Salesforce, inc. (20240257803). QUESTION DETECTION MODEL FOR CALL TRANSCRIPT simplified abstract
Contents
QUESTION DETECTION MODEL FOR CALL TRANSCRIPT
Organization Name
Inventor(s)
Yizhak Elboher of Tel Aviv (IL)
Gaia Steinberg of Tel Aviv (IL)
QUESTION DETECTION MODEL FOR CALL TRANSCRIPT - A simplified explanation of the abstract
This abstract first appeared for US patent application 20240257803 titled 'QUESTION DETECTION MODEL FOR CALL TRANSCRIPT
The patent application describes a method for categorizing a sentence as a question using multiple models.
- The models used include an inverse text normalization (ITN) model, a sentence embeddings model, and a term frequency inverse document frequency (TFIDF) model.
- The output of the ITN model is processed using a finite state transducer (FST), while the output of the sentence embeddings model and TFIDF model are processed using logistic regression (LR) models.
- A support vector machine (SVM) is then applied to the output of the FST and LR models to determine whether the sentence is a question.
Potential Applications: - Natural language processing systems - Question-answering systems - Sentiment analysis tools
Problems Solved: - Accurately categorizing sentences as questions - Improving the performance of question detection algorithms
Benefits: - Enhanced accuracy in identifying questions - Increased efficiency in processing natural language queries
Commercial Applications: - Chatbots and virtual assistants - Customer service automation - Search engine optimization tools
Questions about the technology: 1. How does the use of multiple models improve the accuracy of categorizing sentences as questions? 2. What are the potential limitations of this method in real-world applications?
Original Abstract Submitted
disclosed are some implementations of systems, apparatus, methods and computer program products for categorizing a sentence as a question. rather than using a single model, several different models are leveraged to determine whether a sentence is a question. for example, the models can include an inverse text normalization (itn) model, a sentence embeddings model, and a term frequency inverse document frequency (tfidf) model. the output of an itn model is processed using a finite state transducer (fst) while the output of the sentence embeddings model and tfidf model are processed using logistics regression (lr) models. a support vector machine (svm) is then applied to the output of the fst and lr models to determine whether the sentence is a question.