Dropbox, Inc.

Ermo Wei

Jongmin Baek

Jiarui Ding

Scott McCrae

Yuyang Guo

Ranjitha Gurunath Kulkarni

Yu Zhang

Morgan Zerby

Li Chen Koh

Mingye Xia

Xingyu Xiang

Jialiang Li

Kaiyue Sun

Shubham Goel

Maor Bar Asher

Prasang Upadhyaya

Aleksander Dash

Simon Shubbar

Wesley Liao

Win Suen

Tristan Inghelbrecht

The present disclosure relates to systems, methods, and non-transitory computer-readable media that utilize machine-learning models to classify content items and automatically organize the content items within a file structure according to their content item classifications. For instance, a content item classification system generates one or more content item classification models to determine classifications for content items and/or folders. In some instances, the classification system detects when new content items are added to a smart folder, determines destination folders to which the content items belong based on classifying the content items, and automatically moves the content items accordingly. In various instances, the classification system generates and utilizes a classification model to organize content items into dynamically-generated folders. In example implementations, the classification system generates and utilizes a classification model to automatically organize existing content items into existing folders.

Classifying and organizing digital content items automatically utilizing content item classification models

The present disclosure relates to systems, methods, and non-transitory computer-readable media that utilize machine learning models to generate identifier embeddings from digital content identifiers and then leverage these identifier embeddings to determine digital connections between digital content items. In particular, the disclosed systems can utilize an embedding machine-learning model that comprises a character-level embedding machine-learning model and a word-level embedding machine-learning model. For example, the disclosed systems can combine a character embedding from the character-level embedding machine-learning model and a token embedding from the word-level embedding machine-learning model. The disclosed systems can determine digital connections between the plurality of digital content items by processing these identifier embeddings for a plurality of digital content items utilizing a content management model. Based on the digital connections, the disclosed systems can surface one or more digital content suggestions to a user interface of a client device.

Utilizing machine-learning models to generate identifier embeddings and determine digital connections between digital content items

The present disclosure is directed toward systems, methods, and non-transitory computer-readable media for generating a dynamic facet by using a large language model. For example, the disclosed systems extract raw facet data from a plurality of content items stored in a content management system. In addition, the disclosed systems determine one or more facet content groups by grouping the plurality of content items according to the raw facet data. Further, the disclosed systems generate a facet prompt from the one or more facet content groups. Moreover, the disclosed systems generate a dynamic facet by providing the facet prompt to a large language model.

Dynamic facet generation using large language models

The present disclosure relates to systems, methods, and non-transitory computer-readable media that can leverage a natural language model to determine a most probable candidate sequence of tokens and thereby generate a predicted user activity. In particular, the disclosed systems can tokenize activity event vectors to generate a series of sequential tokens that correspond to recent user activity of one or more user accounts. In addition, the disclosed systems can, for each candidate (e.g., hypothetical) user activity, augment the series of sequential tokens to include a corresponding token. Based on respective probability scores for each of the augmented series of sequential tokens, the disclosed systems can identify as the predicted user activity, a candidate user activity corresponding to one of the augmented series of sequential tokens associated with a highest probability score. Based on the predicted user activity, the disclosed systems can surface one or more suggestions to a client device.

Utilizing a natural language model to determine a predicted activity event based on a series of sequential tokens

One or more embodiments of a synchronization system facilitate selectivity synchronizing digital content items from a collection of digital content items to a local storage of a client device. In particular, one or more embodiments described herein collect and analyze recall data for users of a digital content management system with respect to digital content items to determine synchronization scores for the digital content items. One or more embodiments described herein further include selectively identifying a subset of the digital content items based on the synchronization scores to recommend for synchronization to a local storage of a client device.

Selectively identifying and recommending digital content items for synchronization

Growing community of inventors

San Bruno, CA, United States of America

Ermo Wei