All the sessions from Transform 2021 are readily available on-demand now. Watch now.
Large cloud vendors like Amazon, Google, Microsoft, and IBM supply APIs enterprises can use to take benefit of effective AI models. But comparing these models — each in terms of functionality and expense — can be difficult without the need of thorough arranging. Moreover, the siloed nature of the APIs tends to make it tough to unify services from unique vendors into a single app or workflow without the need of custom engineering work, which can be pricey.
These challenges inspired Samy Melaine and Taha Zemmouri to discovered Eden AI (previously AI-Compare) in 2020. The platform draws on AI APIs from a variety of sources to permit organizations to mix and match models to suit their use case. Eden AI not too long ago launched what it calls an AI management platform, which the firm claims simplifies the use — and integration — of different models for finish customers.
Prior to beginning Eden AI, Melaine and Zemmouri worked as consultants on information science and AI projects at startup DataGenius, exactly where they collaborated with corporations to leverage AI models from different providers. While there, they realizated information processing projects normally involve a number of unique APIs and acquiring the very best functionality — and expense savings — may call for combining models from unique providers, like Google Cloud’s optical character recognition combined with IBM’s translation and Amazon Web Services’ keyword extraction.
Unifying AI models
Eden AI’s new providing, which can deployed in the public cloud or on-premises, connects to AI models for pc vision, all-natural language processing, speech recognition, and machine mastering to permit customers to access, test, and evaluate the expenses of vendor-distinct models. From a dashboard, organizations can build AI projects, produce keys, full payment, upload datasets, and handle and delete AI projects from unique providers.
For instance, making use of Eden AI, a firm could feed a document in Chinese into Google Cloud Platform’s optical character recognition service to extract its contents. Then it could have an IBM Watson model translate the extracted Chinese characters into English words and queue up an Amazon Web Services API to analyze for key phrases.
Eden AI tends to make income by charging providers a commission on the revenues generated by its platform.
“We are driven by the desire to democratize the use of AI by offering our users simplified access to these technologies. In particular, we want to highlight innovative providers offering high-performance engines,” Eden AI writes on its web site. “We believe that the use of AI engines will continue to develop in companies and that it becomes a commodity that must be easily integrated. We are therefore working to guarantee to facilitate the test and use of different AI engines.”
In the future, Eden AI aims to integrate open supply models and construct an algorithm that can automatically recommend the very best AI model — educated in element on metadata from Eden AI’s consumers. The firm also plans to allow customers to deploy their personal algorithms on the platform, which they’ll be in a position to benchmark against third-party models to gauge their relative functionality.