Neuralmagic Inc.

Burst Corporation

Mark Kurtz

Nir N Shavit

Alexander Matveev

Justin Kopinsky

Dan Alistarh

Rati Gelashvili

Jayachandran Ramadorai

Smitha Gopu

Himanshu Roongta

Aditya Chaganti

A system and method of accelerating execution of a NN model, by at least one processor may include: receiving a first matrix A, representing elements of a kernel K of the NN model and a second matrix B, representing elements of an input I to kernel K; producing from matrix A, a group-sparse matrix A', comprising G tensors of elements. The number of elements in each tensor is defined by, or equal to a number of entries in each index of an input tensor register used for a specific Single Instruction Multiple Data (SIMD) tensor operation, and all elements of A&#x2032; outside said G tensors are null. The system and method may further include executing kernel K on input I, by performing at least one computation of the SIMD tensor operation, having as operands elements of a tensor of the G tensors and corresponding elements of the B matrix.

System and method of accelerating execution of a neural network

In one embodiment, a method is disclosed for providing UGC to a media production or broadcast destination for potential inclusion in a broadcast or other media distribution. Video processing parameters of the destination are provided. An application is customized using the parameters to provide a customized application without requiring any additional development for the customization. A token and a link are then transmitted to the destination. The destination accesses the link, and the customized application is downloaded along with an API to the destination. The customized application periodically polls the API for UGC and associated metadata. It then, guided by the parameters, facilitates the ingestion of the media and metadata in the proper format into the appropriate place within the downstream broadcast or media production system.

Techniques for integration of media content from mobile device to broadcast

In one embodiment, a montage of media (videos, images, audio, etc.) is created automatically from media submitted by different users and related to a common theme or event. A bubble is created for each theme or event. The users submit the media over the Internet or another network with an indication of a particular bubble the media pertains to. The media in the bubble is examined for objects, audio, engagement metrics, faces, and various other accompanying markers, which are then used to identify media or fragments of media for inclusion in an automatic montage, with music and transition effects.

Techniques for creation of auto-montages for media content

Growing community of inventors

Medford, MA, United States of America

Mark Kurtz