latentbrief
Back to news
Research1w ago

MIT and Microsoft Unveil System to Optimize AI Workflows

MIT News AI1 min brief

In brief

  • Researchers from MIT and Microsoft have developed a new system called Murakkab that streamlines the design and deployment of complex AI workflows.
    • These workflows, which involve chaining together multiple AI models and tools, are increasingly critical for tasks like video analysis but often suffer from inefficiencies leading to wasted resources.
  • The Murakkab system allows developers to describe their desired workflow in plain language without specifying detailed technical requirements upfront.
    • It then automatically selects the optimal models, tools, hardware configurations, and resource allocations based on user priorities such as cost or speed.
  • The innovation significantly reduces computational units needed, cutting energy consumption and costs without compromising performance.
    • This breakthrough addresses a major challenge for cloud providers who rely on these workflows to power their services while facing growing concerns over energy usage and resource allocation efficiency.
  • The system dynamically adjusts configurations in real-time to ensure optimal resource use, making it a win-win solution for both developers and service providers.
    • This development marks an important step toward more efficient AI-powered systems, with potential applications across industries.
  • As cloud providers continue to adopt advanced workflows, tools like Murakkab could play a crucial role in minimizing costs and improving sustainability.

Terms in this brief

Murakkab
A system developed by MIT and Microsoft to optimize AI workflows by automatically selecting the best models, tools, hardware configurations, and resource allocations based on user priorities like cost or speed. It allows developers to describe their desired workflow in plain language without detailed technical requirements upfront.

Read full story at MIT News AI

More briefs