Abstract: 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? 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.

Abstract: 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? 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.

Abstract: 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? 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.

Abstract: A system and a method of training a Neural network (NN) model may include, receiving a pretrained NN model, that may include a plurality of layers, each associated with an activation matrix; selecting at least one, and performing an iterative training process on the layer. The iterative training process may include, applying an activation threshold to the activation matrix of the layer; measuring an accuracy value of the NN model; retraining the layer, while using a bimodal regularization function of one or more activation matrices of the NN model; and repeating the applying, measuring and retraining, while each repetition uses different activation threshold values. This repetition may be repeated until a maximal value of the activation threshold, where the NN model still converges, is found.

Abstract: 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? 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.

Abstract: 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.

Abstract: 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.

Abstract: A framed PV module rooftop and ballasted ground mount racking system by which the mounting structure also functions as a wire management system and operates as an electrical raceway that includes a plurality of North-South mounting structures, oriented in a North-South direction, each of the North-South mounting structures including a North-South open-air channel for receiving wires, and each of the North-South mounting structures configured for supporting a solar module at a selected tilt angle; a plurality of East-West mounting structures, oriented in an East-West direction, each of the East-West mounting structures including an East-West open-air channel for receiving wires, wherein the coupling of said North-South mounting structures to said East-West mounting structures forms a continuous, open-air, electrical raceway with intersections within an array for module connectivity.

Abstract: A framed PV module rooftop and ballasted ground mount racking system by which the mounting structure also functions as a wire management system and operates as an electrical raceway that includes a plurality of North-South mounting structures, oriented in a North-South direction, each of the North-South mounting structures including a North-South open-air channel for receiving wires, and each of the North-South mounting structures configured for supporting a solar module at a selected tilt angle; a plurality of East-West mounting structures, oriented in an East-West direction, each of the East-West mounting structures including an East-West open-air channel for receiving wires, wherein the coupling of said North-South mounting structures to said East-West mounting structures forms a continuous, open-air, electrical raceway with intersections within an array for module connectivity.

Abstract: 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.

Abstract: 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.

Abstract: A fluid disinfection module (10c) comprises a first housing (214) and a second housing (216) arranged in spaced-apart relationship to each other. The first housing (214) includes tubular support members (218) extending laterally therefrom for respective slidable mateable engagement with corresponding tubular support members (220) also laterally extending from second housing (216), to vary the lateral distance x between the first and second housings (214) and (216). A locking mechanism (260) preferably maintains tubular support members (218,220) in fixed relationship with respect to each other. First housing (214), second housing (216), corresponding tubular support members (218) and (220), and cross support member (230) define a framework for supporting one or more lamp assemblies comprising a UV lamp (18) and a protective sleeve (224) disposed about each UV lamp (18).

Abstract: Reinforced high early-strength cementitious members are made from water and strength generating ingredients, which include a hydraulic cement, a polycarboxylate dispersant and structural synthetic fiber. The strength generating ingredients provide corrosion resistance by avoiding the use of steel reinforcing members, increased ductility and load bearing capacity while developing a high early-strength of at least about 1,400 pounds per square inch (psi) flexural strength and at least about 7,500 pounds per square inch (psi) compressive strength within about 24 hours after placing the cementitious composition in a mold.

Abstract: Reinforced high early-strength cementitious members are made from water and strength generating ingredients, which include a hydraulic cement, a polycarboxylate dispersant and structural synthetic fiber. The strength generating ingredients provide corrosion resistance by avoiding the use of steel reinforcing members, increased ductility and load bearing capacity while developing a high early-strength of at least about 1,400 pounds per square inch (psi) flexural strength and at least about 7,500 pounds per square inch (psi) compressive strength within about 24 hours after placing the cementitious composition in a mold.

Abstract: A fluid disinfection module (10) comprises a first housing (12) and a second housing (20) arranged in spaced-apart relationship to each other. The first housing (12) includes tubular support members (28) and (28a) extending laterally therefrom for respective slidable mateable engagement with corresponding tubular support members (30) and (30a) also laterally extending from second housing (20), for varying the lateral distance x between the first and second housings (12) and (20). A locking device maintains tubular support members (28) (28a) and (30) (30a) in fixed relationship with respect to each other. First housing (12), second housing (20), and corresponding tubular support members (28) (28a) and (30) (30a) define a framework for supporting one or more ultraviolet radiation sources (18) and (26) that are detachably mounted at one end thereof to housings (12) and (20), respectively.