Despite significant progress over the last decade, current AI (artificial intelligence) models can perform poorly and unpredictably when deployed in the real-world, mounting increasing concerns on their safety and reliability. Such reliability issues affect the use of AI across virtually all domains – engineering, healthcare, financial services, autonomous cars, and many others.
The newly established spinoff LatticeFlow is developing a product that enables AI teams to build and deploy AI models that are safe, reliable, and trustworthy. The product empowers enterprises to assess and improve their AI models and datasets, identify critical failure modes and safeguard AI models deployed in production. Large enterprises and governments in Europe and the US, including Swiss Federal Railways (SBB), Germany’s Federal Cyber Security Bureau, and the US Army are already using the system.
Ilir Fetai and Andre Roger, who led the Center of Competence in Machine Perception at Swiss railways, said: “Machine Learning (ML) is one of the core topics at SBB, as we see a huge potential in its application for an improved, intelligent and automated monitoring of our railway infrastructure. The project on robust and reliable AI with LatticeFlow, ETH, and Siemens has a crucial role in enabling us to fully exploit the advantages of using ML.”
Fresh capital to accelerate the development
LatticeFlow has now closed $2.8M in a financing round to accelerate its product development as the company embarks on a new startup journey. The round was led by btov, a European venture capital firm known for its investments in DeepL, SumUp and Volocopter, and Global Founders Capital, the largest tech fund in Europe that previously backed Zalando, Revolut, and Slack.
This is not the first time the scientists take the entrepreneurial route. Vechev previously co-founded DeepCode, an ETH spin-off that built the first AI code review system and was acquired by the cybersecurity unicorn Snyk in 2020. Together with Tsankov, Vechev also co-founded ChainSecurity, an ETH spin-off that became the world-leader in the formal smart contract security audits field, acquired by PwC in 2020.
Experienced team with a sturdy track record
Creating more reliable ML models is challenging as it requires a new generation of tools that leverage techniques from diverse areas such as symbolic reasoning and machine learning. LatticeFlow’s founders are positioned to address this problem as they possess an exceptional track record and leadership in the area of trustworthy AI. For example, in 2017, professor Martin Vechev and his team introduced ERAN, a system that enabled, for the first time, the verification of large deep learning models with millions of parameters. In July 2020, ERAN achieved the highest score at VNN-COMP'20, the first competition for certifying neural networks, outperforming teams from MIT, Oxford, UIUC, and others.
The founders are also involved in the broader mission to accelerate interdisciplinary research and adopt robust and reliable AI. In October 2020, Professor Andreas Krause, together with ETH Zurich, founded the new ETH AI Center where trustworthy AI is a core topic. Krause is also a founding academic co-director of the Swiss Data Science Center whose goal is to boost AI adoption in science and industry.