公开(公告)号：US12265528B1

公开(公告)日：2025-04-01

申请号：US18187553

申请日：2023-03-21

Applicant: Amazon Technologies, Inc.

Inventor： Wuwei Lan ,  Patrick Ng ,  Zhiguo Wang ,  Ramesh M. Nallapati ,  Henghui Zhu ,  Anuj Chauhan ,  Sudipta Sengupta ,  Stephen Michael Ash ,  Bing Xiang ,  Gregory David Adams

IPC: G06F16/00 ,  G06F16/22 ,  G06F16/242 ,  G06F16/2457 ,  G06F16/248 ,  G06F16/25 ,  G06N3/0455 ,  G06N3/0499

Abstract: Techniques for handling natural language query processing are described. In some examples, a sequence-to-sequence model is used to handle a natural language query. Post-processing of a result of the sequence-to-sequence model utilizes fine-grained information from an entity linker. In some examples, the sequence-to-sequence model and aspects of a natural language query pipeline are used to handle a natural language query.

公开(公告)号：US11526557B2

公开(公告)日：2022-12-13

申请号：US16697979

申请日：2019-11-27

Applicant: Amazon Technologies, Inc.

Inventor： Zhiguo Wang ,  Zhiheng Huang ,  Ramesh M. Nallapati ,  Bing Xiang

IPC: G06F16/9038 ,  G06F16/908 ,  G06F16/93 ,  G06N20/00

Abstract: Techniques for displaying a search are described. An exemplary method includes receiving a search query, performing the search query on a plurality of documents, the documents including text passages, to generate a search query result, determining an aspect of the search query result that has a confidence value that exceeds a first confidence threshold with respect to its relevance to the search query; and, displaying the search result including an emphasis on the aspect of the result exceeds the first confidence threshold.

公开(公告)号：US11475067B2

公开(公告)日：2022-10-18

申请号：US16698080

申请日：2019-11-27

Applicant: Amazon Technologies, Inc.

Inventor： Cicero Nogueira Dos Santos ,  Xiaofei Ma ,  Peng Xu ,  Ramesh M. Nallapati ,  Bing Xiang ,  Sudipta Sengupta ,  Zhiguo Wang ,  Patrick Ng

IPC: G06F40/30 ,  G06F16/9032 ,  G06K9/62 ,  G06F16/9038 ,  G06N20/00 ,  G06F16/903 ,  G06F16/93 ,  G06F40/20

Abstract: Techniques for generation of synthetic queries from customer data for training of document querying machine learning (ML) models as a service are described. A service may receive one or more documents from a user, generate a set of question and answer pairs from the one or more documents from the user using a machine learning model trained to predict a question from an answer, and store the set of question and answer pairs generated from the one or more documents from the user. The question and answer pairs may be used to train another machine learning model, for example, a document ranking model, a passage ranking model, a question/answer model, or a frequently asked question (FAQ) model.

公开(公告)号：US12259914B1

公开(公告)日：2025-03-25

申请号：US17240724

申请日：2021-04-26

Applicant: Amazon Technologies, Inc.

Inventor： Yifan Gao ,  Henghui Zhu ,  Ramesh M. Nallapati ,  Patrick Ng ,  Cicero Nogueira Dos Santos ,  Zhiguo Wang ,  Feng Nan ,  Dejiao Zhang ,  Andrew Oliver Arnold ,  Bing Xiang

IPC: G06F16/332 ,  G06F16/3329 ,  G06F40/20 ,  G06N3/045 ,  G06N5/04 ,  G06N20/00

Abstract: Techniques for predicting an answer to a question using a machine learning model are described. In some examples, the model predicts one or more answers to the question by: predicting at least two answers to the question using a first component of the question-answer model from a set of passages, generating, using a second component of the question-answer model, at least one question for each of the predicted at least two answers, and performing roundtrip predictions until each generated question only has one answer.

公开(公告)号：US12062368B1

公开(公告)日：2024-08-13

申请号：US17039613

申请日：2020-09-30

Applicant: Amazon Technologies, Inc.

Inventor： Anuroop Arora ,  Atul Deo ,  Ramesh M. Nallapati ,  Henghui Zhu ,  Arvind Arikatla ,  Sai Bharadwaj Kanduri ,  Srikanth Prabala ,  Dejiao Zhang

IPC: G10L15/18 ,  G06N20/00 ,  G06Q30/01 ,  G10L15/06 ,  G10L15/08 ,  G10L15/26

CPC classification number: G10L15/1822 ,  G06N20/00 ,  G10L15/063 ,  G10L15/26 ,  G06Q30/01 ,  G10L2015/0631 ,  G10L2015/088

Abstract: Systems and methods to detect themes in contacts data. Contacts data may be encoded as text (e.g., chat logs), audio (e.g., audio recordings), and various other modalities. Text-based transcripts of contacts data may be parsed into turns, an issue turn may be detected using a machine learning model, a key phrase may be extracted from the issue turn. Key phrases from across multiple contacts data may be clustered to identify themes.

公开(公告)号：US20230418567A1

公开(公告)日：2023-12-28

申请号：US17847115

申请日：2022-06-22

Applicant: Amazon Technologies, Inc.

Inventor： Praphruetpong Athiwaratkun ,  Yuchen Tian ,  Mingyue Shang ,  Zijian Wang ,  Ramesh M. Nallapati ,  Parminder Bhatia ,  Andrew Oliver Arnold ,  Bing Xiang ,  Sudipta Sengupta ,  Yanitsa Donchev ,  Srinivas Iragavarapu ,  Matthew Lee ,  Vamshidhar Krishnamurthy Dantu ,  Atul Deo ,  Ankur Deepak Desai

IPC: G06F8/33

CPC classification number: G06F8/33

Abstract: Pre-fix matching may constrain the generation of next token predictions. Input text to perform a next token prediction may be received. Multiple tokens may be determined from the input text, including a partial token. From possible tokens, one or more matching possible tokens with the partial token may be identified. Next token predictions may then be filtered using the identified possible tokens in order to ensure that the partial token is matched.

公开(公告)号：US11366855B2

公开(公告)日：2022-06-21

申请号：US16697948

申请日：2019-11-27

Applicant: Amazon Technologies, Inc.

Inventor： Jean-Pierre Dodel ,  Zhiheng Huang ,  Xiaofei Ma ,  Ramesh M. Nallapati ,  Krishnakumar Rajagopalan ,  Milan Saini ,  Sudipta Sengupta ,  Saurabh Kumar Singh ,  Dimitrios Soulios ,  Ankit Sultania ,  Dong Wang ,  Zhiguo Wang ,  Bing Xiang ,  Peng Xu ,  Yong Yuan

IPC: G06F16/00 ,  G06F16/901 ,  G06N3/04 ,  G06F16/2457 ,  G06F16/903

Abstract: Techniques for searching documents are described. An exemplary method includes receiving a document search query; querying at least one index based upon the document search query to identify matching data; fetching the identified matched data; determining one or more of a top ranked passage and top ranked documents from the set of documents based upon one or more invocations of one or more machine learning models based at least on the fetched identified matched data and the document search query; and returning one or more of the top ranked passage and the proper subset of documents.