Microsoft Technology Licensing, LLC

Abed El Kader Asi

Roy Eisenstadt

Alexander Tsvetkov

Dean Geckt

Royi Ronen

Yarin Kuper

A computerized method categorizes questions from a transcript and provides those categorized questions to targets. A transcript associated with a meeting is obtained and a question in the obtained transcript is detected by a question detection model. A question vector embedding of the detected question is generated using a question clustering model and a category of the detected question is determined by the question clustering model using the generated question vector embedding and a plurality of category vector embeddings. A target of the detected question is identified using the determined category and the detected question is provided to the identified target via a question provision interface. Further, the question clustering model is used to generate the category vector embeddings from example questions of category question catalogs, such that the set of categories for which the question clustering model is configured can be efficiently modified.

Categorizing questions from a transcript using vector embeddings for information retrieval

Example solutions for reducing the likelihood of hallucinations by language models, such as large language models (LLMs) are disclosed. By injecting a sufficient range and quantity of curated factual data into a prompt, the likelihood of a hallucination by an LLM may be reduced. This enables language models to be used in a wider range of settings, in which fabrication of facts is problematic, while reducing the need for a human to carefully check the generated text for accuracy. Examples include: generating a summary of a transcript using a summarization model; extracting topic-specific data from stored data using a scoring model; dynamically generating a language model prompt using the topic-specific data and the summary; and generating an output text using a language model and the language model prompt.

Dynamic construction of large language model prompts

The disclosure herein describes interpreting attention-based decisions of summarization outputs generated by a deep learning model. A decision interpretation model obtains attention values defining connections between input tokens associated with a source text and output tokens for a selected portion of a summary associated with the source text. The input tokens having the highest attention values indicating the strongest connections between the input tokens of the source text and an output token of the summary are selected as primary tokens. A semantic similarity between the primary tokens for each attention head and an output token is calculated. The model selects the primary tokens having the closest semantic similarity with the summary portion. A visual cue is generated on or within a portion of the source text corresponding to the primary tokens. The visual cue identifies dominant words in the source text used to explain the summary portion.

Growing community of inventors

Sammamish, WA, United States of America

Abed El Kader Asi