Methods and apparatus for training a matrix-based differentiable program using a photonics-based processor. The matrix-based differentiable program includes at least one matrix-valued variable associated with a matrix of values in a Euclidean vector space. The method comprises configuring components of the photonics-based processor to represent the matrix of values as an angular representation, processing, using the components of the photonics-based processor, training data to compute an error vector, determining in parallel, at least some gradients of parameters of the angular representation, wherein the determining is based on the error vector and a current input training vector, and updating the matrix of values by updating the angular representation based on the determined gradients.

The invention is directed to solving an optimization problem. The method operates a photonic crossbar array structure including N input lines and M output lines, which are interconnected at junctions via N×M photonic memory devices, where N≥2 and M≥2. The photonic memory devices are programmed to store respective weights in accordance with the optimization problem. The photonic crossbar array structure is operated as follows. First, the method determines values of L input vectors of N components each, where L≥2. Second, based on the determined values, N electromagnetic signals are generated, where each of the generated signals multiplexes L input signals encoded at respective wavelengths, so as for the N electromagnetic signals to map the L input vectors of N components each. Third, the N electromagnetic signals generated are applied to the N input lines of the photonic crossbar array structure.

The invention is directed to solving an optimization problem. The method operates a photonic crossbar array structure including N input lines and M output lines, which are interconnected at junctions via N×M photonic memory devices, where N≥2 and M≥2. The photonic memory devices are programmed to store respective weights in accordance with the optimization problem. The photonic crossbar array structure is operated as follows. First, the method determines values of L input vectors of N components each, where L≥2. Second, based on the determined values, N electromagnetic signals are generated, where each of the generated signals multiplexes L input signals encoded at respective wavelengths, so as for the N electromagnetic signals to map the L input vectors of N components each. Third, the N electromagnetic signals generated are applied to the N input lines of the photonic crossbar array structure.

Techniques related to a system for news classification comprising one or more non-transitory memory devices and one or more hardware processors configured to execute instructions from the one or more non-transitory memory devices to cause the system to receive an article, the article including text, extract text from the received article, store the extracted text in a database, determine a set of potential target entities based on the extracted text, determine a classification of the article for each potential target entity of the set of potential target entities for a category, valence, presence of litigation, rumor, or opinion based on the extracted text, associate the classification of the article, along with a probability of the determined classification of the article for each potential target entity, assign the classification of the article if the probability of the classification is greater than a threshold probability, and store the classification of the article and the probability.

Conventional techniques for automatically evaluating and grading assignments are generally ill-suited to evaluation of coursework submitted in media-rich form. For courses whose subject includes programming, signal processing or other functionally expressed designs that operate on, or are used to produce media content, conventional techniques are also ill-suited. It has been discovered that media-rich, indeed even expressive, content can be accommodated as, or as derivatives of, coursework submissions using feature extraction and machine learning techniques. Accordingly, in on-line course offerings, even large numbers of students and student submissions may be accommodated in a scalable and uniform grading or scoring scheme. Instructors or curriculum designers may adaptively refine assignments or testing based on classifier feedback. Using developed techniques, it is possible to administer courses and automatically grade submitted work that takes the form of media encodings of artistic expression, computer programming and even signal processing to be applied to media content.

A software application that is able to manage a collection of robot repairing resources can be used to assist technicians in repairing and solving hardware or software malfunctions within an electro-mechanical robot. The software application is able to simultaneous monitor multiple different electro-mechanical robots by receiving diagnostic information from them. The software application can identify a malfunction within one of the electro-mechanical robots by comparing its diagnostic information against a set of robot repair manuals. The software application will then select an optimal AI algorithm that has the best chance of repairing the hardware or software malfunction. The software application continues by implementing the optimal AI algorithm with a set of cloud accessible robot repairing applications, a technician's intervention, or a combination thereof.

The AI-Based Neighbor Discovery Search Engine Apparatuses, Methods and Systems (“ANDSE”) transforms embedding neural network training request, object search request inputs via ANDSE components into embedding neural network response, object search response outputs. An embedding neural network training request associated with a set of context objects is obtained. Sample similarity evaluation metrics are determined. For each context object, a set of positive target samples that satisfy the sample similarity evaluation metrics for the respective context object is determined. For each context object and each positive target sample in the respective set of positive target samples, a training example comprising the respective context object and a positive target sample is added to a training set. Configuration parameters for an embedding neural network are determined. The embedding neural network is trained using training examples in the training set. A datastructure that stores the adjusted weights of the embedding neural network is generated.

Embodiments of the present invention provide systems and methods to robustly inter-convert between polarization-entangled photon pairs and time-entangled photon pairs, such that produced polarization-entangled photons pairs can be converted into time-entangled photon pairs, stored as time-entangled photon pairs to preserve the entanglement for longer periods of time, and then converted back to polarization-entangled photon pairs when ready for manipulation, processing, and measurement by a quantum application.

Embodiments of the present invention provide systems and methods to robustly inter-convert between polarization-entangled photon pairs and time-entangled photon pairs, such that produced polarization-entangled photons pairs can be converted into time-entangled photon pairs, stored as time-entangled photon pairs to preserve the entanglement for longer periods of time, and then converted back to polarization-entangled photon pairs when ready for manipulation, processing, and measurement by a quantum application.

Embodiments of the invention relate to a time-division multiplexed neurosynaptic module with implicit memory addressing for implementing a universal substrate of adaptation. One embodiment comprises a neurosynaptic device including a memory device that maintains neuron attributes for multiple neurons. The module further includes multiple bit maps that maintain incoming firing events for different periods of delay and a multi-way processor. The processor includes a memory array that maintains a plurality of synaptic weights. The processor integrates incoming firing events in a time-division multiplexing manner. Incoming firing events are integrated based on the neuron attributes and the synaptic weights maintained.