9.3 C
London
Monday, November 28, 2022

HiddenLayer emerges from stealth to protect AI models from attacks – londonbusinessblog.com

Must read

Having a funny bot like Chandler Bing’s can make you a hot item on Tinder

Dinner dates may be dead, as picnics, coffee dates, standups, long drives, and cooking were among the most popular to date ideas. Going...

Dell Technologies demonstrates that network infrastructure spending is robust

Supplier of hardware and infrastructure solutions Dell Technologies (NASDAQ: DELL) is a diversified technology company consisting of two main segments, infrastructure and customer solutions....

Top 5 Financially Exhausting Divorce Settlements in History

The concept of "broken families" has been around since the dawn of humanity. In fact, divorce is a normal part of life these...
Shreya Christinahttps://londonbusinessblog.com
Shreya has been with londonbusinessblog.com for 3 years, writing copy for client websites, blog posts, EDMs and other mediums to engage readers and encourage action. By collaborating with clients, our SEO manager and the wider londonbusinessblog.com team, Shreya seeks to understand an audience before creating memorable, persuasive copy.

As AI-powered services like OpenAI’s GPT-3 grow in popularity, they are becoming an increasingly attractive attack vector. Even if shielded behind an API, hackers can try to reverse engineer the models underlying these services or use “conflicting” data to tamper with them. According to According to Gartner, 30% of all AI cyberattacks by 2022 will employ these techniques, along with data poisoning, in which bad data is injected into the dataset used to train models to attack AI systems.

As in any industry, combating security threats is a never-ending task. But Chris Sestito claims that his platform, Hidden Layercan simplify things for AI-as-a-service vendors by automatically identifying malicious activity against models and responding to attacks.

HiddenLayer emerged from stealth today with $6 million in seed funding from Ten Eleven Ventures, Secure Octane and other investors. Sestito, the former director of threat research at Cylance and VP of engineering at Qualys, co-founded the company several months ago with Tanner Burns and Jim Ballard. Burns and Ballard also worked at Qualys and Cylance and spent time together at BlackBerry, where Ballard was a team leader for data management and Burns was a threat researcher.

“Virtually all business organizations have made significant contributions to machine learning to give itself an advantage — whether that value is in the form of product differentiation, monetization, cost savings, or efficiency,” Sestito told londonbusinessblog.com in an email interview. “Disappointing machine learning attacks can do the same damage we’ve seen with traditional cyberattacks, including exposing customer data and destroying production systems. In fact, at HiddenLayer, we believe we are not far from seeing machine learning models paid back to their organizations.”

HiddenLayer claims its technology can protect models from attacks without access to a vendor’s raw data or algorithms. By analyzing model interactions — in other words, the data entered into the model (e.g., an image of cats) and the predictions the model makes (e.g., the caption “cats”) — to discover patterns that could be malicious HiddenLayer can work “non-invasively” and without prior knowledge of training data, Sestito said.

“Adversarial machine learning attacks aren’t as loud as ransomware — you have to look for them to catch them in time,” Sestito said. “HiddenLayer has focused on a research approach that allows us to publish our findings and train the world to be prepared.”

Mike Cook, an AI researcher who is part of the Knives and brushes collectively, said it’s unclear whether HiddenLayer is doing anything “truly groundbreaking or new.” (Cook is not affiliated with HiddenLayer.) Still, he notes there’s an upside to what HiddenLayer appears to be doing: trying to pool knowledge about attacks on AI and make them more widely accessible.

“The AI ​​boom is still booming, but much of that knowledge about how modern machine learning works and how best to use it is still largely hidden from those with specialist knowledge. Memorable examples for me include researchers who manage to extract individual training data from OpenAI’s GPT-2 and GPT-3 systems,” Cook told londonbusinessblog.com via email. “When expert knowledge is inaccessible and difficult to obtain, sometimes a business just needs convenient ways to reach it.”

HiddenLayer is currently pre-revenue and has no clients, although Sestito says the startup has engaged several “high-profile” design partners. Ultimately, Cook believes its success will depend less on HiddenLayer’s technology and more on whether the threat of attack is as great as the company claims.

“I don’t know how often attacks on machine learning systems occur [at present]. Tricking a spam filter into letting an email through is different in scale and seriousness from extracting proprietary data from a large language model,” Cook said.

At his point, it is difficult to identify real-life examples of attacks on AI systems. Research on the topic has exploded, with more than 1,500 articles on AI security published in 2019 on the scientific publishing site Arxiv.org, up from 56 in 2016, according to a study from Adversara. But there’s little public coverage of attempts by hackers to attack, say, commercial facial recognition systems — assuming such attempts happen at all.

Sestito claims that the threat – regardless of its magnitude today – will grow with the AI ​​market, implicitly in HiddenLayer’s favor. He acknowledges that several startups are already offering products designed to make AI systems more robust, including Robust Intelligence, CalypsoAI, and Troj.ai. But Sestito claims that HiddenLayer stands alone in its AI-driven detection and response approach.

“PwC believes AI will be a $15.7 trillion market by 2030. We absolutely need to defend this technology now,” said Sestito. “Our biggest goal by far is to educate the market about this new threat. The commitment to AI and machine learning is relatively new to many organizations and few have focused on defending those assets. With every new technology comes new attack vectors; this is the same struggle on a new frontier.”

Austin, Texas-based HiddenLayer currently has 11 employees and expects to end 2022 with 14.

More articles

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Latest article

Having a funny bot like Chandler Bing’s can make you a hot item on Tinder

Dinner dates may be dead, as picnics, coffee dates, standups, long drives, and cooking were among the most popular to date ideas. Going...

Dell Technologies demonstrates that network infrastructure spending is robust

Supplier of hardware and infrastructure solutions Dell Technologies (NASDAQ: DELL) is a diversified technology company consisting of two main segments, infrastructure and customer solutions....

Top 5 Financially Exhausting Divorce Settlements in History

The concept of "broken families" has been around since the dawn of humanity. In fact, divorce is a normal part of life these...

The Untold Truth of Master P

Who is Master P's daughter Cymphonique Miller? Wiki bio Cymphonique Miller was born in New Orleans, Louisiana, USA, on August 1, 1996 – her...