Abstract: A method including building a recommendation triggering model. The method can include receiving, via a user device of a user through a network, an add-to-cart command associated with an anchor item for the user. The method further can include determining, in real-time after receiving the add-to-cart command, a recommendation for one or more complementary items of the anchor item for the user. The method also can include determining, in real-time after determining the recommendation, a recommendation confidence for the recommendation. The method additionally can include after determining the recommendation confidence, when the recommendation confidence is positive, transmitting, in real-time through the network, the one or more complementary items to be presented to the user via the user device. The method likewise can include after determining the recommendation confidence, when the recommendation confidence is not positive, refraining from transmitting the one or more complementary items to the user.

Abstract: Embodiments provide methods and systems for generating domain-specific text summary. Method performed by processor includes receiving request to generate text summary of textual content from user device of user and applying pre-trained language generation model over textual content for encoding textual content into word embedding vectors. Method includes predicting current word of the text summary, by iteratively performing: generating first probability distribution of first set of words using first decoder based on word embedding vectors, generating second probability distribution of second set of words using second decoder based on word embedding vectors, and ensembling first and second probability distributions using configurable weight parameter for determining current word. First probability distribution indicates selection probability of each word being selected as current word.

Abstract: Audio distortion compensation methods to improve accuracy and efficiency of audio content identification are described. The method is also applicable to speech recognition. Methods to detect the interference from speakers and sources, and distortion to audio from environment and devices, are discussed. Additional methods to detect distortion to the content after performing search and correlation are illustrated. The causes of actual distortion at each client are measured and registered and learnt to generate rules for determining likely distortion and interference sources. The learnt rules are applied at the client, and likely distortions that are detected are compensated or heavily distorted sections are ignored at audio level or signature and feature level based on compute resources available. Further methods to subtract the likely distortions in the query at both audio level and after processing at signature and feature level are described.

Abstract: Various embodiments include a camera with folded optics and lens shifting capabilities. In some examples, a folded optics arrangement of the camera may include one or more lens elements and light path folding elements (e.g., prisms). Some embodiments include voice coil motor (VCM) actuator arrangements, carrier arrangements, and/or suspension arrangements to provide autofocus (AF) and/or optical image stabilization (OIS) movement. Furthermore, some embodiments include position sensor arrangements for position sensing with respect to AF and/or OIS movement.

Abstract: A computer-implemented method includes using a cache replacement algorithm to forcefully evict target data from a cache. Using the cache replacement algorithm includes selectively accessing data in the cache. A computer program product includes one or more computer readable storage media and program instructions collectively stored on the one or more computer readable storage media. The program instructions include program instructions to perform the foregoing method. A system includes a processor and logic integrated with the processor, executable by the processor, or integrated with and executable by the processor. The logic is configured to perform the foregoing method.

Abstract: Systems and methods include receiving a domain for a determination of a likelihood the domain is malicious or benign; obtaining data associated with the domain including log data from a cloud-based system that performs monitoring of a plurality of users; analyzing the domain with a plurality of components to assess the likelihood, wherein at least one of the plurality of components is a trained machine learning model; and combining results of the plurality of components to predict the likelihood the domain is malicious or benign.

Abstract: In one aspect, a method includes identifying an upcoming content-modification opportunity on a channel, where the upcoming content-modification opportunity relates to a modifiable content-segment transmitted by a content-distribution system on the channel, and where the modifiable content-segment includes an overlay content portion and a non-overlay content portion. The method also includes responsive to identifying the upcoming content-modification opportunity on the channel, identifying a content-presentation device tuned to the channel. The method also includes determining location data associated with the overlay content portion.

Abstract: The present invention relates to the field of electronic transactions, and provides methods and systems for reducing user interventions necessary for authentication of transactions. In an embodiment the invention implements an authentication server configured to receive from a terminal device an electronic payment transaction request and device IDs corresponding to devices having wireless communication capability that are detected within a wireless communication range of the terminal device. The authentication server selects a payment account for implementing the requested electronic payment transaction, wherein selection of the payment account is based on a determination that one or more of the device IDs received from the terminal device are associated with the identified payment account, and thereafter generates a payment transaction authentication decision.

Abstract: The overall architecture and details of a scalable video fingerprinting and identification system that is robust with respect to many classes of video distortions is described. In this system, a fingerprint for a piece of multimedia content is composed of a number of compact signatures, along with traversal hash signatures and associated metadata. Numerical descriptors are generated for features found in a multimedia clip, signatures are generated from these descriptors, and a reference signature database is constructed from these signatures. Query signatures are also generated for a query multimedia clip. These query signatures are searched against the reference database using a fast similarity search procedure, to produce a candidate list of matching signatures. This candidate list is further analyzed to find the most likely reference matches. Signature correlation is performed between the likely reference matches and the query clip to improve detection accuracy.

Abstract: A system and method are disclosed associated with a multi-tenant cloud computing environment. The system may receive information about a serverless function workload (e.g., a NodeJS, Java function or ABAP workload) to be launched in the cloud computing environment. A tenant associated with the serverless function workload to be launched may be identified and, based at least in part on the identified tenant, an objective function (e.g., throughput, latency, cost, etc.) for the serverless function workload to be launched may be identified. A recommendation service platform may then iteratively configure tuning parameters of the cloud computing environment using Bayesian optimization (e.g., to reach a global optimum using a Gaussian process) and the determined objective function. The system may then arrange for the serverless function workload to be executed in the cloud computing environment in accordance with the configured tuning parameters.

Abstract: Photonic processors are described. The photonic processors described herein are configured to perform matrix-matrix (e.g., matrix-vector) multiplication. Some embodiments relate to photonic processors arranged according to a dual-rail architecture, in which numeric values are encoded in the difference between a pair optical signals (e.g., in the difference between the powers of the optical signals). Relative to other architectures, these photonic processors exhibit increased immunity to noise. Some embodiments relate to photonic processors including modulatable detector-based multipliers. Modulatable detectors are detectors designed so that the photocurrent can be modulated according to an electrical control signal. Photonic processors designed using modulatable detector-based multipliers are significantly more compact than other types of photonic processors.

Abstract: A user's session of a web application or a website in a web browser is recorded and replayed while protecting private and sensitive data from unauthorized access. All the captured data needed to re-create (replay) the user's session in the browser itself is recorded and exported on demand. The need to transmit potentially sensitive and private data continuously to external server(s) is eliminated while still guaranteeing availability of a record of user activity leading up to any point of interest during the user's session. By encrypting recording information and redacting all non-layout content (e.g. text nodes, images, inputs) from the browser DOM before capturing the DOM, the visual layout of the page is maintained and the probability of leaking the user's sensitive or private information is reduced. The replaying user is still able to derive meaningful information about the user's interaction with the web application or website without jeopardizing privacy.

Abstract: When a media client is receiving a media stream, the media client determines an upcoming time point at which the media client is to perform a content revision involving insertion (e.g., substitution or overlaying) of new content. The media client further determines an advanced time point when the media client should initiate a process of acquiring the new content, setting the advanced time point sufficiently in advance of the upcoming content-revision time point to enable the media client to obtain at least enough of the new content to be able to start the content revision on time. In an example implementation, the media client could determine the advanced time point by predicting how long the content-acquisition process will take, based on consideration of past instances of content acquisition, possibly correlated with operational factors such as content source, processor load, memory load, network speed, and time of day.

Abstract: Some embodiments include a camera voice coil motor (VCM) actuator that includes an additive coil structure for shifting a lens along one or multiple axes. The additive coil structure may include a base portion configured to couple with a lens carrier and at least partially surround a perimeter of the lens carrier. In various examples, the additive coil structure may include folded portions that individually include a respective coil that is located proximate a respective magnet. According to various embodiments, the additive coil structure may be formed using an additive process.

Abstract: A system and method are disclosed to facilitate automated database system workload classification. A utilization metrics data source may contain utilization metrics vectors associated with workloads of the database system. A variational autoencoder may receive utilization metrics vectors from the utilization metrics data source and encode the utilization metrics vectors into latent vector features. Moreover, the variational autoencoder may be trained to generate appropriate distributions around the latent vector features. A synthetic workload creation platform receives information about the distributions around the latent vector features and samples different values from the distributions to create synthetic workload vectors. A workload classification platform, trained using the synthetic workload vectors, may then generate workload classification labels for workloads running on the database system (e.g., to tune database parameters as appropriate).

Abstract: A carrier head for a chemical mechanical polisher includes a base, a substrate mounting surface, an annular inner ring and an outer ring. The inner ring has a lower surface configured to contact an upper surface of a substrate positioned on the substrate mounting surface, an outer surface, and an inwardly facing surface extending downwardly from the lower surface and is configured to circumferentially surround the edge of the substrate, the inner ring vertically movable relative to the substrate mounting surface. The outer ring has an inner surface circumferentially surrounding the inner ring, an outer surface, and a lower surface to contact the polishing pad, and the outer ring is vertically movable relative to and independently of the substrate mounting surface and the inner ring.

Abstract: A method of battery management of an electronic device, the electronic device, and a non-transitory computer readable recording medium are provided. The method includes detecting, by the electronic device, that a battery management criterion is met, enabling at least one of a plurality of battery management functions, in response to detecting that the battery management criterion is met, and performing at least one action corresponding to the enabled at least one of the plurality of the battery management functions.

Abstract: Frames from an image stream or streams are processed by independently operating digital signal processors (DSPs), with only frame checking microprocessors operating in a lockstep mode. In one example, two DSP are operating on alternate frames. Each DSP processes the frames and produces prediction values for the next frame. The lockstep microprocessors develop their own next frame prediction. The lockstep processors compare issued frames and previously developed predicted frames for consistency. If the predictions are close enough, the issued frame passes the test. The lockstep processors then compare the issued frame to the preceding two frames for a similar consistency check. If the prior frames are also close enough, the issued frame is acceptable. In another example, hardware checkers are provided to compare the present frame with a larger number of prior frames. The hardware checkers provide comparison results to the lockstep processors to compare against allowable variation limits.

Abstract: Photonic processors are described. The photonic processors described herein are configured to perform matrix-matrix (e.g., matrix-vector) multiplication. Some embodiments relate to photonic processors arranged according to a dual-rail architecture, in which numeric values are encoded in the difference between a pair optical signals (e.g., in the difference between the powers of the optical signals). Relative to other architectures, these photonic processors exhibit increased immunity to noise. Some embodiments relate to photonic processors including modulatable detector-based multipliers. Modulatable detectors are detectors designed so that the photocurrent can be modulated according to an electrical control signal. Photonic processors designed using modulatable detector-based multipliers are significantly more compact than other types of photonic processors.

Abstract: An image signal processor includes a first matrix processing circuit, a post processing circuit, a second matrix processing circuit, and a split visual and analytics circuit. The first matrix processing circuit is configured to receive a plurality of component images generated based on an image captured by an image sensor and generate a plurality of first matrix outputs based on the plurality of component images. The post processing circuit is configured to perform color conversion on the plurality of first matrix outputs to generate a first luminance component of the image and a chrominance component of the image. The second matrix processing circuit is configured to perform color conversion on the plurality of first matrix outputs to generate a second luminance component of the image and a saturation component of the image. The split visual and analytics circuit is configured to generate visual and analytic data of the image.