Amazon Technologies, Inc.

Ganesh Kumar Gella

Rama Krishna Sandeep Pokkunuri

Robert William Serr

Harshal Pimpalkhute

Christopher Pitstick

Yonnas Getahun Beyene

Pratik Prakash Shah

Daniel Paul Kirschner

Venkata Abhinav Sidharth Bhagavatula

Stefano Stefani

John Baker

Yahor Pushkin

Roger Scott Jenke

Kalpesh N Sutaria

Vasanth Philomin

Enrico Sartorello

Nikhil Kandoi

Atacan Conduroglu

Jordan Carl Cherry

Santosh Kumar Ameti

Swaminathan Sivasubramanian

Yi Zhang

Sravan Babu Bodapati

Katrin Kirchhoff

Yaser Al-Onaizan

Kosar Akram Jaff

Swapandeep Singh

Monica Lakshmi Sunkara

Ashish Vishwanath Shenoy

Meghana Puvvadi

Sudhakar Rao Puvvadi

Anubhav Mishra

Manikya Pavan Kiran Pothukuchi

Saab Mansour

Gregory Paul Niejadlik

Salvatore Romeo

Kartik Natarajan

Ramiah Kandasamy

Tania Khattar

Navneet Sabbineni

Cheng Ran Li

Satya Naga Satis Kumar Gunuputi Alluri Venka

Ruhaab Markas

Shahab Shekari

Sara Elisabeth Borowski

Jessica Lauren Fracassini

Jason Douglas Trichel

Jeffrey David Backer

Craig Joseph Mouser

Tengfei Li

Techniques for at least the generation of a chatbot built from a custom vocabulary and to use runtime hints during inference are described. In some examples, the generation of the chatbot includes receiving a request to build a chatbot using a bot definition and a custom vocabulary, wherein the chatbot is to use runtime hints during usage; building the chatbot from the bot definition and custom vocabulary by at least: generating automatic speech recognition (ASR) artifacts to be used in decoding audio input into the chatbot into text for at least one other component of the chatbot to use in determining a next act to be performed, the ASR artifacts including artifacts that use the custom vocabulary and artifacts that do not use the custom vocabulary, and storing the ASR artifacts.

Dynamically selectable automated speech recognition using a custom vocabulary

A system receives one or more transcripts of communications between entities. The system identifies a requested action in the communications based at least in part on a mapping between the requested action and an application programming interface. The system identifies one or more statements eliciting information, based on parameters to the application programming interface. The system generates a definition of an artificial agent based, at least in part, on the requested action and the one more statements eliciting information.

Artificial agent generator

Using a first set of machine learning models, a communication from a user of a restaurant is analyzed at an order coordinator linked via a network to resources of an order management service at a provider network. A response to the communication is prepared using another set of models at the provider network and presented to the user. An order of the user for one or more restaurant menu items is fulfilled, based at least partly on analysis of a second communication received from the user after the response is presented.

Distributed system for automated restaurant order acquisition

Based on analysis of a workload associated with a throttling key of a client request directed to a first service, a scale-out requirement of the throttling key is obtained at respective resource managers of a plurality of other services which are utilized by the first service to respond to client requests. The resource managers initiate, asynchronously with respect to one another, resource provisioning tasks at each of the other services to fulfill the scale-out requirement. A throttling limit associated with the throttling key is updated to a second throttling key after the resource provisioning tasks are completed by the resource managers, and the updated limit is used to determine whether to accept another client request associated with the throttling key.

Automated provisioning techniques for distributed applications with independent resource management at constituent services

Techniques for hosting adding and warming a host are described. In some instances, a method of determining that at least one group of hosts is to be increased by adding an additional host to the group of hosts; sending a request to the group of hosts for a list of machine learning models loaded per host of the group of hosts; receiving, from each host, the list of loaded machine learning models; loading at least a proper subset of list of loaded machine learning models into random access memory of the at least one group; receiving a request to perform an inference; routing the request to the additional host of the group of hosts; performing an inference using the additional host of the group of hosts; and providing a result of the inference to an external entity is described.

Pre-warming scheme to load machine learning models

Methods and systems for adding functionality to an account of a language processing system where the functionality is associated with a second account of a first application system is described herein. In a non-limiting embodiment, an individual may log into a first account of a language processing system and log into a second account of a first application system. While logged into both the first account and the second account, a button included within a webpage provided by the first application may be invoked. A request capable of being serviced using the first functionality may be received by the language processing system from a device associated with the first account. The language processing system may send first account data and the second account data to the first application system to facilitate an action associated with the request, thereby enabling the first functionality for the first account.

Third party account linking for voice user interface

Techniques for hosting machine learning models are described. In some instances, a method of receiving a request to perform an inference using a particular machine learning model; determining a group of hosts to route the request to, the group of hosts to host a plurality of machine learning models including the particular machine learning model; determining a path to the determined group of hosts; determining a particular host of the group of hosts to perform an analysis of the request based on the determined path, the particular host having the particular machine learning model in memory; routing the request to the particular host of the group of hosts; performing inference on the request using the particular host; and providing a result of the inference to a requester is performed.

Hosting machine learning models

Techniques for obtaining information about an artwork and presenting such information may be provided. For example, a visual representation of the artwork, such as an image, may be received and analyzed. Results of the analysis may be categorized into a type and attributes of the artwork and may be presented to a first user for confirmation. Once a confirmation is received, the categorized type and attributes and other information may be presented to a second user searching for the artwork.

Smart categorization of artwork

The technology may provide display of an electronic representation of physical art in a virtual environment so that a customer may visualize how the physical art would appear in a physical environment. Based on a characteristic of the physical art, a virtual environment in which to display an electronic representation of the physical art may be selected. The electronic representation of the physical art may then be included in the virtual environment for display to a customer. The customer may then electronically purchase the physical art via the virtual environment.

Display of an electronic representation of a physical object in a virtual environment

Techniques for providing an image download protection may be provided. In particular, an original image may be protected from being downloaded by providing a substitute image in its place, identified by a plurality of layers that are configured to look substantially similar to the original image. However, when a user device attempts to quickly store the image (e.g., by right-clicking on the image and selecting 'save'), the user device can be hindered from downloading the substantially similar looking image because only one layer may be stored on the user device. In some examples, the layer of the image looks substantially distinct from the original image, but the layer in combination with a plurality of other layers can look substantially similar to the original image.

Image download protection

Methods and systems for selecting and shipping an art sample that will degrade after a pre-defined period of time are provided. A user can request an art sample while browsing for items associated with an electronic marketplace. The art sample, which can be a lower quality replica of an original art piece, may be printed utilizing a printing technique which results in the art sample degrading after a pre-defined period of time. The art sample may then be shipped to the user to aid in determining whether to purchase the original art piece by utilizing the art sample to visualize the potential purchase in the user's own viewing space.

Fine art samples

Growing community of inventors

Redmond, WA, United States of America

Ganesh Kumar Gella