AI model testing firm Robust Intelligence has raised $ 30 million in a Series A funding round led by Tiger Global and Sequoia. The company was co-founded by CEO Yaron Singer, an Israeli professor of computer science and applied mathematics at Harvard University, and plans to expand its business in Israel. The company has raised $ 45 million to date.

Robust Intelligence is developing an AI firewall that wraps around the organization’s artificial intelligence (AI) model, monitoring and remedying data that causes errors and risks in real time. The product inspects the data before it is entered into the AI ​​model using advanced machine learning techniques, alerting and correcting the data in real time, thus preventing the model from going wrong. Robust Intelligence already has dozens of paying customers and employs 50 developers with advanced university degrees at the company’s headquarters in San Francisco.

The company works with dozens of Fortune 500 and Fortune 100 companies, including PayPal, Expedia, Tokio Marine (the world’s largest Japanese insurance company), leading diagnostic and medical device companies, government agencies, etc. In Israel, Robust Intelligence works with leading companies in finance, insurance and more.

Professor Singer said, “We are keen to expand our presence in Israel by recruiting clients as well as talent specializing in software development, algorithms and artificial intelligence. Until recently, the idea of ​​AI Firewall seemed impossible. We invest enormous resources in finding people who believe in themselves and have no limits. Israel is the natural place to find such people.

The Robust Intelligence product enables organizations to minimize the dangers inherent in AI and technology teams to focus on system development, rather than debugging and firefighting models in production. The company’s AI firewall surrounds the AI ​​model while monitoring and even correcting any input that could disrupt the model’s results in real time. The product is based on a system that learns tests and monitoring measures related to the AI ​​model it protects. These tests include “stress tests” for the AI ​​model, which also challenges it with basic errors and exposure to extreme conditions and also monitors input and output anomalies for each model.

Professor Singer said: “The weaknesses and errors of AI models emerge once the product is activated and given the current widespread use of AI in automatic decision making, it is terrifying to know that systems operate without protection. Our company’s mission is to eliminate the risks businesses take when using AI systems. We save our customers a lot of time and money, but more importantly, we protect people from bad decisions that could result from AI mistakes.

As an example, Professor Singer described a client who provides AI models to banks and financial institutions to identify financial fraud. While installing the AI ​​firewall, they discovered that the models were wrong, providing an entry where the data field for the country the transfer originated from was lowercase and not uppercase. In other words, if the data field shows us (underscore) instead of US, the model got it wrong. These mistakes lead to huge financial losses and put financial institutions using AI and their customers at significant risk. “In this case, the error was not intentional,” he explains. After all, the data was compiled from different systems and the format has changed. The problem is, even such minor changes can cause AI models to produce unexpected results. Other examples include the decisions made by credit card and insurance companies because they are based on AI systems based on statistical learning. In such systems, the data used to train the learning machine can be biased in favor of certain populations, creating a system that makes decisions that are also biased in their favor. Such system errors could prevent people from getting a loan for a home or, worse yet, health insurance. “

Professor Singer, 42, was born and raised in Tel Aviv and served in an elite IDF intelligence unit. He then moved to the United States, where he obtained his doctorate. in Computer Science at UC Berkley. In 2011, he founded a startup that uses AI to analyze networks and joined Google. At Google, he worked on algorithms to speed up the formation of AI models to enable the tech giant’s products to design more accurate predictions. It was also around this time that he discovered how easily AI systems could get out of hand using incorrect or misleading data. As a professor at Harvard University School of Computer Sciences, his studies focused on the theoretical limitations of AI models, which are prone to errors, and on the development of advanced optimization methods to correct them.

Professor Singer founded Robust Intelligence in 2019 with Harvard students with whom he has published dozens of articles on the subject. Singer’s team achieved a breakthrough in algorithm planning, helped raise government awareness of the issue, and even secured an initial DARPA grant for the study.

Professor Singer said, “This latest round of investment will significantly accelerate the development of our AI Firewall product and RIME platform as well as our marketing and sales efforts. Our world is adopting AI technologies at an exponential rate and relying on AI. to make critical decisions. The problem, however, is that AI models frequently fail due to bad data, posing a huge risk to businesses and society as a whole. “

Posted by Globes, Israel business news – en.globes.co.il – December 12, 2021.

© Copyright of Globes Publisher Itonut (1983) Ltd., 2021. .