Intel corporation (20240176998). METHODS AND APPARATUS FOR DISCRIMINATIVE SEMANTIC TRANSFER AND PHYSICS-INSPIRED OPTIMIZATION OF FEATURES IN DEEP LEARNING simplified abstract

METHODS AND APPARATUS FOR DISCRIMINATIVE SEMANTIC TRANSFER AND PHYSICS-INSPIRED OPTIMIZATION OF FEATURES IN DEEP LEARNING

Organization Name

intel corporation

Inventor(s)

Anbang Yao of Beijing (CN)

Hao Zhao of Beijing (CN)

Ming Lu of Beijing (CN)

Yiwen Guo of Beijing (CN)

Yurong Chen of Beijing (CN)

METHODS AND APPARATUS FOR DISCRIMINATIVE SEMANTIC TRANSFER AND PHYSICS-INSPIRED OPTIMIZATION OF FEATURES IN DEEP LEARNING - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240176998 titled 'METHODS AND APPARATUS FOR DISCRIMINATIVE SEMANTIC TRANSFER AND PHYSICS-INSPIRED OPTIMIZATION OF FEATURES IN DEEP LEARNING

Simplified Explanation

The patent application describes methods and apparatus for discriminative semantic transfer and physics-inspired optimization in deep learning.

• A computation training method for a convolutional neural network (CNN) involves receiving a sequence of training images in the CNN of a first stage to describe objects of a cluttered scene as a semantic segmentation mask.
• The semantic segmentation mask is then received in a semantic segmentation network of a second stage to produce semantic features.
• By using weights from the first stage as feature extractors and weights from the second stage as classifiers, edges of the cluttered scene can be identified using the semantic features.

Potential Applications

The technology described in the patent application could be applied in various fields such as computer vision, autonomous driving, robotics, and medical imaging.

Problems Solved

This technology helps in improving the accuracy and efficiency of object detection and segmentation in cluttered scenes, which can be challenging for traditional deep learning models.

Benefits

The benefits of this technology include enhanced semantic segmentation, improved object detection in complex environments, and optimized deep learning training processes.

Potential Commercial Applications

Potential commercial applications of this technology include developing advanced surveillance systems, enhancing medical image analysis software, and improving autonomous navigation systems.

Possible Prior Art

One possible prior art for this technology could be the use of transfer learning and optimization techniques in deep learning models for image recognition tasks.

How does this technology compare to existing object detection and segmentation methods in terms of accuracy and efficiency?

The patent application does not provide a direct comparison with existing methods in terms of accuracy and efficiency. Further research or testing may be needed to evaluate the performance of this technology against current approaches.

What are the computational requirements for implementing this technology in real-time applications?

The patent application does not specify the computational requirements for real-time implementation. Understanding the computational resources needed for deploying this technology in real-world scenarios would be crucial for its practical application.

Original Abstract Submitted

methods and apparatus for discrimitive semantic transfer and physics-inspired optimization in deep learning are disclosed. a computation training method for a convolutional neural network (cnn) includes receiving a sequence of training images in the cnn of a first stage to describe objects of a cluttered scene as a semantic segmentation mask. the semantic segmentation mask is received in a semantic segmentation network of a second stage to produce semantic features. using weights from the first stage as feature extractors and weights from the second stage as classifiers, edges of the cluttered scene are identified using the semantic features.