The disclosure herein relates to business content analysis. In particular, the disclosure relates to systems and methods of an expense management system operable to perform automatic business documents' content analysis for generating business reports associated with automated value added tax (VAT) reclaim, Travel and Expenses (T&E) management, Import/Export management and the like. The system is further operable to provide various organizational expense management aspects for the corporate finance department and the business traveler based upon stored data. Additionally, the system is configured to use a content recognition engine, configured as an enhanced OCR mechanism used for extracting tagged text from invoice images and also provides continuous learning mechanism in a structured mode allowing classification of invoice images by type, providing continual process of improvement and betterment throughout.

Systems and methods described herein utilize supervised machine learning to generate a figure-of-speech prediction model for classify content words in a running text as either being figurative (e.g., as a metaphor, simile, etc.) or non-figurative (i.e., literal). The prediction model may extract and analyze any number of features in making its prediction, including a topic model feature, unigram feature, part-of-speech feature, concreteness feature, concreteness difference feature, literal context feature, non-literal context feature, and off-topic feature, each of which are described in detail herein. Since uses of figure of speech in writings may signal content sophistication, the figure-of-speech prediction model allows scoring engines to further take into consideration a text's use of figure of speech when generating a score.

Embodiments of the invention provide a method for mapping a bipartite graph onto a neuromorphic architecture comprising of a plurality of interconnected neuromorphic core circuits. The graph includes a set of source nodes and a set of target nodes. The method comprises, for each source node, creating a corresponding splitter construct configured to duplicate input. Each splitter construct comprises a first portion of a core circuit. The method further comprises, for each target node, creating a corresponding merger construct configured to combine input. Each merger construct comprises a second portion of a core circuit. Source nodes and target nodes are connected based on a permutation of an interconnect network interconnecting the core circuits.

A device includes a plurality of blocks. Each block of the plurality of blocks includes a plurality of rows. Each row of the plurality of rows includes a plurality of configurable elements and a routing line, whereby each configurable element of the plurality of configurable elements includes a data analysis element comprising a plurality of memory cells, wherein the data analysis element is configured to analyze at least a portion of a data stream and to output a result of the analysis. Each configurable element of the plurality of configurable elements also includes a multiplexer configured to transmit the result to the routing line.

A simulation method for a mixed-signal circuit system includes: detecting a plurality of registers and a clock signal included in the mixed-signal circuit system; performing a timing analysis converting operation upon a circuit block coupled between any two register of the plurality of registers to obtain a converted circuit system; and performing a Static Timing Analysis operation upon the converted circuit system; wherein when the circuit block is convertible into a combinational circuit block, the timing analysis converting operation includes: converting the circuit block to the combinational circuit block, wherein the combinational circuit block is logic gate-level.

The disclosure provides an approach for recognizing and analyzing activities. In one embodiment, a learning application trains parameters of a hierarchical model which represents human (or object) activity at multiple levels of detail. Higher levels of detail may consider more context, and vice versa. Further, learning may be optimized for a user-preferred type of inference by adjusting a learning criterion. An inference application may use the trained model to answer queries about variable(s) at any level of detail. In one embodiment, the inference application may determine scores for each possible value of the query variable by finding the best hierarchical event representation that maximizes a scoring function while fixing the value of the query variable to its possible values. Here, the inference application may approximately determine the best hierarchical event representation by iteratively optimizing one level-of-detail variable at a time while fixing other level-of-detail variables, until convergence.

A system and method for selecting at least one relay in a communication network; the network comprises a plurality of nodes; each of the nodes comprises at least one processing unit; each pair of the nodes is characterized by a first connection number; a first score is associated with each pair of (i) first first connection number; and (ii) second first connection number; each of the processing unit is programmed to execute the method.

Disclosed is a system and method for processing account registration by identifying account candidates attempting to open an account as abusive. That is, the present disclosure discusses identifying, and challenging and marking abusive account registration. The present disclosure takes into account users' behaviors on a network and the impact to the cost and/or revenue of the network. The present disclosure is proactive as it allows for actions to be taken at the earliest possible time in the registration process before an account is created. This prevents abusive activity from taking place within the network and effecting services and privileges available to legitimate users. Additionally, the effects of the disclosed systems and methods minimize the negative impacts of abusive activity on normal user accounts.

In one embodiment, a management system determines respective capability information of machine learning systems, the capability information including at least an action the respective machine learning system is configured to perform. The management system receives, for each of the machine learning systems, respective performance scoring information associated with the respective action, and computes a degree of freedom for each machine learning system to perform the respective action based on the performance scoring information. Accordingly, the management system then specifies the respective degree of freedom to the machine learning systems. In one embodiment, the management system comprises a management device that computes a respective trust level for the machine learning systems based on receiving the respective performance scoring feedback, and a policy engine that computes the degree of freedom based on receiving the trust level. In further embodiments, the machine learning system performs the action based on the degree of freedom.

Methods, systems, and computer storage media for implementing neural networks in fixed point arithmetic computing systems. In one aspect, a method includes the actions of receiving a request to process a neural network using a processing system that performs neural network computations using fixed point arithmetic; for each node of each layer of the neural network, determining a respective scaling value for the node from the respective set of floating point weight values for the node; and converting each floating point weight value of the node into a corresponding fixed point weight value using the respective scaling value for the node to generate a set of fixed point weight values for the node; and providing the sets of fixed point floating point weight values for the nodes to the processing system for use in processing inputs using the neural network.