Amazon Technologies, Inc. patent applications on December 26th, 2024

Patent Applications by Amazon Technologies, Inc. on December 26th, 2024

Amazon Technologies, Inc.: 8 patent applications

Amazon Technologies, Inc. has applied for patents in the areas of G10L15/18 (2), G10L15/26 (2), G06F40/35 (1), H04L67/1014 (1), H04L61/4511 (1) G06F40/35 (1), G06N3/098 (1), G06V30/19147 (1), G10L13/033 (1), G10L15/18 (1)

With keywords such as: user, language, data, synthetic, voice, content, techniques, generated, training, and information in patent application abstracts.

Patent Applications by Amazon Technologies, Inc.

20240428002. MEDICAL CONVERSATION SUMMARIZATION STYLE INTELLIGENCE_simplified_abstract_(amazon technologies, inc.)

Inventor(s): Aparna Elangovan of Seattle WA (US) for amazon technologies, inc., Lei Xu of Jersey City NJ (US) for amazon technologies, inc., Devang Kulshreshtha of Montreal (CA) for amazon technologies, inc., Sravan Babu Bodapati of Fremont CA (US) for amazon technologies, inc., Katrin Kirchhoff of Seattle WA (US) for amazon technologies, inc., Sarthak Handa of Seattle WA (US) for amazon technologies, inc.

IPC Code(s): G06F40/35, G06F40/40

CPC Code(s): G06F40/35

Abstract: a medical audio summarization service receives a medical conversation and an indication of a user preferred summarization style selected from a plurality of available summarization styles to generate a medical summary that conforms to the user preferred summarization style. a transcript is generated via a medical audio transcription service, and the transcript is used by a natural language processing engine (including a large language model) to generate the medical summary. the large language model is trained to be used to generate medical summaries that conform to respective ones of a plurality of user preferred summarization styles. the large language model is trained using training data comprising previously generated summaries and summary interaction metadata generated from user edits and/or feedback.

20240428082. EFFICIENT RECOVERY FROM FAILURES DURING DISTRIBUTED TRAINING OF MACHINE LEARNING MODELS_simplified_abstract_(amazon technologies, inc.)

Inventor(s): Zhuang Wang of Kirkland WA (US) for amazon technologies, inc., Zhen Jia of San Jose CA (US) for amazon technologies, inc., Shuai Zheng of Santa Clara CA (US) for amazon technologies, inc., Zhen Zhang of Santa Clara CA (US) for amazon technologies, inc., Xinwei Fu of Fremont CA (US) for amazon technologies, inc., Yida Wang of Palo Alto CA (US) for amazon technologies, inc.

IPC Code(s): G06N3/098

CPC Code(s): G06N3/098

Abstract: a placement plan for training state checkpoints of a machine learning model is generated based at least in part on a number of training servers of a distributed training environment. the plan indicates, with respect to an individual server, one or more other servers at which replicas of training state checkpoints of the individual server are to be stored. during selected periods of one or more training iterations of the model, respective portions of a replica of a training state checkpoint of a first server are transmitted to a second server selected based on the placement plan. after an event causes disruption of the training iterations, one of the checkpoints generated at the first server is retrieved from the second server and used to resume the training iterations.

20240428605. PRIVACY-PRESERVING TRAINING AND EVALUATION OF COMPUTER VISION MODELS_simplified_abstract_(amazon technologies, inc.)

Inventor(s): Kun Li of Sammamish WA (US) for amazon technologies, inc., Pragyana K. Mishra of Seattle WA (US) for amazon technologies, inc.

IPC Code(s): G06V30/19, G06V20/62, G06V20/70, G06V30/14, G06V30/148, G06V30/30

CPC Code(s): G06V30/19147

Abstract: devices and techniques are generally described for privacy preservation for computer vision models. in some examples, a first field of text and a second field of text may be detected in a first image. a first alpha-numeric text string may be detected in the first field and a second alpha-numeric text string may be detected in the second field. a first sub-image including the first alpha-numeric text string may be generated and a second sub-image including the second alpha-numeric text string may be generated. the first sub-image may be sent to a first computing device for annotation and the second sub-image may be sent to a second computing device for annotation.

20240428775. USER-CUSTOMIZED SYNTHETIC VOICE_simplified_abstract_(amazon technologies, inc.)

Inventor(s): Sebastian Dariusz Cygert of Gdansk (PL) for amazon technologies, inc., Daniel Korzekwa of Gdansk (PL) for amazon technologies, inc., Kamil Pokora of Gdansk (PL) for amazon technologies, inc., Piotr Tadeusz Bilinski of Warsaw (PL) for amazon technologies, inc., Kayoko Yanagisawa of Cambridge (GB) for amazon technologies, inc., Abdelhamid Ezzerg of Cambridge (GB) for amazon technologies, inc., Thomas Edward Merritt of Downham Market (GB) for amazon technologies, inc., Raghu Ram Sreepada Srinivas of Snohomish WA (US) for amazon technologies, inc., Nikhil Sharma of Kirkland WA (US) for amazon technologies, inc.

IPC Code(s): G10L13/033, G10L13/047, G10L13/10

CPC Code(s): G10L13/033

Abstract: techniques for generating customized synthetic voices personalized to a user, based on user-provided feedback, are described. a system may determine embedding data representing a user-provided description of a desired synthetic voice and profile data associated with the user, and generate synthetic voice embedding data using synthetic voice embedding data corresponding a profile associated with a user determined to be similar to the current user. based on user-provided feedback with respect to a customized synthetic voice, generated using synthetic voice characteristics corresponding to the synthetic voice embedding data and presented to the user, and the synthetic voice embedding data, the system may generate new synthetic voice embedding data, corresponding to a new customized synthetic voice. the system may be configured to assign the customized synthetic voice to the user, such that a subsequent user may not be presented with the same customized synthetic voice.

20240428783. GENERATIVE LANGUAGE MODELS_simplified_abstract_(amazon technologies, inc.)

Inventor(s): Rahul Gupta of Waltham MA (US) for amazon technologies, inc., Charith Peris of Arlington MA (US) for amazon technologies, inc., Palash Goyal of San Jose CA (US) for amazon technologies, inc., Lisa Bauer of New York NY (US) for amazon technologies, inc., Ninareh Mehrabi of Glendale CA (US) for amazon technologies, inc.

IPC Code(s): G10L15/18, G10L15/26

CPC Code(s): G10L15/18

Abstract: systems and techniques for moderating responses of a generative language model are described herein. some user inputs to a generative language model may include biases, misinformation, and other references to moderated content. to prevent the generative language model from generating responses that promote these forms of moderated content, the techniques described determine a policy corresponding to the determined moderated content category of the user input. the determined policy may correspond to a template of instructions for how the generative language model is to respond to such moderated content. the output of the generative language model may also be moderated before being presented to the user.

20240428787. GENERATING MODEL OUTPUT USING A KNOWLEDGE GRAPH_simplified_abstract_(amazon technologies, inc.)

Inventor(s): Mahdi Namazifar of Oakland CA (US) for amazon technologies, inc., Di Jin of Santa Clara CA (US) for amazon technologies, inc., Yang Liu of Los Altos CA (US) for amazon technologies, inc., Devamanyu Hazarika of Sunnyvale CA (US) for amazon technologies, inc., Dilek Hakkani-Tur of Los Altos CA (US) for amazon technologies, inc., Yubin Ge of Champaign IL (US) for amazon technologies, inc.

IPC Code(s): G10L15/22, G06F40/295, G10L15/183

CPC Code(s): G10L15/22

Abstract: techniques for constraining the results of a generative language model to valid information using knowledge-grounded documentation. a generative language model may generate invalid results, including compound entities and incorrect entity relations. the techniques include, for a given user inquiry, determining a set of documented information, from a particular knowledge base, that corresponds to the user inquiry. the techniques further include determining a subgraph from a knowledge graph representing the knowledge base, as well as determining a trie data structure representation of the set of documented information. the user inquiry and subgraph are provided as input to a trained generative language model for generating a response to the user inquiry. the techniques include using the trie data structure to validate that the generated response corresponds to real information from the set of documented information.

20240428797. SPEECH PROCESSING_simplified_abstract_(amazon technologies, inc.)

Inventor(s): Beiye Liu of Millwood NY (US) for amazon technologies, inc., Wael Hamza of Wylie TX (US) for amazon technologies, inc., Liwei Cai of Cambridge MA (US) for amazon technologies, inc., Konstantine Arkoudas of North Bergen NJ (US) for amazon technologies, inc., Chengwei Su of Lexington MA (US) for amazon technologies, inc., Subendhu Rongali of New York NY (US) for amazon technologies, inc.

IPC Code(s): G10L15/26, G10L15/18

CPC Code(s): G10L15/26

Abstract: techniques for performing spoken language understanding (slu) processing are described. an slu component may include an audio encoder configured to perform an audio-to-text processing task and an audio-to-nlu processing task. the slu component may also include a joint decoder configured to perform the audio-to-text processing task, the audio-to-nlu processing task and a text-to-nlu processing task. input audio data, representing a spoken input, is processed by the audio encoder and the joint decoder to determine nlu data corresponding to the spoken input.

20240430341. PROCESSING DNS QUERIES TO IDENTIFY PRE-PROCESSING INFORMATION_simplified_abstract_(amazon technologies, inc.)

Inventor(s): Ryan F. Watson of Seattle WA (US) for amazon technologies, inc.

IPC Code(s): H04L67/5681, G06F16/955, H04L61/4511, H04L67/1014, H04L67/563

CPC Code(s): H04L67/5681

Abstract: systems and methods for processing a dns query to identify and implement pre-processing information by a dns server component in anticipation of a corresponding content request from a client computing device are provided. the pre-processing information can correspond to identification of content to be preloaded or other actions to be implemented by one or more computing devices in association with an anticipated client content request. based on identification of the content or future actions, a dns server component can provide the pre-processing information to one or more computing devices, such as computing devices of a cdn service provider and/or an original content provider, in advance of a corresponding request for content from the client computing device in order to improve performance associated with responding to the client request.

Amazon Technologies, Inc. patent applications on December 26th, 2024