Google Inc.

Microsoft Technology Licensing, LLC

Waymo LLC

Razvan Mathias

Yaniv Leviathan

Eyal Segalis

Andrew Salamatov

Ofer Ron

Tomer Amiaz

Asaf Aharoni

David Andrew Sterling

Sasha Goldshtein

Giselli Panontini De Souza Sana

Xiaoyong Liu

Jeffrey Korn

Jochen Mathias Eisinger

Johan Engstrom

Alexander Sherman

Emmanuel Christophe

Nicolas Zea

Nicholas Bennett Carter

Joseph Lee

Isaac Supeene

Albert Bachand

Implementations are directed to updating a trained voice bot that is deployed for conducting conversations on behalf of a third-party. A third-party developer can interact with a voice bot development system that enables the third-party developer to train, update, validate, and monitor performance of the trained voice bot. In various implementations, the trained voice bot can be updated by updating a corpus of training instances that was initially utilized to train the voice bot, and updating the trained voice bot based on the updated corpus. In some implementations, the corpus of training instances may be updated in response to identifying occurrence(s) of behavioral error(s) of the trained voice bot while the conversations are being conducted on behalf of the third-party. In additional or alternative implementations, the corpus of training instances may be updated in response to determining the trained voice bot does not include a desired behavior.

Updating trained voice bot(s) utilizing example-based voice bot development techniques

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining the likelihood that a particular event would occur during a navigation interaction using simulations generated by sampling from agent data. In one aspect, a method comprises: identifying an instance of a navigation interaction that includes an autonomous vehicle and agents navigating in an environment; generating multiple simulated interactions corresponding to the instance, comprising, for each simulated interaction: identifying one or more agents; for each identified agent and for each property that characterizes behavior of the identified agent, obtaining a probability distribution for the property; sampling a respective value from each of the probability distributions; and simulating the navigation interaction in accordance with the sampled values; and determining a likelihood that the particular event would occur during the navigation interaction based on whether the particular event occurred during each of the simulated interactions.

Probabilistic simulation sampling from agent data

A computing device provides a first server. The first server receives a primary request associated with a client. In response to receiving the primary request, the first server processes the primary request to generate a response message. As part of processing the primary request, the first server sends a secondary request to a second server after a delay period expires. The second server accesses a resource as part of processing of the secondary request. The duration of the delay period depends at least in part on a recent latency level of the client with regard to the resource. In this way, the first server throttles access to the resource.

Throttling usage of resources

A computer-implemented method for delaying synchronization of a computing device is disclosed according to one aspect of the subject technology. The method comprises receiving an update notification at the computing device, determining whether the computing device is idle, and, if the computing device is idle, then delaying retrieval of synchronization information corresponding to the update notification. The method also comprises determining whether a user has started using the computing device, and, if the user has started using the computing device, then retrieving the synchronization information corresponding to the update notification.

Delay retrieval of synchronization information

A computer-implemented method for deleting synchronization information in a user account is disclosed according to an aspect of the subject technology. The method comprises receiving an instruction from one of a plurality of computing devices to delete synchronization information in the user account, wherein the synchronization information is synchronized across the plurality of computing devices. The method also comprises sending an instruction to each of the other computing devices to disable synchronization, and deleting the synchronization information in the user account.

Efficient clearing of synchronization information

Computer-implemented methods for synchronizing data between a server and a client are provided. In one aspect, a method includes receiving a request from a client to synchronize data. The request includes a synchronization token that includes client recent synchronization signatures, each client recent synchronization signature representing a state of data synchronization on the client for a certain time period. The method also includes comparing each client recent synchronization signature with a corresponding server recent synchronization signature representing data on a server for the same time period. When the comparison indicates that the data on the client for the time period represented by the client recent synchronization signature is not synchronized with corresponding data on the server for the time period, then the method includes sending the data on the server for the time period to the client. Systems and machine-readable media are also provided.

Data synchronization with eventual consistency

Automatically converting data between different schemas. A client request to read data from a front end schema node to a back end schema node is received. The nodes may be constructed according to different schemas, whereby a front end data object and a back end data object are created with a common interface and a common set of property types for wrapping around respective front end and back end data objects. The data is then read from the front end schema node to the back end schema node via the front end data object and the back end data object according to the common interface and the common set of property types. A SchemaState Object provides a listing of property pairs for constructing the front end and back end data objects, each with at least one property from the listing of property pairs.

Converting data between a front end schema node and a back end schema node

Methods, systems, and computer-readable media for synchronizing and sharing of relatively large, modifiable documents with multiple clients utilizing a variety of devices. The method receives instructions from one or more clients regarding a preferable format for use with a respective client device. The method further defines subsections of the shared document based upon semantic breaks within the original document and generates a subsection ID for each subsection. The subsection IDs are sent to the clients who can review the IDs and request to receive one or more subsections on their respective device. The requested subsections are transcoded for use on the respective client devices and sent to the client according to their requests. The clients may then modify the one or more subsections and these modifications are applied to corresponding subsections on other client devices.

Synchronizing modifiable documents with multiple clients using document subsections

Growing community of inventors

Redmond, WA, United States of America

Razvan Mathias