Led by ex-AWS AI exec, Abacus.AI raises $13M to automate neural network creation
Abacus.AI Inc., an artificial intelligence startup led by Amazon Web Services Inc. and Google LLC veterans, today announced it has raised a $13 million funding round from a group of prominent tech industry figures.
Noted venture capitalist Mike Volpi led the round. Volpi, known for his role at early Facebook Inc. investor Index Ventures, was joined in the round by former Google Chairman Eric Schmidt, Yahoo co-founder Jerry Yang, Minted Inc. founder Mariam Naficy and others.
Abacus.AI was until today known as RealityEngines. The startup entered the limelight last year, when it nabbed a $5.25 million seed round before even introducing its first product.
Abacus.AI Chief Executive Officer Bindu Reddy was the general manager for AI verticals at AWS before co-founding the startup. Chief Technology Officer Arvind Sundararajan and research director Siddartha Naidu previously both held senior engineering roles at Google. Sundararajan led development on the Gmail backend, while Naidu is credited as the founder of Google Cloud’s BigQuery data warehouse service.
In conjunction with the funding, Abacus.AI today launched a new offering described as an “autonomous AI service.” It uses a machine learning optimization technology the startup’s researchers first detailed in an academic paper last year to help enterprises develop AI models faster.
According to Abacus.AI, its service can assess the dataset that a company wants to process and the project’s intended use case to find the most suitable network for the task. It then automatically trains the neural network to hone its accuracy and performance. Once the tuning is complete, the ready-to-use AI is made accessible to developers via an application programming interface.
“Abacus.AI’s expert AI engines use use-case specific neural architecture search to find the best neural network for a problem based on the training dataset,” Reddy told SiliconANGLE. “When training data is not sufficient … Abacus.AI uses generative modeling to augment the customer dataset with synthetic data.”
The service also helps with tasks such as monitoring AI performance and retraining neural networks when new data is available, according to Reddy.
Abacus.AI claims that an AI developed with its service can provide accuracy levels comparable to hand-crafted models. The service’s launch into general availability follows a beta that saw the participation of more than 1,200 users from companies that included contract manufacturer Flex Inc., which makes Apple Inc.’s Mac Pro.
Abacus.AI’s research team today also released an open-source library designed to help AI developers remove bias from their models.
The new funding will go toward “expanding our research science and engineering team,” Reddy said. “We also plan to continue to scale our customer base and support language [and] vision use cases.”
Since you’re here …
Show your support for our mission with our one-click subscription to our YouTube channel (below). The more subscribers we have, the more YouTube will suggest relevant enterprise and emerging technology content to you. Thanks!
Support our mission: >>>>>> SUBSCRIBE NOW >>>>>> to our YouTube channel.
… We’d also like to tell you about our mission and how you can help us fulfill it. SiliconANGLE Media Inc.’s business model is based on the intrinsic value of the content, not advertising. Unlike many online publications, we don’t have a paywall or run banner advertising, because we want to keep our journalism open, without influence or the need to chase traffic.The journalism, reporting and commentary on SiliconANGLE — along with live, unscripted video from our Silicon Valley studio and globe-trotting video teams at theCUBE — take a lot of hard work, time and money. Keeping the quality high requires the support of sponsors who are aligned with our vision of ad-free journalism content.