Abstract: To achieve efficient reading of data from a memory including a plurality of banks by specifying different banks and accessing the memory from a plurality of hash computation circuits simultaneously, an information processing device includes a memory 1 including a plurality of banks, a plurality of hash computation circuits 8, and an interconnect 2 respectively connecting the banks in the memory 1 and the hash computation circuits 8 to each other, wherein the hash computation circuits 8 execute control in such a manner that read requests for reading data from the memory 1 respectively include bank numbers for specifying different banks in the same cycle.

Abstract: An optimal timing to perform dressing on a grindstone is detected easily. A machine learning device includes: an input data acquisition unit that acquires input data including an arbitrary grinding condition for an arbitrary work material in a grinding process of an arbitrary grinding machine and grindstone information related to an arbitrary grindstone including at least a degree of necessity of dressing process indicating a degree of necessity of dressing before a grinding process is performed under the grinding condition; a label acquisition unit that acquires label data indicating the degree of necessity of dressing process of the grindstone after the grinding process is performed under the grinding condition included in the input data; and a learning unit that executes supervised learning using the input data acquired by the input data acquisition unit and the label data acquired by the label acquisition unit, and generates a learned model.

Abstract: Transmedia scheduling systems and methods are described in which a user interface is generated via a channel engine that includes first and second channels. The channels are based on one or more channel templates stored in a channel database, and are preferably populated with concurrent transmedia stories. Each of the transmedia stories can be based on at least one transmedia object stored in a transmedia database and may be dynamically generated.

Abstract: Embodiments set forth techniques for managing advertisement auctions on a client device. The method can include the steps of (1) receiving, from a server device, a plurality of objects, where each object is associated with a respective digital asset, and each object includes, in association with the respective digital asset (i) a server-derived digital asset vector, (ii) a server-derived predicted tap-through rate, and (iii) a bid amount. In turn, and for each object of the plurality of objects, the client device (2) generates a respective estimated cost per impression for the object based on the information provided by the server device as well as information derived by the client device. Subsequently, the client device (3) identifies, among the plurality of objects, the object associated with the highest respective estimated cost per impression, and (4) causes an advertisement for the respective digital asset associated with the identified object to be displayed.

Abstract: Embodiments of the present disclosure generally relate to a network-based real-time enterprise travel management system that can facilitate planning, booking, scheduling, and managing travel, while also facilitating expense budgeting and reporting. Embodiments of the present disclosure can provide systems and methods for generating and managing a predictive itinerary associated with travel and/or an event, as well as a centralized marketplace for accessing various services and events.

Abstract: A system and method for securely training a model comprises including the algorithms of the model into a training module, and communicating the training module from a vendor computer system to a customer computer system. The training module is operated on the customer's computer system using the data from the customer, and a trained training module is achieved after operating the training module on the customer's computer system. The trained training module is used to initialize at least one additional module, which is communicated from the customer's computer system to the vendor's computer system, is used to score observations.

Abstract: Embodiments described herein relate to a method for probabilistically forecasting the state of hardware components. The method may include obtaining data items corresponding to a hardware component and performing an analysis of the hardware component. The analysis may include making a variety of probability predictions as to whether a label from among a set of possible labels is likely to be the correct label. The set of probabilities from the aforementioned analysis are then analyzed to determine which predicted label has the tightest range, and the prediction with the tightest range for a certain label is displayed to a user in a ranked fashion that includes a quantity of such probability prediction ranges. Such a display may allow an administrator to take action as to which hardware components should be replaced and in what order.

Abstract: Transmedia scheduling systems and methods are described in which a user interface is generated via a channel engine that includes first and second channels. The channels are based on one or more channel templates stored in a channel database, and are preferably populated with concurrent transmedia stories. Each of the transmedia stories can be based on at least one transmedia object stored in a transmedia database and may be dynamically generated.

Abstract: Methods and apparatus for detecting and correlating content consumption, e.g., viewing of programs and/or accessing websites, across platforms of different types are described. Content consumption, e.g., requests and/or streaming is monitored and recorded on a per subscriber basis for a wide variety of platforms, e.g., set boxes, IP content playback devices such as cell phones or pad devices with IP streaming support, and/or other devices. Web site access and corresponding content consumption is determined based on DNS requests thereby avoiding the need for monitoring at web site servers and allowing centralized monitoring of web site access operations in the form of DNS requests to be detected by a Web service provider who, at least in some embodiments, is also a content provider.

Abstract: The present disclosure provides a method of controlling a probability of a probability-based item provision system, which includes measuring an amount of use of an item provision system during a first time interval, comparing the measurement result with a first threshold, and changing the probability of the item provision system within a predetermined range during a second time interval after the first time interval, based on a result of the comparing.

Abstract: A point applying unit associates a predetermined point according to a type of regular purchase contract with user identification information according to a contractual user, for each predetermined period; a number-of-points information storage unit stores number-of-points information representing the number of points for each reward, in a plurality of providable rewards; a selection information acquisition unit retrieves selection information that is associated with the user identification information and represents a selected reward of the plurality of rewards; and a reward management unit enables the selected reward to be used, based on the user identification information, in a case where the total number of points are set by the selection information retrieved by the selection information acquisition unit and the number-of-points information in the number-of-points information storage unit and are associated with the selected reward has a predetermined relationship with the predetermined point associated b

Abstract: Systems and methods for redeeming a reward held by an individual are described. A method for redeeming a reward includes determining threshold criteria for provision of a targeted reward redemption offer, identifying at least one individual based upon the threshold criteria, determining the targeted reward redemption offer, and providing the targeted reward redemption offer to the identified individual. A response to the provided targeted reward redemption offer may be received, and an account of the identified individual may be adjusted in accordance with the targeted reward redemption offer and the received response.

Abstract: This disclosure involves generating graph data structures that model inter-feature dependencies for use with machine-learning models to predict end-user behavior. For example, a processing device receives an input dataset and a request to modify a first input feature of the input dataset. The processing device uses a graph data structure that models the inter-feature dependencies to modify the input dataset by propagating the modification of the first input feature to a second input feature dependent on the first input feature. The modification to the second input feature is a function of at least (a) the value of the first input feature and (b) a weight assigned to an edge linking the first input feature to the second input feature within the directed graph. The processing device then applies a trained machine-learning model to the modified input dataset to generate a prediction of an outcome.

Abstract: Methods and system for improved aircraft safety are described herein. A machine learning model may be trained using historical flight data for a number of previously flown flights where an automation function autonomously disengaged. The trained machine learning model may be used to generate probabilistic alert rules. The probabilistic alert rules may be used by a computing device onboard an aircraft to provide contextual information to flight crew relating to engagement status for one or more automation functions of the aircraft.

Abstract: A machine learning model is learned using secure vertical federated learning by receiving, by a network machine learning model, from a plurality of private machine learning models, a set of private machine learning model outputs. The set of private machine learning model outputs is based on data owned exclusively by each of the plurality of private machine learning models. The set of private machine learning model machine learning outputs is aligned based on sample IDs of the data. The network machine learning model, a prediction, the prediction being the output of the network model based on the set of private machine learning model outputs. Transmitting, by the network model, the prediction, to one of the plurality of private machine learning models, the one of the plurality of private machine learning models comprising labels.

Abstract: The Double Actor Critic (DAC) reinforcement-learning algorithm affords stable policy improvement and aggressive neural-net optimization without catastrophic overfitting of the policy. DAC trains models using an arbitrary history of data in both offline and online learning and can be used to smoothly improve on an existing policy learned or defined by some other means. Finally, DAC can optimize reinforcement learning problems with discrete and continuous action spaces.

Abstract: An adaptive system and a method for adjusting a parameter in the adaptive system includes operating an adaptive system with an output signal produced from an input signal applied to an input, in which a parameter with a finite range is determined based on a difference between the output signal and a target output signal. In one example the parameter with the finite range is a fixed-point parameter or an analog parameter. The parameter is accessed from the adaptive system. A probability distribution of the parameter is assigned. The finite range for the parameter is updated based on the probability distribution which has been assigned. The probability distribution function may be updated alone with the finite range of the parameter. The probability distribution may be derived from one or more historical values of the parameter, and a plurality of system parameters belonging to an identical category of data as the parameter.

Abstract: A data processing pipeline may be generated to include an orchestrator node, a preparator node, and an executor node. The preparator node may generate a training dataset. The executor node may execute machine learning trials by applying, to the training dataset, a machine learning model and/or a different set of trial parameters. The orchestrator node may identify, based on a result of the machine learning trials, a machine learning model for performing a task. The execution of the data processing pipeline may be optimized. Examples of optimizations include pooling multiple machine learning trials for execution at a single executor node, executing at least some machine learning trials using a sub-sample of the training dataset, and adjusting a proportion of trial parameters sampled from a uniform distribution to avoid a premature convergence to a local minima within the hyper-parameter space for generating the machine learning model.

Abstract: Systems, methods, and computer-readable media are disclosed for optimizing a programmatic advertising system by continuously adjusting certain policies with respect to the advertising bids to manage and balance commitments made to publishers and/or advertisers. The policies may be modified to adjust an attach rate and/or a guaranteed bid rate to balance various type of advertising bids (e.g., open auction, private market place deals, preferred deals, etc.) and ensure that commitments publishers and/or advertisers are satisfied. The system may continuously track new bids and may adjust the policies in real time or near real time to achieve certain targets for submissions, impressions, and other well-known advertising metrics.

Abstract: An information processing apparatus includes: a processor configured to: when a request for recommendation information and situation information are input, identify one or more actions associated with a situation indicated by the situation information based on action history information in which a past action is associated with a situation where the past action is executed, the situation information indicating a situation of a user; for each target action among the one or more identified actions, calculate a first probability and a second probability, the first probability being a probability that each target action is included as a choice for the user, the second probability being a probability that the user selects each target action when the each target action is included as the choice; and output an action with a relatively low first probability and a relatively high second probability, as the recommendation information, among the one or more actions.