Abstract: Described is a machine learning system for binary classifications. The system improves interpolation schemes used in isotonic regressions by providing a continuous function that also is monotonic. The system outputs a probability estimating function on a signal space that is both monotonic and varies continuously with the input signals. More specifically, described is an interpolation function that is continuous and piecewise linear on Delaunay simplices. Accordingly, the resulting probability estimation function may more accurately match actual probabilities especially when training data is sparse.

Abstract: Systems and methods for reducing a churn rate associated with subscribers of social network services are described. In some example embodiments, the systems and methods may access activity information associated with a former subscriber of a social network service, compare the accessed activity information to activity information associated with subscribers of the social network service, identify one or more differences between the activity information associated with the former subscriber of the social network service and the activity information associated with the subscribers of the social network service, and perform an action based on the identified one or more differences.

Abstract: A computer-implemented method and apparatus for supporting decisions in sequential clinical risk assessment examinations, the method comprising receiving one or more first test results and a question, both associated with a patient; and assessing by a processor associated with a computing platform, information gain provided by a second test which may be performed for the patient, as the conditional mutual information between a second test and the question, using the first test result.

Abstract: A recommended tire selection system 100 comprises a driving condition obtaining unit 101 that obtains a driving condition of a vehicle on which the recommended tire is loaded, a retread necessity obtaining unit 103 that obtains the retread necessity to replace the tread; a performance decision unit 105 that decides a rolling resistance and a wear resistance of the recommended tire based on the obtained driving condition, and decides a casing durability based on the obtained retread necessity; and a selection unit 107 that selects a recommended tire that is a combination of a tread and a casing, fulfilling the decided rolling resistance, wear resistance, and casing durability.

Abstract: The present invention discloses methods, systems, and tools for evaluating a number of value significance measures of ontological subjects of compositions or networks. The method breaks a composition into its constituent ontological subjects of different orders and builds a participation matrix indicating the participation of ontological subjects of the composition in other ontological subjects, i.e. the partitions, of the composition. Using the participation information of the OSs into each other, an association strength matrix is built from which the value significance measures of the partitions of the composition are calculated. The methods systematically calculate the value significances of the ontological subjects of different orders of the composition. Various systems for implementing the methods and some exemplary applications and services are disclosed.

Abstract: A conversation-simulating system facilitates simulating an intelligent conversation with a human user. During operation, the system can receive a user-statement from the user during a simulated conversation, and generates a set of automatic-statements that each responds to the user-statement. The system then determines a set of behavior-characteristics for the user, and computes relevance scores for the automatic-statements based on the behavior-characteristics. Each relevance score indicates an outcome quality that the user is likely to perceive for the automatic-statement as a response to the user-statement. The system selects an automatic-statement that has a highest relevance score from the set of automatic-statements, and provides the selected automatic-statement to the user.

Abstract: A method for providing a conditional scored list of applications for use in recommending applications includes storing on a cloud computing service a conditional probability table across a set of available applications provided by the cloud computing service. The cloud computing service receives a request to provide a scored list of applications for a user, retrieves a set of user-installed applications for the user, and calculates a total conditional probability for each application in the set of available applications. The cloud computing service then constructs the scored list of applications from the set of available applications, where a score of each application is its corresponding total conditional probability, and outputs the scored list of applications.

Abstract: The location of the memory controllers within the on-chip fabric of multiprocessor architectures plays a central role in latency bandwidth characteristics of the processor-to-memory traffic. Intelligent placement substantially reduces the maximum channel load depending on the specific memory controller configuration selected. A variety of simulation techniques are used along and in combination to determine optimal memory controller arrangements. Diamond-type and diagonal X-type memory controller configurations that spread network traffic across all rows and columns in a multiprocessor array substantially improve over other arrangements. Such placements reduce interconnect latency by an average of 10% for real workloads, and the small number of memory controllers relative to the number of on-chip cores opens up a rich design space to optimize latency and bandwidth characteristics of the on-chip network.

Abstract: System and method for incrementally varying input levels and recording respective output temperatures at the varied levels, storing, in a training table, first correlations between the varied levels of the input parameter and defined temperature ranges, detecting an input parameter level during operation and re-recording the output temperature at the detected level, generating a second correlation between the detected level and a second temperature range, determining whether the second temperature range is different from a first temperature range from among the defined temperature ranges, the first temperature range correlating to the level of the input parameter corresponding to the detected level, updating the training table when the second temperature range is different from the first temperature range, predicting an expected temperature range, and dynamically scaling, based on the expected temperature range, the detected level of the input parameter to control the output temperature in real-time.

Abstract: Disclosed is a method of recovering from overflow of a hardware dynamically reconfigurable NFA cell array, to find matches within a symbol stream to regular expression or similar rules without missing matches due to overflow. Upon overflow, active states are selected to spill from the cell array, saving state information and spill position. Scanning continues a limited distance, with additional overflow spills possible, to a selected end of segment position where all active end states are removed and recorded. A re-scan of the segment from the first overflow position begins with each previously spilled state re-injected at the same position it was spilled from. At the end of the segment, saved end states are re-injected and scanning continues. RE-scans may iterate if there was additional overflow. NFA states may be assigned color codes, with connected states receiving the same color, to aid in efficient overflow recovery.

Abstract: Systems, methods, means, computer program code and computerized processes include receiving a first set of de-identified transaction data from a first transaction data source, receiving a second set of de-identified transaction data from a second transaction data source, filtering the first and second sets of de-identified transaction data to identify transactions associated with at least a first entity and to create first and second filtered data sets, removing data associated with an identifier field for each of the transactions in the first filtered data set to created a de-identified first data set, removing data associated with an identifier field for each of the transactions in the second filtered data set to create a de-identified second data set, and processing the first and second de-identified data sets using a probabilistic engine to establish a linkage between data in each data set.

Abstract: A modeling framework for evaluating the impact of weather conditions on farming and harvest operations applies real-time, field-level weather data and forecasts of meteorological and climatological conditions together with user-provided and/or observed feedback of a present state of a harvest-related condition to agronomic models and to generate a plurality of harvest advisory outputs for precision agriculture. A harvest advisory model simulates and predicts the impacts of this weather information and user-provided and/or observed feedback in one or more physical, empirical, or artificial intelligence models of precision agriculture to analyze crops, plants, soils, and resulting agricultural commodities, and provides harvest advisory outputs to a diagnostic support tool for users to enhance farming and harvest decision-making, whether by providing pre-, post-, or in situ-harvest operations and crop analyzes.

Abstract: A method for supporting to collect hard negative image is provided. The method includes the steps of: (a) a classifier receiving from a hard negative proposer, and classifying, hard negative image candidate corresponding to a certain label candidate or a specific label candidate judged to have semantic or visual relationship with a target object; and (b) the classifier (i) allowing the hard negative proposer to select an additional label candidate whose similarity to the certain label candidate or the specific label candidate exceeds a preset degree of similarity, if a percentage or a number of the hard negative image candidate mistaken for having the target object is judged to satisfy a prefixed condition, (ii) receiving at least one additional hard negative image candidate corresponding to the selected additional label candidate and (iii) classifying the additional hard negative image candidate.

Abstract: Systems and methods are provided for expert systems for well completion using Bayesian decision networks to determine well completion recommendations. The well completion expert system includes a well completion Bayesian decision network (BDN) model that receives inputs and outputs recommendations based on Bayesian probability determinations. The well completion BDN model includes a treatment fluids section, a packer section, a junction classification section, a perforation section, a lateral completion section, and an open hole gravel packing section.

Abstract: A dynamic random access memory (DRAM) comprising a programmable intelligent search memory (PRISM) for regular expression search using non-deterministic finite state automaton and further comprising a cryptography processing engine for performing encryption and decryption, said PRISM and cryptography processing engines creating a secure DRAM for use in a system.

Abstract: Methods, systems and apparatus, including computer programs encoded on a computer storage medium, for receiving aggregate user data, the aggregate user data corresponding to response rate of one or more answering users responding to requests, processing the aggregate user data to generate an analytical model, the analytical model describing a probability that an average answering user will respond to a request based on response rates of one or more time intervals, receiving a request, identifying a plurality of answering users, processing the analytical model to identify a sub-set of answering users of the plurality of answering users, and transmitting the request to each answering user of the sub-set of answering users.

Abstract: A medical general intelligence computer system and computer-implemented methods analyze morpho-physiological numbers for determining a risk of an emergent disease state, determining an emergent disease state, predicting a pre-emergent disease state, determining a pre-emergent disease state, and/or predicting a risk of a pre-emergent disease state.

Abstract: Disclosed is a network computer system which may comprise a hardware processor which may comprise a programmable intelligent search memory for content search. The programmable intelligent search memory may perform regular expression based search. The programmable intelligent search memory may use at least one regular expression. The regular expression may be converted into at least one non-deterministic finite state automata (NFA) representing the functionality of the regular expression.

Abstract: Disclosed is a method and system for matching a complex NFA state comprising a spinner followed by a character class sequence which may be represented by the general regular expression form [S] {N,M}[A0][A1] . . . [Ak?1]. An input transition activates the spinner and the spin count increments with successive matches of the spin class [S]. When the spin count is between N and M, sequence matching begins. Several base sequence CCLs are compared in parallel with a corresponding window of input symbols. If all match, a signal enters a delay line until the end of the base sequence. When the signal exits the delay line, extended sequence CCLs are accessed from a table sequentially and compared with successive input symbols. After the final extension CCL matches, an output transition is signaled. For short sequences, unused base sequence CCLs may be configured with look-ahead classes.

Abstract: A method is provided for estimating past data by identifying a high frequency data set for a defined time period. A pattern is calculated for the high frequency data set and then the pattern is applied to a low frequency data set in a past time period to estimate a high frequency query point.