Asapp, Inc.

Tao Lei

Joseph Ellsworth Hackman

Christopher David Fox

Shawn Henry

Satchuthananthavale Rasiah Kuhan Branavan

Kelsey Taylor Ball

Howard Chen

Jonathan David Weese

Lili Yu

Frederick William Poe Heckel

Sida Wang

Alan Nicolas Hampton

Kevin Yang

Mitchell Steven Morgan

Lisa Lijia Jiang

Priya Sawant

Kyle Swanson

Adrien Brochard

Julian Martin Eisenschlos

A message may be suggested to a user participating in a conversation using one or more neural networks where the suggested message is adapted to the preferences or communication style of the user. The suggested message may be adapted to the user with a user embedding vector that represents the preferences or communication style of the user in a vector space. To suggest a message to the user, a conversation feature vector may be computed by processing the text of the conversation with a neural network. A context score may be computed for one or more designated messages, where the context score is computed by processing the user embedding vector, the conversation feature vector, and a designated message feature vector with a neural network. A designated message may be selected as a suggested message for the user using the context scores. The suggestion may then be presented to the user.

Customized message suggestion with user embedding vectors

A company may implement automated workflows for convenience of users or to reduce support costs. For example, allowing a user to change an address using an automated workflow may be faster or less expensive than with a human agent. In some instances, a first communications session may be started between a first user and a second user. During the first communications session, one or more communications may be processed to select an intent of the first user and a value of an information item communicated by the first user. An automated workflow may be selected to continue assisting the first user, and the first user may be transferred to a second communications session with the automated workflow. The automated workflow may be initialized with the value of the information item that was provided during the first communications session so that the first user does not need to repeat information.

Natural language processing for information extraction

A description of a conversation may be generated to allow a person to understand important aspects of the conversation without needing to review the conversation. The conversation description may be generated by identifying one or more events that occurred in the conversation and then generating the description using the identified events. A set of possible events may be determined in advance for a particular application. The events may be identified by using an event neural network for each event. Each event neural network may process the messages of the conversation to generate an event score that indicates a match between the conversation and the corresponding event. The event scores may then be used to select one or more events. Message scores from the event neural network of a selected event may then be used to select one or more messages of the conversation as a rationale for the selected event.

Generating conversation descriptions using neural networks

The amount of time required to train a neural network may be decreased by modifying the neural network to allow for greater parallelization of computations. The computations for cells of the neural network may be modified so that the matrix-vector multiplications of the cell do not depend on a previous cell and thus allowing the matrix-vector computations to be performed outside of the cells. Because the matrix-vector multiplications can be performed outside of the cells, they can be performed in parallel to decrease the computation time required for processing a sequence of training vectors with the neural network. The trained neural network may be applied to a wide variety of applications, such as performing speech recognition, determining a sentiment of text, determining a subject matter of text, answering a question in text, or translating text to another language.

Fast neural network implementations by increasing parallelism of cell computations

A user of an automated natural language system may submit an ambiguous or incomplete request, and interactive techniques may be used to obtain clarification information from the user and then determine a response for presentation to the user. A user's initial request may be processed to compute a category score for each possible category of request. The category scores may be processed to determine if clarification of the request is needed. Where clarification is needed, one or more tags may be selected to determine a clarification question to be presented to the user. For example, a tag clarification score may be computed for each tag that indicates a value of the tag in clarifying the request. After receiving the clarification information from the user, one or more categories may be selected or, where needed, additional clarification information may be obtained.

Clarification of natural language requests using neural networks

Text may be modified according to a specified attribute value, such as changing a 1-star review to a 5-star review. To modify the text, an input sequence of tokens may be obtained corresponding to the text. Mask scores may be computed for the tokens by processing the input sequence of tokens with masking neural network. One or more tokens may be selected using the mask scores, and a masked sequence of tokens may be generated by replacing the selected tokens with a mask token. The masked sequence of tokens may be processed by a language model neural network to select a replacement token for each of the mask tokens. Modified text may then be generated using the selected replacement tokens. The modified text may be used for any appropriate application, such as suggesting messages to users participating in a conversation.

Modifying text according to a specified attribute

A company may implement automated workflows for convenience of users or to reduce support costs. For example, allowing a user to change an address using an automated workflow may be faster or less expensive than with a human agent. Companies may provide support over different types of communications channels with different capabilities, such as voice channels and text channels. Instead of implementing different workflows for different channels, a company may separate aspects of the workflow that are common to different channels from aspects of the workflow that are different for different channels. For example, a workflow may be implemented to determine an action in response to a received communication where the action may be used with multiple channels. The action may then be used to select an action implementation that is specific to a channel.

Automated communications over multiple channels

Initialization of automated workflows

A message may be suggested to a user participating in a conversation using one or more neural networks where the suggested message is adapted to the preferences or communication style of the user. The suggested message may be adapted to the user with a user embedding vector that represents the preferences or communication style of the user in a vector space. To suggest a message to the user, a conversation feature vector may be computed by processing the text the conversation with a neural network. A context score may be computed for one or more designated messages, where the context score is computed by processing the user embedding vector, the conversation feature vector, and a designated message feature vector with a neural network. A designated message may be selected as a suggested message for the user using the context scores. The suggestion may then presented to the user.

Machine learning models may be used during a communications session to process natural language communications and perform actions relating to the communications session. For example, a machine learning model may be used to provide an automated response to a user, to suggest a completion of text being entered by a user, or to provide information about a relevant resource. Machine learning models may rely on machine learning model data that is updated during a communications session as communications are processed by the machine learning model. To improve the performance of a machine learning model when a user leaves a first communications session and enters a second communications session, the machine learning model data may be stored during a first communications session and then retrieved during the second communications session to initialize a machine learning model for the second communications session.

Maintaining machine language model state across communications channels

The present disclosure describes a system, method, and computer program for matching an input file to one of a plurality of datastore files and displaying the rationale for the match. Neural networks are trained to create vector representations of objects in the input file and the datastore files. The cost of each possible pairing of vector representations between the input file and a datastore file is computed, and an optimal transport algorithm is used to identify the vector pairings that result in the lowest total cost of alignment. The datastore file with the lowest total cost of alignment to the input file is identified as the best matching file. The alignment results are used to display the rationale for the match. To constrain the alignment results of the optimal transport algorithm, one or more dummy points and, in certain embodiments, duplicate points are added to one or both of the vector sets to achieve alignments with the desired sparsity patterns.

File matching with rationale alignment using neural networks and an optimal transport algorithm

Text of an incomplete message entered by a user may be processed using a neural network to suggest messages similar to the message the user is in the process of entering. Word embeddings may be obtained for the words of the text that represent the words in a first vector space. The word embeddings may then be processed by the neural network to compute an input message feature vector that represents the incomplete message in a second vector space. The input message feature vector may be used to select a first designated message as a suggestion from a set of designated messages, and the first designated message may be selected using a similarity score computed from the input message feature vector and a first designated message feature vector corresponding to the first designated message. The first designated message may then be presented as a suggestion to the user.

Processing an incomplete message with a neural network to generate suggested messages

A company may desire to maintain a quality level for messages sent by customer service representatives to customers. The company may receive a message input by a customer service representative, modify the message with one or more neural networks, and transmit the modified message to a customer. To modify a message, an input vector may be created for each word of the message where the input vector is created using a word embedding of the word and a feature vector that represents the characters of the word. The input vectors for the words of the message may be sequentially processed with an encoding neural network to compute a message encoding vector that represents the message. The message encoding vector may then be processed by a decoding neural network to sequentially generate the words of a modified message. The modified message may then be transmitted to the customer.

Growing community of inventors

Jersey City, NJ, United States of America

Tao Lei