GSI Technology. has been granted a patent for a machine learning method that involves extracting features from inputs to create fixed-size keys. These keys are organized to minimize distance between similar inputs and maximize distance between dissimilar ones, enabling efficient K-nearest neighbor searches in constant time, regardless of dataset size. GlobalData’s report on GSI Technology gives a 360-degree view of the company including its patenting strategy. Buy the report here.
Smarter leaders trust GlobalData
Access deeper industry intelligence
Experience unmatched clarity with a single platform that combines unique data, AI, and human expertise.
According to GlobalData’s company profile on GSI Technology, Convolutional neural networks (CNNs) was a key innovation area identified from patents. GSI Technology's grant share as of July 2024 was 58%. Grant share is based on the ratio of number of grants to total number of patents.
Machine learning method using fixed-size neural network keys
The granted patent US12073328B2 outlines a method for machine learning that leverages a neural network to enhance the efficiency of data processing through an associative memory array. The method begins with the extraction of features from a training set of inputs, where each input generates a unique feature set that serves as a key in a key-value pair. The neural network organizes these keys such that the distance between keys representing similar inputs is minimized, while the distance between keys for dissimilar inputs is maximized. This arrangement allows for the storage of the dataset in columns of an associative memory array, enabling the activation of multiple rows to implement a K-nearest neighbor processor. This processor operates in constant time relative to the fixed size of the keys, independent of the dataset size, facilitating the identification of K keys that are similar to a given query key.
Additionally, the patent describes the implementation of a SoftMax unit within the associative memory array to process the K keys and produce query results. The method further defines the distance between keys based on the specific machine learning operation to be performed, which may include vector generation or classification. A similar approach is applied in a second method where training feature sets are generated to capture similarities between pairs of training inputs associated with the same object. These feature sets are also stored in the associative memory array, allowing for efficient searching and retrieval in constant time, regardless of the number of feature sets. The patent emphasizes the importance of maintaining a fixed size for the feature sets and keys, ensuring consistent performance across various machine learning tasks.
To know more about GlobalData’s detailed insights on GSI Technology, buy the report here.
Data Insights
From
The gold standard of business intelligence.
Blending expert knowledge with cutting-edge technology, GlobalData’s unrivalled proprietary data will enable you to decode what’s happening in your market. You can make better informed decisions and gain a future-proof advantage over your competitors.
GlobalData, the leading provider of industry intelligence, provided the underlying data, research, and analysis used to produce this article.
GlobalData Patent Analytics tracks bibliographic data, legal events data, point in time patent ownerships, and backward and forward citations from global patenting offices. Textual analysis and official patent classifications are used to group patents into key thematic areas and link them to specific companies across the world’s largest industries.
