Qualcomm incorporated (20240118923). ROBUST SCHEDULING WITH GENERATIVE FLOW NETWORKS simplified abstract

ROBUST SCHEDULING WITH GENERATIVE FLOW NETWORKS

Organization Name

qualcomm incorporated

Inventor(s)

Corrado Rainone of Haarlem (NL)

Wei David Zhang of Amsterdam (NL)

Roberto Bondesan of London (GB)

Markus Peschl of Amsterdam (NL)

Mukul Gagrani of Milpitas CA (US)

Wonseok Jeon of San Diego CA (US)

Edward Teague of San Diego CA (US)

Piero Zappi of La Jolla CA (US)

Weiliang Zeng of San Diego CA (US)

Christopher Lott of San Diego CA (US)

ROBUST SCHEDULING WITH GENERATIVE FLOW NETWORKS - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240118923 titled 'ROBUST SCHEDULING WITH GENERATIVE FLOW NETWORKS

Simplified Explanation

The abstract describes a processor-implemented method for generating schedules for tasks associated with artificial neural networks and selecting the best schedule based on testing.

• The method involves generating schedules for tasks associated with artificial neural networks.
• Each schedule assigns operations of the neural network to specific processors on a hardware device.
• The schedules are tested on the hardware device or a model of the device to determine their efficiency.
• The best schedule is selected based on the testing results, meeting specific selection criteria.

Potential Applications

This technology could be applied in various fields such as:

• High-performance computing
• Artificial intelligence
• Robotics

Problems Solved

This technology addresses the following issues:

• Optimizing task scheduling for neural network operations
• Improving hardware device efficiency
• Enhancing overall system performance

Benefits

The benefits of this technology include:

• Faster task execution
• Better resource utilization
• Increased system reliability

Potential Commercial Applications

With its ability to optimize task scheduling for neural networks, this technology could be valuable in industries such as:

• Data centers
• Autonomous vehicles
• Medical imaging

Possible Prior Art

One possible prior art for this technology could be research papers or patents related to task scheduling optimization for neural networks on hardware devices.

How does this technology compare to existing scheduling methods for neural networks on hardware devices?

This article does not provide a direct comparison with existing scheduling methods. It would be helpful to understand the specific advantages or differences this technology offers.

What are the specific selection criteria used to choose the best schedule?

The abstract mentions selecting a schedule based on testing and a selection condition, but the exact criteria are not detailed. Understanding the selection process would provide insights into the decision-making algorithm used in this technology.

Original Abstract Submitted

a processor-implemented method includes generating, by a scheduling model, a group of schedules from a computation graph associated with a task, each node on the computation graph being associated with an operation of an artificial neural network, each schedule of the group of schedules associating each node of the computation graph with a processor of a group of processors of a hardware device. the processor-implemented method also includes testing one or more schedules of the group of schedules on the hardware device or a model of the hardware device. the processor-implemented method further includes selecting a schedule of the one or more schedules based on testing the one or more schedules, the selected schedule satisfying a selection condition.