18323473. ENCODING METHOD AND ENCODING CIRCUIT simplified abstract (MACRONIX INTERNATIONAL CO., LTD.)

ENCODING METHOD AND ENCODING CIRCUIT

Organization Name

MACRONIX INTERNATIONAL CO., LTD.

Inventor(s)

Tzu-Hsiang Su of Hsinchu City (TW)

Jeng-Lun Shieh of Taichung City (TW)

Shanq-Jang Ruan of New Taipei City (TW)

ENCODING METHOD AND ENCODING CIRCUIT - A simplified explanation of the abstract

This abstract first appeared for US patent application 18323473 titled 'ENCODING METHOD AND ENCODING CIRCUIT

The application provides an encoding method and circuit that involves linear conversion, comparison, binding, addition, and operating functions to generate an output vector.

• Linear conversion of input into a first vector based on a weight by a convolution layer.
• Comparison of the first vector with a reference value to generate a second vector using an activation function.
• Binding the second vector with a random vector to produce multiple binding results.
• Adding the binding results to generate an adding result.
• Operating the adding result with a Signum function and a normalization function to create an output vector.

Potential Applications: This technology can be applied in image and speech recognition systems, natural language processing, and data compression algorithms.

Problems Solved: This technology addresses the need for efficient encoding methods in various machine learning and artificial intelligence applications.

Benefits: The technology offers improved accuracy, speed, and efficiency in data encoding processes, leading to enhanced performance in AI systems.

Commercial Applications: This technology can be utilized in industries such as healthcare, finance, automotive, and cybersecurity for tasks like image analysis, voice recognition, and data encryption.

Prior Art: Researchers can explore prior art related to convolutional neural networks, activation functions, and encoding methods in the field of artificial intelligence and machine learning.

Frequently Updated Research: Stay updated on the latest advancements in convolutional neural networks, activation functions, and encoding techniques to enhance the performance of AI systems.

Questions about Encoding Method and Circuit: 1. How does the activation function impact the generation of the second vector? 2. What are the potential challenges in implementing this encoding method in real-world applications?

Original Abstract Submitted

The application provides an encoding method and an encoding circuit. The encoding method includes: performing linear conversion on an input into a first vector based on a weight by a convolution layer; comparing the first vector generated from the convolution layer with a reference value to generate a second vector by an activation function; binding the second generated by the activation function with a random vector to generate a plurality of binding results; adding the binding results to generate an adding result; and operating the adding result by a Signum function and a normalization function to generate an output vector.