Microsoft announced yesterday several new features for developers at its Azure Open Source Day 2023. The company explained that these capabilities enable software developers to quickly build intelligent applications.
“Most companies have leveraged AI to improve efficiency and costs. Large AI applications leveraging natural language processing (NLP), automatic speech recognition (ASR), and text-to-speech (TTS) are becoming prevalent, but what powers these applications is the underlying infrastructure optimized for large AI workloads. Azure is the best place to build AI workloads,” explained Eric Boyd, CVP for AI Platform.
First up, Microsoft has announced the public preview of new Vision Services that help developers to create computer vision apps across industries. Vision Services provide several features such as background removal, smart cropping, automatic captioning, image search, model customization, and video summarization. Microsoft has also introduced new AI controls to help users with real-time alerts, environment analysis, and movement tracking.
Microsoft announced its plans to launch foundation models in Azure Machine Learning. The new service will enable users to deploy pre-trained open-source foundation models for various machine-learning tasks, such as multi-modality and natural language processing. It eliminates the need for users to manage and optimize dependencies manually.
Microsoft unveils new Responsible AI Toolbox additions
Microsoft has recently introduced two new open-source tools to increase the adoption of responsible AI practices. The Responsible AI Mitigations Library makes it easier to mitigate errors in Machine Learning models.
Finally, the Responsible AI Tracker extension lets users view models, code, and visualization artifacts within the same framework. Microsoft claims that it should help to speed up the model interaction and evaluation process. You can check out Microsoft’s blog post to learn more about new capabilities announced at the Azure Open Source Day 2023.