IBM has patented a method and system for training a spoken language understanding (SLU) model without the need for a transcript. The system extracts entity and intent labels from speech recordings to train the model. The computing device includes a processor, network interface, and engine for training the SLU model. GlobalData’s report on International Business Machines gives a 360-degree view of the company including its patenting strategy. Buy the report here.
Smarter leaders trust GlobalData
According to GlobalData’s company profile on International Business Machines, M2M communication interfaces was a key innovation area identified from patents. International Business Machines's grant share as of February 2024 was 75%. Grant share is based on the ratio of number of grants to total number of patents.
Training spoken language understanding model without speech recording transcript
A recently granted patent (Publication Number: US11929062B2) discloses a computing device equipped with a processor, network interface, and an engine designed for training a spoken language understanding (SLU) model. The device receives natural language training data over a network, including speech recordings and semantic entities or intents associated with each recording. It then extracts entity labels, values, and intent labels from the data to train the SLU model without requiring a transcript of the speech recordings. The patent emphasizes the innovative approach of training the model based on semantic entities and intents, streamlining the training process for improved language understanding.
Furthermore, the patent details additional features of the computing device, such as the arrangement of semantic entities, pre-processing alignment, and utilization of different types of training data. The training data can be derived from user-administrator transaction records, enhancing the model's ability to comprehend real-world interactions. The patent also highlights the use of neural network processing, including techniques like connectionist temporal classification (CTC) and recurrent neural network transducer (RNN-T), for efficient extraction of entity and intent labels. Additionally, the patent describes the incorporation of transfer learning by initializing the SLU model with an automatic speech recognition (ASR) model. This method aims to enhance the model's performance during an active phase, where it can accurately interpret raw spoken language data without the need for transcripts, providing a comprehensive understanding of intents and semantic entities within the spoken language.
To know more about GlobalData’s detailed insights on International Business Machines, buy the report here.
Premium 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.
