and other issues are identified on scans. V7 is also starting to see activity with tech and tech-forward companies looking at how to apply AI in a wide variety of other applications, including companies building engines to create images out of natural language commands and industrial applications. It’s not disclosing a full list of customers and those evaluating its tech but the list numbers more than 300 clients and includes GE Healthcare, Paige AI and Siemens, alongside other Fortune 500 companies and larger privately held businesses.

Radical Ventures and Temasek are co-leading this round, w1ith Air Street Capital, Amadeus Capital Partners and Partech (three previous backers) also participating, along with a number of individuals prominent in the world of machine learning and AI.

They include Francois Chollet (the creator of Keras, the open source Python neural network library), Oriol Vinyals (a principal research scientist at DeepMind), Jose Valim (creator of the Elixir programming language), Ashish Vaswani (a co-founder of Adept AI who had previously been at Google Brain, where he invented Transformers) and unnamed others from OpenAI, Twitter and Amazon.

CEO Alberto Rizzoli said in an interview that this is the largest Series A funding round in this category to date, and it will be used both to hire more engineers as well as to build out its business operations to take on a new wave of customer interest with an emphasis on the U.S.

He declined to comment on 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 that ARR grew three-fold in 2022.

There have been a number of other startups that have emerged to help improve the efficiency of training AI data and to address the wider area of AI modeling.

SuperAnnotate, which has raised about $18 million per PitchBook, is one of V7’s closer rivals. (One example of that: V7 lays out how the two services compare on its site, and SuperAnnotate has been in touch to explain how the comparison is not accurate.)

Others include Scale AI, which initially focused on the automotive sector but has since branched 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.

As with these companies, V7 — named in reference to AI being the “seventh” area for processing images after the six areas in the human brain that form its visual cortex (V1 through V6) — is building services to solve a specific challenge: the concept of the training model and how data is fed into it is inefficient and can be improved.

V7’s specific USP is automation. It estimates that around 80% of an engineering team’s time is spent on managing training data: labeling, identifying when something is incorrectly labeled, rethinking categorizations and so on, and so it has built a model to automate that process.

It calls the process it has come up with “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 that it takes just 100 “human-guided” examples for its automated labelling to kick into action.

Investors are betting that shortening the time between AI models being devised and applied will drive more business for the company.

“Computer vision is being deployed at scale across industries, delivering innovation and breakthroughs, and a fast growing $50 billion market. Our thesis for V7 is that the breadth of applications, and the speed at which new products are expected to be launched in the market, call for a centralised platform that connects AI models, code, and humans in a looped ecosystem,” said Pierre Socha, a partner at Amadeus Capital Partners, in a statement.

V7 describes the process as “autopilot” but co-pilot might be more accurate: The idea is that anything flagged as unclear is routed back to humans to evaluate and review. It doesn’t so much replace those humans as makes it easier for them to get through workloads more efficiently. (It can also work better than the humans at times, so the two used in tandem could be helpful to double check each other’s work.) Below is an example of how the image training is working on a scan to detect pneumonia.

Image Credits: V7 labs

Considering the many areas where AI is being applied to improve how images are processed and used, Rizzoli said the decision to double down on the field of medicine initially was partly to keep the startup’s feet on the ground, and to focus on a market that might not have ever built this kind of technology in-house, but would definitely want to use it.

“We decided to focus on verticals that are already commercializing AI-based applications, or where a lot of work on visual processing is being done, but by humans,” he said. “We didn’t want to be tied to moonshots or projects that are being run out of big R&D budgets because that means someone is looking to fully solve the problem themselves, and they are doing something more specialized, and they may want to have their own technology, not that of a third party like us.”

And in addition to companies’ search for “their own secret sauce,” some projects might never see the light of day outside of the lab, Rizzoli added. “We are instead working for actual applications,” he said.

Image Credits: V7 Labs (opens in a new window)

In another regard, the startup represents a shift we’re seeing in how information is being sourced and adopted among enterprises. Investors think that the framework that V7 is building speaks to how data will be ingested by enterprises in the future.

“V7 is well-positioned to become the industry-standard for managing data in modern AI workflows,” said Parasvil Patel, a partner with Radical Ventures, in a statement. Patel is joining V7’s board with this round.

“The number of problems that are now solvable with AI is vast and growing quickly. As businesses of all sizes race to capture these opportunities, they need best-in-class data and model infrastructure to deliver outstanding products that continuously improve and adapt to real-world needs,” added Nathan Benaich of Air Street Capital, in a statement. “This is where V7’s AI Data Engine shines. No matter the sector or application, customers rely on V7 to ship robust AI-first products faster than ever before. V7 packages the industry’s rapidly evolving best practices into multiplayer workflows from data to model to product.”

More TechCrunch

Looking Glass makes trippy-looking mixed-reality screens that make things look 3D without the need of special glasses. Today, it launches a pair of new displays, including a 16-inch mode that…

Looking Glass launches new 3D displays

Replacing Sutskever is Jakub Pachocki, OpenAI’s director of research.

Ilya Sutskever, OpenAI co-founder and longtime chief scientist, departs

Intuitive Machines made history when it became the first private company to land a spacecraft on the moon, so it makes sense to adapt that tech for Mars.

Intuitive Machines wants to help NASA return samples from Mars

As Google revamps itself for the AI era, offering AI overviews within its search results, the company is introducing a new way to filter for just text-based links. With the…

Google adds ‘Web’ search filter for showing old-school text links as AI rolls out

Blue Origin’s New Shepard rocket will take a crew to suborbital space for the first time in nearly two years later this month, the company announced on Tuesday.  The NS-25…

Blue Origin to resume crewed New Shepard launches on May 19

This will enable developers to use the on-device model to power their own AI features.

Google is building its Gemini Nano AI model into Chrome on the desktop

It ran 110 minutes, but Google managed to reference AI a whopping 121 times during Google I/O 2024 (by its own count). CEO Sundar Pichai referenced the figure to wrap…

Google mentioned ‘AI’ 120+ times during its I/O keynote

Firebase Genkit is an open source framework that enables developers to quickly build AI into new and existing applications.

Google launches Firebase Genkit, a new open source framework for building AI-powered apps

In the coming months, Google says it will open up the Gemini Nano model to more developers.

Patreon and Grammarly are already experimenting with Gemini Nano, says Google

As part of the update, Reddit also launched a dedicated AMA tab within the web post composer.

Reddit introduces new tools for ‘Ask Me Anything,’ its Q&A feature

Here are quick hits of the biggest news from the keynote as they are announced.

Google I/O 2024: Here’s everything Google just announced

LearnLM is already powering features across Google products, including in YouTube, Google’s Gemini apps, Google Search and Google Classroom.

LearnLM is Google’s new family of AI models for education

The official launch comes almost a year after YouTube began experimenting with AI-generated quizzes on its mobile app. 

Google is bringing AI-generated quizzes to academic videos on YouTube

Around 550 employees across autonomous vehicle company Motional have been laid off, according to information taken from WARN notice filings and sources at the company.  Earlier this week, TechCrunch reported…

Motional cut about 550 employees, around 40%, in recent restructuring, sources say

The keynote kicks off at 10 a.m. PT on Tuesday and will offer glimpses into the latest versions of Android, Wear OS and Android TV.

Google I/O 2024: Watch all of the AI, Android reveals

Google Play has a new discovery feature for apps, new ways to acquire users, updates to Play Points, and other enhancements to developer-facing tools.

Google Play preps a new full-screen app discovery feature and adds more developer tools

Soon, Android users will be able to drag and drop AI-generated images directly into their Gmail, Google Messages and other apps.

Gemini on Android becomes more capable and works with Gmail, Messages, YouTube and more

Veo can capture different visual and cinematic styles, including shots of landscapes and timelapses, and make edits and adjustments to already-generated footage.

Google Veo, a serious swing at AI-generated video, debuts at Google I/O 2024

In addition to the body of the emails themselves, the feature will also be able to analyze attachments, like PDFs.

Gemini comes to Gmail to summarize, draft emails, and more

The summaries are created based on Gemini’s analysis of insights from Google Maps’ community of more than 300 million contributors.

Google is bringing Gemini capabilities to Google Maps Platform

Google says that over 100,000 developers already tried the service.

Project IDX, Google’s next-gen IDE, is now in open beta

The system effectively listens for “conversation patterns commonly associated with scams” in-real time. 

Google will use Gemini to detect scams during calls

The standard Gemma models were only available in 2 billion and 7 billion parameter versions, making this quite a step up.

Google announces Gemma 2, a 27B-parameter version of its open model, launching in June

This is a great example of a company using generative AI to open its software to more users.

Google TalkBack will use Gemini to describe images for blind people

Google’s Circle to Search feature will now be able to solve more complex problems across psychics and math word problems. 

Circle to Search is now a better homework helper

People can now search using a video they upload combined with a text query to get an AI overview of the answers they need.

Google experiments with using video to search, thanks to Gemini AI

A search results page based on generative AI as its ranking mechanism will have wide-reaching consequences for online publishers.

Google will soon start using GenAI to organize some search results pages

Google has built a custom Gemini model for search to combine real-time information, Google’s ranking, long context and multimodal features.

Google is adding more AI to its search results

At its Google I/O developer conference, Google on Tuesday announced the next generation of its Tensor Processing Units (TPU) AI chips.

Google’s next-gen TPUs promise a 4.7x performance boost

Google is upgrading Gemini, its AI-powered chatbot, with features aimed at making the experience more ambient and contextually useful.

Google’s Gemini updates: How Project Astra is powering some of I/O’s big reveals