CVEDIA, a computer vision AI provider, has solved the domain adaptation gap using its proprietary synthetic data pipeline. The domain gap, a major bottleneck in AI development, is the inability for algorithms trained on synthetic data to perform as well as those trained on real data. CVEDIA is claiming a precision improvement of 170% while sustaining a gain of 160% on recall over benchmarks. Solving this problem will allow AI technologies to scale without the burdens of data collection and labeling.
The company has recently announced several high-profile clients who have integrated the breakthrough technology into their products, including smart city systems provider Cubic, smart home technology company Vivint, and defense supplier FLIR, recently acquired by Teledyne Technologies.
The domain gap has long been a problem posed to computer vision engineers – significant because data poses extreme barriers to project completion. VentureBeat reports that only 13% of data science projects are completed¹ – often due to data issues including high expenses, the extensive time involved to collect it, and staffing requirements associated with technical utilization. By solving for the domain adaptation gap, clients will be able to sidestep the entire data process with completely synthetic AI, built without ever collecting data.
"We're the first synthetic data company that can create resilient AI with zero data," says Arjan Wijnveen, CEO and Co-founder of CVEDIA. "We've improved our technology to a point where it usually works even better than real data, because synthetic data can eliminate bias. We have clients that have been waiting for this technology to exist because it opens up possibilities for new products and markets."