Abstract: A communication server facilitates a video call between client devices of a plurality of participants. During the video call, the communication server identifies an in-call activity to recommend to the plurality of participants in the video call based on user information. The in-call activity is selected to be relevant to each of the plurality of participants and jointly recommended to the plurality of participants of the video call. To identify the recommended in-call activity, the communication server may determine common interests among the plurality of participants and select an in-call activity that is associated with the common interests. After the recommended in-call activity is selected, an indication of the recommended in-call activity is provided to the client devices to enable the client devices to display a user interface including the in-call activity to the participants during the video call.

Abstract: A system and method for creating a machine learning (ML) classifier for a database uses a weakly-supervised training data set created automatically from database items on the basis of a human-created keyword set. The automatically created training data set is used to construct one or more deep learning classifier checkpoints, which can then be compared with one another and with a classifier based on the original keyword set in order to select a classifier for use by other users viewing the database.

Abstract: Mechanisms for implementing a text encoder and text encoder operations are provided. A contrastive machine learning training operation trains an encoder of a machine learning computer model, to learn a sense and similarity preserving embedding that operates to encode input natural language text data to generate encoded natural language text data based on a sense attribute of one or more terms in the input natural language text data. The contrastive machine learning training operation learns to separate positive samples in training data from negative samples in the training data. The trained encoder processes a term specified in an input natural language text to generate an encoded natural language text based on the learned embedding and inputs, to a downstream computing system, the encoded natural language text to cause the downstream computing system to perform a computer natural language processing operation based on the embedding.

Abstract: A system and method for creating a machine learning (ML) classifier for a database uses a weakly-supervised training data set created automatically from database items on the basis of a human-created keyword set. The automatically created training data set is used to construct one or more deep learning classifier checkpoints, which can then be compared with one another and with a classifier based on the original keyword set in order to select a classifier for use by other users viewing the database.

Abstract: A method of registering an attribute in a speech synthesis model, an apparatus of registering an attribute in a speech synthesis model, an electronic device, and a medium are provided, which relate to a field of an artificial intelligence technology such as a deep learning and intelligent speech technology. The method includes: acquiring a plurality of data associated with an attribute to be registered; and registering the attribute in the speech synthesis model by using the plurality of data associated with the attribute, wherein the speech synthesis model is trained in advance by using a training data in a training data set.

Abstract: Systems and methods are provided for automatic evaluation of argument critique essays written by young students in response to prompts. A transformer pre-trained for natural language processing is employed as a machine learning model, which is fine-tune with a first training dataset comprising unannotated argument critique essays written by college students, and then fine-tuned with a second training dataset comprising annotated argument critique essays written by middle school students, where each sentence in the second training dataset is annotated for the presence of valid critiques to prompts. The fine-tuned machine learning model is used to classify each sentence in an essay to be evaluated as either containing a valid critique or not.

Abstract: A sound source tracking method adapted to an ongoing video conference comprising: obtaining a streaming signal of the video conference from an internet; performing a video conference procedure to obtain an audio signal from the streaming signal and send the audio signal to a speaker; performing an audio tracking procedure to obtain the audio signal outputted from the video conference procedure to the communication device and send the audio signal to a sound source tracking camera; playing the audio signal to generate a far-end sound; recording a field sound comprising at least one of the far-end sound and a local-end sound; and performing a comparing procedure to determine a shooting direction of the sound source tracking camera, wherein the shooting direction is adjusted so as not to shoot the speaker when a similarity of the far-end sound and the audio signal is greater than a threshold.

Abstract: Location determination and telephone number distribution for emergency calls is enabled by a telephony system which maintains multiple pools of telephone numbers. Each pool corresponds to a different region such that the pools of telephone numbers are defined at the region-level rather than at the site-level. The telephony system determines the location of a calling device initiating an emergency call regardless of whether the calling device is at a known site. Based on the determined location of the calling device, one of the pools of telephone numbers which corresponds to that location is selected. The telephony system thereafter distributes a telephone number for the calling device to use for the emergency call from that selected pool of telephone numbers to facilitate an emergency call between the calling device and a local public safety answering point.

Abstract: As part of a dialog session between a user and an automated assistant, implementations can process, using a streaming ASR model, a stream of audio data that captures a portion of a spoken utterance to generate ASR output, process, using an NLU model, the ASR output to generate NLU output, and cause, based on the NLU output, a stream of fulfillment data to be generated. Further, implementations can further determine, based on processing the stream of audio data, audio-based characteristics associated with the portion of the spoken utterance captured in the stream of audio data. Based on the audio-based characteristics and/the stream of NLU output, implementations can determine whether the user has paused in providing the spoken utterance or has completed providing of the spoken utterance. If the user has paused, implementations can cause natural conversation output to be provided for presentation to the user.

Abstract: One example method includes receiving, by a computing device, audio during a video conference having a plurality of participants, the audio comprising spoken words by a user of the computing device; recognizing one or more words from the spoken words; identifying one or more keywords within the one or more recognized words; accessing a set of rules comprising one or more rules, each rule of the one or more rules associated with an application of a set of applications, and at least one rule of the one or more rules associated with a functionality of a respective application; determining a context associated with the one or more keywords; determining an application to execute based on the one or more keywords, the context, and the one or more rules, wherein determining the application comprises determining a functionality of the application to invoke; and in response to receiving user confirmation of the functionality of the application to invoke, executing the application and invoking the functionality.

Abstract: A system is provided for determining subscription data when a user requests to receive an output in the future when an event occurs. The system may determine an output type based on the capabilities of the output device and a trigger type. The system may determine a trigger type based on the priority of the triggering event. The system may also determine how many times the subscription is to be executed. Using this information, the system creates the subscription so that the user may receive a notification or an announcement when an event occurs.

Abstract: Systems and methods for providing an online to offline service in response to a voice request from a user terminal are provided. A method includes: receiving a voice request from a user terminal; in response to the voice request, updating a customized recognition model trained using data of a plurality of points of interest associated with the user terminal; obtaining a general recognition model trained using data from general public; determining a literal destination associated with the voice request based at least on the voice request, the customized recognition model and the general recognition model.

Abstract: Embodiments of this disclosure disclose an information processing method, apparatus and a non-transitory computer readable medium. The method includes: obtaining a target text sequence corresponding to to-be-processed text information; obtaining a context vector according to the target text sequence; determining a logical similarity corresponding to the target text sequence according to the context vector and the target text sequence; and encoding the target text sequence corresponding to target text information by using the logical similarity to obtain a text encoding result. In this embodiment of this disclosure, a context vector related to a discrete sequence is used to encode the discrete sequence, to strengthen the dependence between elements in the discrete sequence, thereby enhancing the performance of a neural network model and improving the learning capability of the model.

Abstract: Transcribed text and physiological data of a remote video conference participant are transmitted to a local device separately from the video data, which depicts the remote party during a time interval. An image of the video data is captured at a time instant within the time interval. A value of a remote party feature is determined remotely using the video data. The remote party feature can be the remote party's heart rate at the time instant. The value of the feature is received onto the local device. Audio data captures sounds spoken by the remote party and is converted by the remote device into words of text. The audio data converted into a particular word was captured at the time instant. The particular word is received onto the local device. The particular word and the value of the feature are displayed in association with one another on the local device.

Abstract: Embodiments of the disclosure provide an information processing method, an information processing apparatus, and a storage medium. The method includes: obtaining source data; encoding sub-data in the source data based on a target word feature vector to obtain hidden feature vectors corresponding to the sub-data, the target word feature vector representing a sentiment feature standard; obtaining a word feature vector corresponding to the source data based on the hidden feature vectors corresponding to the sub-data; and inputting the word feature vector into a preset sentiment classification network to obtain a result of sentiment polarity prediction of the source data. According to the embodiments of the disclosure, the accuracy of sentiment polarity prediction may be improved.

Abstract: Methods, systems, apparatuses, and computer-readable media are provided for managing extended reality video conferences. In one implementation, the computer-readable medium may include instructions to cause a processor to: receive a request to initiate a video conference between a plurality of participants; receive image data captured by at least one image sensor associated with a wearable extended reality appliance, the image data reflecting a layout of a physical environment in which the wearable extended reality appliance is located; analyze the image data to identify at least one interference region in the physical environment; receive visual representations of the plurality of participants; and cause the wearable extended reality appliance to display the visual representations of the plurality of participants at multiple distinct locations other than in the at least one interference region.

Abstract: Example embodiments describe means (200) for performing i) pre-compensating (210, N sets of K1 tone data values (220) for crosstalk between N communication lines; the N sets of K1 tone data values pertaining to respective N terminal nodes of a digital communication system; ii) calculating (215) from the pre-compensated N sets of K1 tone data values (221) N sets of first time domain symbols (225); iii) calculating (283) a second time domain symbol (284) from a set of K2 tones values (280); the K2 tone data values pertaining to a selected one of the N terminal nodes; and iv) adding (212) the second time domain symbol in a weighted manner to the first time domain symbols such that the second time domain symbol is added to the first time domain symbol for the selected terminal node and to at least one other of the first time domain symbols for the respective other terminal nodes.

Abstract: Methods and systems are disclosed for enabling the generation of a token corresponding to a tone generated by a telephony system, comprising receiving one or more dual tone multi-frequency (DTMF) tones generated by a telephony system, generating a token based on the one or more DTMF tones; and transmitting the generated token to a merchant system.

Abstract: Implementations relate to enabling of authorization of certain automated assistant functions via one or more modalities available within a vehicle. Implementations can eliminate wasting of computational and communication resources by at least allowing other users to authorize execution of certain input commands from a user, without requesting the user to re-submit the commands. The vehicle can include a computing device that provides access to restricted data, which can be accessed in order for an action to be performed by the automated assistant. However, when a restricted user requests that the automated assistant perform an action involving accessing the restricted data, the automated assistant can be authorized or unauthorized to proceed with fulfilling the request via a modality controlled by an unrestricted user.

Abstract: The subject technology provides a video conferencing application in which a live incoming or outgoing video stream can be supplemented with supplemental content, such as stickers, animations, etc., from within the video conferencing application. In this manner, a user participating in a video conferencing session with a remote user can add stickers, animations, and/or adaptive content to an outgoing video stream being captured by the device of the user, or to an incoming video stream from the device of the remote user, without having to locally cache/store a video clip before editing, and without having to leave the video conferencing session (or the video conferencing application) to access a video editing application.