Uber Technologies’ artificial intelligence (AI) data services business, Uber AI Solutions, has added new solutions and now makes them available to AI labs and enterprises in 30 countries.
Uber AI Solutions offers other businesses solutions that Uber has developed over the past decade while using data and AI in its own operations, the company said in a Friday (June 20) press release.
“We’re bringing together Uber’s platform, people and AI systems to help other organizations build smarter AI more quickly,” Megha Yethadka, general manager and head of Uber AI Solutions, said in the release. “With today’s updates, we’re scaling our platform globally to meet the growing demand for reliable, real-world AI data.”
One solution offered by Uber AI Solutions is a platform that connects enterprises to global talent who can provide annotation, translation and editing for multilingual and multimodal content, according to the release. The available talent includes experts in coding, finance, law, science and linguistics.
The company also offers datasets to train large AI models for generative AI, mapping, speech recognition and other use cases; task flows, annotations, simulations and multilingual support to help train AI agents; and its own internal platforms for managing large-scale annotation projects and validating AI outputs, per the release.
“With these advancements, Uber AI Solutions is poised to become the human intelligence layer for AI development worldwide — combining software, operational expertise and its massive global scale,” the release said.
The AI industry has faced a shortage of high-quality data for training AI models, PYMNTS reported in July.
While the internet generates enormous amounts of data daily, quantity doesn’t necessarily translate to quality when it comes to training AI models. Researchers need diverse, unbiased and accurately labeled data, and that combination is becoming increasingly scarce.
In another, separate development in this space, SandboxAQ said Wednesday (June 18) that it launched a dataset designed to help researchers advance AI models in drug discovery.
Generated with the use of SandboxAQ’s AI large quantitative model capabilities and Nvidia’s development platform for AI training and fine-tuning, the dataset enables researchers to train AI models to accurately predict protein-ligand binding affinities at least 1,000 times faster than traditional physics-based methods.
