International business machines corporation (20240289683). SELF-SUPERVISED TERM ENCODING WITH CONFIDENCE ESTIMATION simplified abstract

SELF-SUPERVISED TERM ENCODING WITH CONFIDENCE ESTIMATION

Organization Name

international business machines corporation

Inventor(s)

Francesco Fusco of Zurich (CH)

Diego Matteo Antognini of Ruvigliana (CH)

SELF-SUPERVISED TERM ENCODING WITH CONFIDENCE ESTIMATION - A simplified explanation of the abstract

This abstract first appeared for US patent application 20240289683 titled 'SELF-SUPERVISED TERM ENCODING WITH CONFIDENCE ESTIMATION

Simplified Explanation: This patent application describes a method and computer program product for generating a model with a term encoder by training the model on a dataset to associate terms with embeddings, minimizing distances between embeddings, and predicting confidence scores.

• Training the model on a dataset with terms and embeddings
• Generating second embeddings using a term encoder from word subunits
• Minimizing distances between first and second embeddings
• Predicting confidence scores based on minimized distances
• Deploying the model as an executable algorithm to infer embeddings and confidence scores from input terms

Key Features and Innovation: - Utilizes a term encoder to generate embeddings from word subunits - Trains the model to minimize distances between embeddings - Predicts confidence scores based on minimized distances - Allows users to infer embeddings and confidence scores from input terms

Potential Applications: - Natural language processing - Sentiment analysis - Text classification - Information retrieval

Problems Solved: - Efficiently generating embeddings from input terms - Improving accuracy in predicting confidence scores - Enhancing the performance of executable algorithms

Benefits: - Improved accuracy in inferring embeddings - Enhanced performance in predicting confidence scores - Increased efficiency in natural language processing tasks

Commercial Applications: Potential commercial applications include: - Sentiment analysis tools for businesses - Text classification algorithms for content management systems - Information retrieval systems for search engines

Questions about the Technology: 1. How does the term encoder generate embeddings from word subunits? 2. What are the specific applications of the model in natural language processing tasks?

Frequently Updated Research: Stay updated on advancements in natural language processing techniques for generating embeddings and predicting confidence scores.

Original Abstract Submitted

according to one embodiment, a method and computer program product for generating a model including a term encoder is provided. the embodiment may include training the model on a training dataset that associates training terms with first embeddings of the training terms. the training includes generating, with the term encoder, second embeddings from numerical representations of word subunits of the training terms with an objective of minimizing distances between the first embeddings and the second embeddings. the word subunits form part of a predetermined set of word subunits. the training includes predicting confidence scores based on the minimized distances. the embodiment may include deploying the model as part of an executable algorithm to allow a user to infer third embeddings and corresponding confidence scores from any input terms written based on word subunits of the predetermined set.