Abstract: The techniques described herein relate to methods and apparatus for interferometric modulation. An apparatus includes an interferometric device comprising a first optical path and a second optical path, and at least one Franz-Keldysh (FK) modulator disposed in either the first optical path or the second optical path of the interferometric device. The interferometric device receives input light, wherein a first portion of the input light travels along the first optical path of the interferometric device, and a second portion of the input light travels along the second optical path of the interferometric device. The FK modulator modulates an intensity of either the first portion of the input light or the second portion of the input light.

Abstract: Methods, apparatus, systems, and articles of manufacture are disclosed to improve media identification. An example apparatus includes a hash handler to generate a first set of reference matches by performing hash functions on a subset of media data associated with media to generate hashed media data based on a first bucket size, a candidate determiner to identify a second set of reference matches that include ones of the first set, the second set including ones having first quantities of hits that did not satisfy a threshold, determine second quantities of hits for ones of the second set by matching ones to the hash tables based on a second bucket size, and identify one or more candidate matches based on at least one of (1) ones of the first set or (2) ones of the second set, and a report generator to generate a report including a media identification.

Abstract: The invention relates to combinations of peptides derived from a portion of an amino sequence of a grass pollen allergen, e.g. the allergens Phl p 1, Phl p 2, Phl p 3, Phl p 4 and/or Phl p 5, or a peptide variant thereof. Such peptides comprise at least one T cell epitope mid a high number of patients in a worldwide population will have HLA Class II alleles with the potential to bind the peptides of the peptide combinations. The invention also relates to the use of such peptide combinations in relieving an immune response caused by grass pollen species.

Abstract: The disclosed embodiments include a method performed by server computer(s). The method includes obtaining private healthcare insurance claims data and public healthcare procedure code data from one or more data source(s), generating a model based the training data, and determining a multiplier for each of the many facilities based on the model. The model provides predictive results for variability between healthcare facilities thereby enabling consumer research in light of considerations in facility variability. The multiplier is indicative of a value used to scale the public healthcare procedure code data for a facility such that a healthcare service associated with the facility can be estimated based on the multiplier.

Abstract: The disclosed technology teaches finding contents of one or more virtual machines running on one or more cloud servers. The disclosed technology includes a cloud snapshot metadata manager periodically instantiating indexing virtual machines on the cloud servers. The indexing virtual machines compile metadata of one or more virtual machines on the cloud server from one or more snapshot file systems of the virtual machines. The indexing virtual machines then transmit the compiled metadata to the cloud snapshot metadata manager. The cloud snapshot metadata manager forwards the metadata to a client and causing the client to create an index of the compiled metadata from multiple virtual machine snapshots on multiple clouds and responding to queries from a user using the index of compiled metadata, without requiring the user to instantiate or to attach to the snapshots.

Abstract: Token based dynamic data indexing with integrated security includes generating attribute tokens based on a database, each attribute token including an ordered sequence of symbols, generating an index indexing the attribute tokens, such that the index includes a root node, a first branch depending from the root node representing a first symbol from a first attribute token, and a security bitmask for the first attribute token, such that a security bitmask for the first branch at the root node is a hierarchical logical disjunction based on the security bitmask for the first attribute token, such that generating a response receiving an information retrieval request indicating a search symbol and a search context, is based on a determination that a horizontal logical disjunction of a vertical logical conjunction of the security bitmask for the first branch and a security bitmask for the information retrieval request indicates authorization.

Abstract: Systems and methods for forecasting future values of data streams are disclosed. One example method may include receiving information characterizing each of a plurality of forecasting models, retrieving historical data for each of a plurality of data streams, determining one or more constraints, dynamically selecting one of the plurality of forecasting models for each of the data streams based on accuracy metrics for the forecasting models, estimating cost metrics associated with each forecasting model, dynamically selecting the forecasting model based at least in part on the accuracy metrics, the cost metrics, and the determined constraints, and forecasting a first subsequent value of each data stream using the corresponding selected forecasting model.

Abstract: Embodiments described herein provides a training mechanism that transfers the knowledge from a trained BERT model into a much smaller model to approximate the behavior of BERT. Specifically, the BERT model may be treated as a teacher model, and a much smaller student model may be trained using the same inputs to the teacher model and the output from the teacher model. In this way, the student model can be trained within a much shorter time than the BERT teacher model, but with comparable performance with BERT.

Abstract: An embodiment provided herein preprocesses the input samples to the classification neural network, e.g., by adding Gaussian noise to word/sentence representations to make the function of the neural network satisfy Lipschitz property such that a small change in the input does not cause much change to the output if the input sample is in-distribution. Method to induce properties in the feature representation of neural network such that for out-of-distribution examples the feature representation magnitude is either close to zero or the feature representation is orthogonal to all class representations. Method to generate examples that are structurally similar to in-domain and semantically out-of domain for use in out-of-domain classification training. Method to prune feature representation dimension to mitigate long tail error of unused dimension in out-of-domain classification. Using these techniques, the accuracy of both in-domain and out-of-distribution identification can be improved.

Abstract: Systems and methods for forecasting cashflows across one or more accounts of a user disclosed. One example method may include retrieving a data set for each of a plurality of accounts from a database, constructing a graph including a plurality of nodes linked together by a multitude of edges, wherein each node identifies a time series value corresponding to one of the accounts, and each edge indicates a time series value of a corresponding set of transactions occurring between a corresponding pair of accounts, determining a plurality of constraints, determining a specified loss function based on the plurality of constraints, back-propagating a derivative of the specified loss function into a deep neural network (DNN) to determine a set of neural network parameters, forecasting, using the DNN, a time sequence for one or more of the nodes and one or more of the edges, and providing the forecasted time sequences to the user.

Abstract: A system may include a quantum model engine configured to generate a quantum computing model to represent an electronic design automation (EDA) process for a circuit design. The EDA process may be a multi-patterning process to assign colors to geometric elements of the circuit design, and the quantum computing model may include an objective function that specifies a cost value for a given state of the quantum computing model. Generation of the quantum computing model may include adaptively determining a penalty term in the objective function based on a circuit analysis of the circuit design. The quantum model engine may also be configured to generate a color assignment for the geometric elements of the circuit design through the quantum computing model. The system may also include a manufacture support engine configured to use the color assignment to support manufacture of circuit layers of the circuit design through multiple manufacturing steps.

Abstract: Embodiments described herein are generally directed to an improved vector normalization instruction. An embodiment of a method includes responsive to receipt by a GPU of a single instruction specifying a vector normalization operation to be performed on V vectors: (i) generating V squared length values, N at a time, by a first processing unit, by, for each N sets of inputs, each representing multiple component vectors for N of the vectors, performing N parallel dot product operations on the N sets of inputs. Generating V sets of outputs representing multiple normalized component vectors of the V vectors, N at a time, by a second processing unit, by, for each N squared length values of the V squared length values, performing N parallel operations on the N squared length values, wherein each of the N parallel operations implement a combination of a reciprocal square root function and a vector scaling function.

Abstract: An embodiment proposed herein uses sparsification techniques to train the neural network with a high feature dimension that may yield desirable in-domain detection accuracy but may prune away dimensions in the output that are less important. Specifically, a sparsification vector is generated based on Gaussian distribution (or other probabilistic distribution) and is used to multiply with the higher dimension output to reduce the number of feature dimensions. The pruned output may be then used for the neural network to learn the sparsification vector. In this way, out-of-distribution detection accuracy can be improved.

Abstract: Methods, systems, and devices for game matchmaking are described. The methods, systems, and devices for game matchmaking may include determining a ranking of a first user of a set of users in a game environment, monitoring, via a sensor, a performance attribute of the first user while the first user plays a game in the game environment, modifying the ranking of the first user in the game environment based on the monitored performance attribute of the first user, and, in some examples, matching the first user with a second user of the set of users based on the modified ranking of the first user.

Abstract: Examples described herein relate to a decompression engine that can request compressed data to be transferred over a memory bus. In some cases, the memory bus is a width that requires multiple data transfers to transfer the requested data. In a case that requested data is to be presented in-order to the decompression engine, a re-order buffer can be used to store entries of data. When a head-of-line entry is received, the entry can be provided to the decompression engine. When a last entry in a group of one or more entries is received, all entries in the group are presented in-order to the decompression engine. In some examples, a decompression engine can borrow memory resources allocated for use by another memory client to expand a size of re-order buffer available for use. For example, a memory client with excess capacity and a slowest growth rate can be chosen to borrow memory resources from.

Abstract: A method of creating a polymer surface with surface structures is disclosed. The method includes creating a mold, forming a metal sheet into the molds, creating a surface structure on a surface of the metal sheet by exposing the surface to laser pulses, and bringing a curable polymer to be in contact with the surface of the metal sheet containing the surface structure, curing the curable polymer, and separating the cured polymer from the metal sheet, resulting in a polymer surface containing the surface structure. The polymer surfaces with the surface structures can be hydrophobic or superhydrophobic depending on the micro and nano features contained by the surface structures.

Abstract: Embodiments provide for controlling power production in an islanded microgrid system while maintaining the system frequency and implementing desired load sharing between different types of generating resources and energy storage systems. Embodiments include a controller in communication with the resources to control operation of the resources and operative to determine an optimal load balance based on load and renewable generation forecast information; transmit load and generation schedules to a generation controller that operates the resources in accordance with the schedules; calculate a frequency set point for energy storage resources in the microgrid based on the optimal load balance of the energy storage resources, droop settings, rated power, and the frequency of the energy storage resources; and to control the frequency set point of the energy storage resources to achieve the optimal load balance. Numerous additional aspects are disclosed.

Abstract: An engine system includes an air handling control unit which controls a plurality of air handling actuators responsible for maintaining flow of air and exhaust gas within the engine system. The engine system has a plurality of sensors whose sensor signals at least partially define a current state of the engine system. The air handling control unit includes a controller which controls the air handling actuators of the engine system as well as a processing unit coupled to the sensors and the controller. The processing unit includes an agent which learns a policy function that is trained to process the current state, determines a control signal to send to the controller by using the policy function after receiving the current state as an input, and outputs the control signal to the controller. Then, the agent receives a next state and a reward value from the processing unit and updates the policy function using a policy evaluation algorithm and a policy improvement algorithm based on the received reward value.

Abstract: Some embodiments include a camera voice coil motor (VCM) actuator configured to shift a lens and/or an image sensor along multiple axes. The VCM actuator may include a bottom flexure and a top flexure that connect one or more dynamic members to one or more static members. The VCM actuator may include stationary magnets and coils held by dynamic members. In some cases, the VCM actuator may be configured to move the image sensor along an optical axis, to move the image sensor in directions orthogonal to the optical axis, and/or to tilt the image sensor relative to the orthogonal axis. In some examples, the VCM actuator may be configured to move the image sensor in directions orthogonal to the optical axis, to move the lens along the optical axis, and/or to tilt the lens relative to the optical axis.

Abstract: In one aspect, a method includes an input buffer receiving content; accessing the content from the input buffer and extracting from the accessed content, in-band message data that indicates a start time associated with a content-modification opportunity for a content-presentation device; the input buffer transmitting the content to a delay buffer; the delay buffer receiving the content from the input buffer and, after a delay time-period, transmitting the content to an output buffer; the output buffer receiving the content from the delay buffer and outputting for presentation a portion of the content; and after outputting for presentation the portion of the content and proximate a time-point associated with the indicated start-time, starting to perform a content-modification operation in connection with the content-modification opportunity.