Microsoft has announced that Lobe, a machine learning tool that helps people apply deep learning and AI models quickly – without the need of writing code – into tools they are developing, is now available with image classification support. This essentially means that people can import images of the things they want Lobe to recognize, and the free app automatically selects the right machine learning architecture to begin training a machine learning model. The company says that it is making the app available this week in public preview, and it can be downloaded on Windows or Mac computers for free.
“Today, Lobe supports image classification but plans to expand to other models and data types in the future,” Jennifer Langston, who writes about Microsoft research and innovation, said in a blog post. Lobe’s visual interface allows developers to create apps with features such as reading handwriting, recognizing hand gestures, hearing music, etc. You can download Lobe for Windows and Mac from the Lobe page.
According to Microsoft, “Lobe automatically selects the right machine learning architecture and starts training without any setup or configuration.” Users can evaluate the model’s strengths and weaknesses via real-time visual results and modify the model to boost performance. Once they’re done, users can export their models to a variety of industry-standard formats and ship it on any platform they chose, according to Microsoft’s Lobe web page.
Microsoft acquired the AI-focused firm Lobe back in 2018 as a part of its strategy and CEO Satya Nadella’s vision to build AI into everything that the company makes. Lobe helps developers, as well as people who do not have any data science experience, to apply deep learning and AI models quickly into tools that can be used in industries like healthcare or agriculture. Lobe leverages the tech giant’s AI research, global infrastructure, and developer tools.
Microsoft says that Lobe uses “open-source machine learning architectures and transfer learning to train custom machine learning models on the user’s own machine”. The company claims that the data is kept private and there is no requirement for an Internet connection or logins. With image classification support, people can just simply import images of the things they want Lobe to be trained with, and the training is done automatically. Furthermore, it complements Azure AI’s services for customers looking to leverage cloud computing capabilities, the company says.