Intersil Americas Inc.

Google Inc.

Warren Craddock

David W Ritter

Robert David Zucker

Steven G Herbst

Itaru Hiromi

Philip V Golden

Fabien Viger

Pradeep Bhardwaj

Elias Andrikopoulos

Systems and methods are provided for determining an order of execution of a neural network. For instance, data indicative of a neural network and data indicative of an amount of available memory in a constrained memory space can be obtained. The neural network can include a plurality of operators. An order of execution associated with the neural network can then be determined. The order of execution specifies an order in which to execute each of the plurality of operators. The order of execution is determined based at least in part on the available memory in the constrained memory space. In particular, one or more graph search algorithms can be performed on a graph that is representative of the neural network to obtain the order of execution.

Determining orders of execution of a neural network

Described herein are optical proximity detectors, methods for use therewith, and systems including an optical proximity detector. Such optical proximity detectors include an analog front-end and a digital back-end. In certain embodiments, the digital back-end includes a dynamic gain and phase offset corrector, a cross-talk corrector, a phase and magnitude calculator, and a static phase offset corrector. The dynamic gain and phase offset corrector corrects for dynamic variations in gain and phase offset of the analog front-end due to changes in temperature and/or operating voltage levels. The crosstalk corrector corrects for electrical and/or optical crosstalk associated with the analog front-end. The phase and magnitude calculator calculates phase and magnitude values in dependence on the corrected versions of digital in-phase and quadrature-phase signals received from the analog front-end. The static phase offset corrector corrects for a static phase offset of the optical proximity detector.

Optical proximity detectors

Provided herein are methods and systems that provide automatic compensation for frequency attenuation of a video signal transmitted over a cable. In accordance with an embodiment, a system includes an equalizer and a compensation controller. The equalizer receives a video signal that was transmitted over a cable, provides compensation for frequency attenuation that occurred during the transmission over the cable, and outputs a compensated video signal. The compensation controller automatically adjusts the compensation provided by the equalizer based on comparisons of one or more portions of the compensated video signal to one or more reference voltage levels. The compensating is selectively locked and reset in response to specific conditions being detected, e.g., a locking condition and a reset condition.

Cable equalization locking

Provided herein are methods and systems that provide automatic compensation for frequency attenuation of a video signal transmitted over a cable. In accordance with an embodiment, a system includes an equalizer and a compensation controller. The equalizer receives a video signal that was transmitted over a cable, provides compensation for frequency attenuation that occurred during the transmission over the cable, and outputs a compensated video signal. The compensation controller automatically adjusts the compensation provided by the equalizer based on comparisons of one or more portions of the compensated video signal to one or more reference voltage levels.

Systems and methods for cable equalization

Systems and methods to provide automatic compensation for frequency attenuation of a video signal transmitted over a cable are provided. Frequency compensation is applied to a received video signal that was transmitted over the cable to thereby produce a compensated video signal. A shape of a horizontal sync portion of the compensated video signal is analyzed. The frequency compensation is automatically adjusted based on the shape of the horizontal sync portion of the compensated video signal. This abstract is not intended to be a complete description of the various embodiments of the present invention.

Growing community of inventors

Alameda, CA, United States of America

Warren Craddock