18096650. APPARATUS FOR PROCESSING A DEEP LEARNING MODEL AND A METHOD THEREOF simplified abstract (HYUNDAI MOTOR COMPANY)

APPARATUS FOR PROCESSING A DEEP LEARNING MODEL AND A METHOD THEREOF

Organization Name

HYUNDAI MOTOR COMPANY

Inventor(s)

Jin Sol Kim of Hwaseong-si (KR)

APPARATUS FOR PROCESSING A DEEP LEARNING MODEL AND A METHOD THEREOF - A simplified explanation of the abstract

This abstract first appeared for US patent application 18096650 titled 'APPARATUS FOR PROCESSING A DEEP LEARNING MODEL AND A METHOD THEREOF

Simplified Explanation

The patent application describes an apparatus for processing a deep learning model by optimizing memory usage and processing time for each layer of the model.

• The controller in the apparatus detects memory usage, processing times for different memories, and determines the optimal memory for each layer based on an objective function.

• This innovation aims to improve the efficiency and performance of deep learning models by dynamically selecting the most suitable memory for each layer.

• By optimizing memory usage and processing time, the apparatus can enhance the overall speed and accuracy of deep learning model computations.

• Potential applications of this technology include accelerating training and inference processes in various deep learning applications such as image recognition, natural language processing, and autonomous driving systems.

• The problems solved by this technology include reducing memory bottlenecks and improving overall performance in deep learning models.

• The benefits of this technology include faster processing speeds, improved accuracy, and more efficient memory utilization in deep learning tasks.

Potential Commercial Applications

The potential commercial applications of this technology in optimizing memory usage for deep learning models can be seen in industries such as healthcare (medical image analysis), finance (fraud detection), and e-commerce (recommendation systems).

Possible Prior Art

One possible prior art in this field is the use of memory optimization techniques in traditional machine learning algorithms to improve performance and efficiency. However, the specific approach of dynamically selecting optimal memory for each layer of a deep learning model may be a novel aspect of this innovation.

Unanswered Questions

How does this technology compare to existing memory optimization techniques in deep learning models?

The article does not provide a direct comparison with existing memory optimization techniques in deep learning models. Further research or a comparative study would be needed to evaluate the effectiveness and efficiency of this innovation in comparison to other methods.

What impact could this technology have on the energy consumption of deep learning systems?

The article does not address the potential impact of this technology on the energy consumption of deep learning systems. Understanding how memory optimization affects energy usage could be crucial for assessing the overall sustainability and cost-effectiveness of implementing this innovation.

Original Abstract Submitted

An apparatus for processing a deep learning model includes a first memory, a second memory, and a controller. The controller is configured to, for each layer of the deep learning model, detect memory usage, a first processing time corresponding to the first memory being used, and a second processing time corresponding to the second memory being used, and determine an optimal memory for each layer of the deep learning model based on an objective function.