New Galileo accurate Computer Vision Intelligence uses Machine Learning
Galileo, the first-ever machine-learning (ML) data intelligence company for unstructured data, today announced the launch of its proprietary data-quality intelligence platform, called Galileo Data Intelligence for Computer Vision.
The first-ever solution to solve for data quality issues across the entire ML workflow, the Galileo platform will allow data scientists and ML engineers to automate the ‘needle in the haystack’ approach, reducing model production time by 10x, improving model accuracy by 15% across the board and reducing data labeling costs for human-labeled datasets by 40%.
As the global datasphere expands, 80% of the anticipated 163 zettabytes available by 2025 will be unstructured, increasing the risk of errors and model production inefficiencies by forcing data scientists and ML engineers to manually track down and diagnose problems within models. The vast majority — 84% — of data scientists and ML engineers report that this ‘needle in a haystack’ approach to model error detection is “an issue for their teams at least some of the time,” according to a recent survey.
By adding just a few lines of Python code during the model training process, the innovative Galileo Data Intelligence for Computer Vision platform automatically identifies problematic data that negatively impacts model performance, then suggests effective solutions for data-science teams to seamlessly address the issue.
With the Galileo platform, engineers will be able to address a major bottleneck in the data-science workflow, which will allow for more efficiency and accuracy in iterations as well as in image classification, object detection and semantic segmentation (pixel-level) models.
“Data science applications across industries are rapidly expanding. Unfortunately, so too are the challenges for ML and data science practitioners, many of whom are forced to spend untold amounts of time managing data quality issues to create high-quality models — an issue our team has experienced firsthand, and one we sought to resolve by founding Galileo,” said Vikram Chatterji, co-founder and CEO of Galileo. “Galileo Data Intelligence for Computer Vision will create significant efficiencies for our customers — allowing data scientists to work more quickly and effectively than ever before across the cyclical ML workflow, whether that be data preparation ahead of labeling, during training iterations or in monitoring production models.”
Galileo Data Intelligence for Computer Vision is specifically tailored for data science teams that are passionate about taking control of their data and maximizing the performance of their models.
This innovative platform is ideal for teams building image classification, object detection and semantic segmentation models across a broad range of industries, including e-commerce, retail technology, video intelligence, autonomous vehicles, healthcare, customer experience and more. Galileo’s growing list of satisfied customers, which includes Uniphore, Headspace Health, Spectrum Labs, Involve AI, Legartis, Enterpret and numerous other software companies dedicated to enhancing data science and ML team operations, is a testament to the platform’s effectiveness.
“Voxel leverages computer vision to fundamentally change how companies manage their physical operations by enhancing security cameras with real-time AI,” said Harishma Dayanidhi, co-founder of Voxel, a Galileo customer. “To do this well, it is critical to identify the data that severely degrades the model performance. Galileo empowers our image perception team by automatically surfacing such erroneous data and fixing them — thereby allowing us to build better, safer models 10x faster!”
