Box Limited

Google Inc.

Bryan Huh

Florian Jourda

Pierre-Alexandre Masse

Jeremy Glassenberg

Justin Tim Peng

Hieu Minh Nguyen

Winston Yih-Jiun Huang

Victor De Vansa Vikramaratne

Tanooj Luthra

Ritik Malhotra

Dirk Padfield

Victor DeVansa Vikramaratne

Noah Murad

Edward Lo

An automated speech recognition (ASR) transcript of at least a portion of a media content is obtained from an ASR tool. Suggested words are received for corrected words of the ASR transcript of the media content. Features are obtained using at least the suggested words or the corrected words. The features include features relating to sound similarities between the suggested words and the corrected words. The features are input into a machine learning (ML) model to obtain a determination regarding a validity of the suggested words. Responsive to the suggested words constituting a valid suggestion, the suggested words are incorporated into the ASR transcript. At least a portion of the ASR transcript is transmitted to a user device in conjunction with at least a portion of the media content.

False suggestion detection for user-provided content

Systems, methods and computer program products for high-performance, low latency start-up of large shared media files. A method for low latency startup with low defect playback commences upon identifying a first media file having a first format to be converted to a second media file having a second format. A scheduler divides the first media file into multiple partitions separated by partition boundaries. The method continues by converting the partitions into respective converted partitions that comport with the second format. Determinations as to the position of the partition boundaries is made based on measurable conditions present at a particular moment in time. Different formats receive different treatment based on the combination of characteristics of the first format, characteristics of the second format, as well as on characteristics of measurable conditions present at the moment in time just before conversion of a segment.

Low latency and low defect media file transcoding using optimized storage, retrieval, partitioning, and delivery techniques

Systems for dynamic watermarking within a cloud-based collaboration environment. Storage facilities are managed by servers to provide cloud-based storage services to two or more client entities as a service across a network. The servers maintain versions of editable documents or photo images or video clips using a version indication. Dynamic watermarking applies a first watermark image to a first version of the editable document and grants edit access to a first client entity of the two or more client entities. While the first client entity has edit access to the first version of the editable document the system generates a second version of the editable document and applies a second watermark image to the second version of the editable document before delivering the second version to a second client entity. A security watermark is based on a user credential, and a leak source can be determined based the user credential.

Event-driven generation of watermarked previews of an object in a collaboration environment

Growing community of inventors

San Jose, CA, United States of America

Bryan Huh