Twitter, Inc.

Arvind Thiagarajan

Rohit Jain

Parag Agrawal

Mike Jahr

Milind Ganjoo

Sean Moon

Allen Chen

Tommy Chong

Xiangyi Zheng

Andrew Hazen Schlaikjer

Minali Aggarwal

A method for ranking messages of conversation graphs in a messaging platform includes classifying messages of a conversation graph into a plurality of sections based on content quality of the messages, where the plurality of sections include a first section and a second section. The first section has messages from the conversation graph determined as higher quality than messages of the second section. The method includes determining a plurality of predictive outcomes for each of a plurality of messages of the conversation graph using predictive models, ranking the messages of the first section based on the predictive outcomes for the messages of the first section, and transmitting at least a subset of the messages of the first section to be rendered on a client application according to the rank.

Hard and soft ranking messages of conversation graphs in a messaging platform

An example system includes a memory store of aggregate definitions. Each aggregate definition specifies a key value, an output store, a feature, a half-life value, and an aggregate operation metric to apply to a cross of the feature and the half-life value to generate aggregate metrics. The system also includes an aggregation engine that generates aggregate feature records from the input source based on the aggregate definitions and stores the aggregate feature records in the output store. An aggregate feature record includes an aggregate of the metric for the feature decayed over time using the half-life. The system also includes a query service that identifies, using the aggregate definitions, responsive aggregate feature records that satisfy parameters of a received request, applies the half-life to the responsive feature records, and provides the responsive feature records to a requester, the requester using the responsive feature records as input for a neural network.

Aggregate features for machine learning

According to an aspect, a method for ranking messages of conversation graphs in a messaging platform using machine-learning signals includes receiving a conversation view request to retrieve messages of a conversation graph stored on a messaging platform and determining a plurality of predictive outcomes for each of a plurality of messages of the conversation graph using predictive models. The determining step may include obtaining a plurality of signals relevant to the predictive models, the plurality of signals including data structure-related signals relating to the conversation graph and inputting the plurality of signals to the predictive models. The method includes ranking the plurality of messages based on the predictive outcomes and transmitting at least a subset of the plurality of messages to be rendered on a client application according to the rank.

Ranking messages of conversation graphs in a messaging platform using machine-learning signals

According to an aspect, a method for ranking messages of conversation graphs in a messaging platform using predictive outcomes includes receiving, over a network, a conversation view request to retrieve messages of a conversation graph stored on a messaging platform, and determining, by the messaging platform, a plurality of predictive outcomes for each of a plurality of messages of the conversation graph using predictive models. The plurality of predictive outcomes includes a reciprocal engagement probability. The method includes ranking, by the messaging platform, the plurality of messages based on the predictive outcomes, and transmitting, by the messaging platform, over the network, at least a subset of the plurality of messages to be rendered on a client application according to the rank.

Ranking messages of conversation graphs in a messaging platform using predictive outcomes

Growing community of inventors

Seattle, WA, United States of America

Arvind Thiagarajan