Amazon Technologies, Inc.

Karthik Ramakrishnan

Michael Martin George

Rohit Prasad

Hugh Evan Secker-Walker

Stan Weidner Salvador

Andrew Smith

Christopher Schindler

Rafal Kuklinski

Bjorn Hoffmeister

Nikko Strom

Ryan Paul Thomas

Vivek Yadav

Premkumar Natarajan

Yue Wu

Ivan Borges Oliveira

Rajiv Reddy

Philip Mark Kaminsky

Ayush Jaiswal

Manik Kumar

Phillip Oliver Kriett

Akhand Pratap Singh

Georgios Patsakis

Shengbin Yang

Andrew V Goldberg

Gourav Datta

Quico Pepijn Spaen

Diwakar Tiwari

Prateek Singhal

A system configured to perform style-aware listener animation. By representing different listening styles (e.g., facial expressions) using an embedding space, a single model can be trained to generate unique facial animations for a number of distinct listeners. Thus, individual listening styles can be associated with a listener identifier, enabling the system to (i) animate a plurality of different listeners with unique nonverbal behavior and/or (ii) select a particular listener identifier or desired type of listener style with which to animate. This enables the model to be generalized to new listeners to generate additional listener facial responses without needing training data for each new listener. The model may process a listener representation style or listener identifier, along with input data corresponding to a speaker talking, to generate unique facial animation responsive to the speech.

Listener animation

One challenge for middle-mile route planning is that the set of loads changes significantly between daily planning and execution. Systems and methods are provided for optimizing a transportation plan for a transportation network based on these load changes. The disclosed system re-optimizes a solution by starting from a previously existing plan and previously generated columns (e.g., candidate routes). The disclosed techniques significantly improve the compute time of the system to generate transportation plans that are optimized accordingly to an optimization parameter. The system takes into account the current execution status associated with a given entry of the plan to determine whether the entry should be re-optimized. Entries corresponding to tours that have already commenced, may be at least partially ignored for re-optimization consideration. The disclosed techniques enable state-aware, adaptive re-optimization for even tours that are in-progress or have been tendered.

Techniques for reactive route planning

Techniques for ensuring content output to a user conforms to a quality of the user's speech, even when a speechlet or skill ignores the speech's quality, are described. When a system receives speech, the system determines an indicator of the speech's quality (e.g., whispered, shouted, fast, slow, etc.) and persists the indicator in memory. When the system receives output content from a speechlet or skill, the system checks whether the output content is in conformity with the speech quality indicator. If the content conforms to the speech quality indicator, the system may cause the content to be output to the user without further manipulation. But, if the content does not conform to the speech quality indicator, the system may manipulate the content to render it in conformity with the speech quality indicator and output the manipulated content to the user.

Content output management based on speech quality

Features are disclosed for selecting and using multiple transforms associated with a particular remote device for use in automatic speech recognition ('ASR'). Each transform may be based on statistics that have been generated from processing utterances that share some characteristic (e.g., acoustic characteristics, time frame within which the utterances where processed, etc.). When an utterance is received from the remote device, a particular transform or set of transforms may be selected for use in speech processing based on data obtained from the remote device, speech processing of a portion of the utterance, speech processing of prior utterances, etc. The transform or transforms used in processing the utterances may then be updated based on the results of the speech processing.

Generation and use of multiple speech processing transforms

Speech recognition may be improved using data derived from an utterance. In some embodiments, audio data is received by a user device. Adaptation data may be retrieved from a data store accessible by the user device. The audio data and the adaptation data may be transmitted to a server device. The server device may use the audio data to calculate second adaptation data. The second adaptation data may be transmitted to the user device. Synchronously or asynchronously, the server device may perform speech recognition using the audio data and the second adaptation data and transmit speech recognition results back to the user device.

Using adaptation data with cloud-based speech recognition

In automatic speech recognition, certain parsing information, such as rules and tags, may be embedded into a finite state transducer (FST) to produce FST output that includes speech recognition results along with codes indicating parsing results of the recognized speech. The codes in the FST output may be formatted using a markup language, such as XML or JSON, for processing by a later application. The FST may be constructed according to a grammar defining the parsing information. The codes for inclusion in the FST output may be embedded into arcs of the FST and then included in the FST output when the speech recognition engine traverses the arcs of the FST.

Parse information encoding in a finite state transducer

Growing community of inventors

Bellevue, WA, United States of America

Karthik Ramakrishnan