Abstract: A mechanism is provided in a data processing system for performing regression testing on a question answering system instance. The mechanism trains a machine learning model for a question answering system using a ground truth virtual checksum as part of a ground truth including domain-specific ground truth. The ground truth virtual checksum comprises a set of test questions, an answer to each test question, and a confidence level range for each answer to a corresponding test question. The mechanism runs regression test buckets across system nodes with domain-specific corpora and receiving results from the system nodes. Each system node implements a question answering system instance of the question answering system by executing in accordance with the machine learning model and by accessing domain-specific corpora. Each test bucket includes a set of questions matching a subset of questions in the ground truth virtual checksum.

Abstract: Described is a system for decoding spiking reservoirs even when the spiking reservoir has continuous synaptic plasticity. The system uses a set of training patterns to train a neural network having a spiking reservoir comprised of spiking neurons. A test pattern duration d is estimated for a set of test patterns P, and each test pattern is presented to the spiking reservoir for a duration of d/P seconds. Output spikes from the spiking reservoir are generated via readout neurons. The output spikes are measured and the measurements are used to compute firing rate codes, each firing rate code corresponding to a test pattern in the set of test patterns P. The firing rate codes are used to decode performance of the neural network by computing a discriminability index (DI) to discriminate between test patterns in the set of test patterns P.

Abstract: Described is a neuromorphic system implemented in hardware that implements neuron membrane potential update based on the leaky integrate and fire (LIF) model. The system further models synapse weights update based on the spike time-dependent plasticity (STDP) model. The system includes an artificial neural network in which the update scheme of neuron membrane potential and synapse weight are effectively defined and implemented.

Abstract: Systems and methods for embodiments of a graph based artificial intelligence systems for identity management are disclosed. Embodiments of the identity management systems disclosed herein may utilize a network graph approach to analyzing identities or entitlements of a distributed networked enterprise computing environment. Specifically, in certain embodiments, an artificial intelligence based identity management systems may utilize the peer grouping of an identity graph (or peer grouping of portions or subgraphs thereof) to identify roles from peer groups or the like.

Abstract: Provided is a process, including: obtaining a set of historical geolocations; segmenting the historical geolocations into a plurality of temporal bins; determining pairwise transition probabilities between a set of geographic places based on the historical geolocations; configuring a compute cluster by assigning subsets of the transition probabilities to computing devices in the compute cluster; receiving a geolocation stream indicative of current geolocations of individuals; selecting a computing device in the compute cluster in response to determining that the computing device contain transition probabilities for the received respective geolocation; selecting transition probabilities applicable to the received respective geolocation from among the subset of transition probabilities assigned to the selected computing device; predicting a subsequent geographic place based on the selected transition probabilities.

Abstract: A method, computer program product, and system perform computations using a sparse convolutional neural network accelerator. A first vector comprising only non-zero weight values and first associated positions of the non-zero weight values within a 3D space is received. A second vector comprising only non-zero input activation values and second associated positions of the non-zero input activation values within a 2D space is received. The non-zero weight values are multiplied with the non-zero input activation values, within a multiplier array, to produce a third vector of products. The first associated positions are combined with the second associated positions to produce a fourth vector of positions, where each position in the fourth vector is associated with a respective product in the third vector. The products in the third vector are transmitted to adders in an accumulator array, based on the position associated with each one of the products.

Abstract: A testing framework associated with a decision metaphor model tool reads table profile files to generate requests for a test of a decision metaphor. The testing framework sends the requests for the test to a decision engine and receives responses for the requests for comparison against expected values and possible errors. The testing framework also outputs an output file that includes a result of the test, where the output file is formatted in a computer-displayable and user-readable graphical format.

Abstract: This disclosure relates to predicting and managing supply chain network risks. In one embodiment, a processor-implemented method obtains identifiers for supply chain contributors and parameters; and a query. The method performs a natural language processing algorithm on the query to extract text components, which it analyzes to identify supply chain component clusters and risk identifiers. It also includes executing a machine learning technique for learning of the risk identifiers and generating co-occurrence rules between the risk identifiers, as well as associated rule support and rule confidence parameters. It further includes sorting the co-occurrence rules to generate a prioritized rules list, and generating a risk prediction model for the supply chain using the prioritized rules list, using a classifier algorithm.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training and applying a machine learning model. One of the methods includes the actions of obtaining a collection of training data, the training data comprising collection of data points associated with a labeled set of real property parcels; training a machine learning model using the training data, the machine learning model being trained to generate a likelihood with respect to a parameter from input data associated with a specific parcel of real property, wherein training includes optimizing the model using a Markov chain optimization that seeks to minimize error in the model where the model is underpinned by one or more non-differentiable functions; receiving a plurality of data points associated with an input parcel of real property; and using the optimized model to generate a likelihood for the parameter for the input parcel of real property.

Abstract: A speech recognition neural network system includes an encoder neural network and a decoder neural network. The encoder neural network generates an encoded sequence from an input acoustic sequence that represents an utterance. The input acoustic sequence includes a respective acoustic feature representation at each of a plurality of input time steps, the encoded sequence includes a respective encoded representation at each of a plurality of time reduced time steps, and the number of time reduced time steps is less than the number of input time steps. The encoder neural network includes a time reduction subnetwork, a convolutional LSTM subnetwork, and a network in network subnetwork. The decoder neural network receives the encoded sequence and processes the encoded sequence to generate, for each position in an output sequence order, a set of substring scores that includes a respective substring score for each substring in a set of substrings.

Abstract: A machine learning device which learns to determine at least one arc welding condition includes a state observation unit which observes a state variable consisting of at least one physical quantity regarding the arc welding and the at least one arc welding condition at least during or after the arc welding, and a learning unit which learns a change in the at least one physical quantity observed by the state observation unit and the at least one arc welding condition in association with each other.

Abstract: The present disclosure describes a system, method, and computer program for classifying monitored assets based on user labels and for detecting potential misuse of monitored assets based on said classifications. Machine-learning-based modeling is used to classify one or more types of monitored assets with a select user label. A data model is created that reflects monitored assets used by users associated with the select user label. Each a time a user with the select user label accesses an applicable type of monitored asset, the data model is updated to reflect the event. The data model is used to classify one or more monitored assets with the select user label. If a user without the select user label uses a monitored asset classified with the select user label, a potential misuse of the monitored asset is detected.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for system modernization using machine learning are disclosed. In one aspect, a method includes the actions of generating training data. The actions include generating a first model, a second model, and a third model. The actions include receiving data that is related to technological capabilities of an application and data that is related to business priorities of the application. The actions include applying the first model to the data that is related to the technological capabilities of the application and the second model to the data that is related to business priorities of the application. The actions include generating a modification recommendation for the application. The actions further include providing, for output, the modification recommendation for the application. The actions include receiving feedback data that indicates a level of acceptance of the modification recommendation for the application.

Abstract: A machine learning model is trained by defining a scenario including models of vehicles and a typical driving environment. A model of a subject vehicle is added to the scenario and sensor locations are defined on the subject vehicle. A perception of the scenario by sensors at the sensor locations is simulated. The scenario further includes a model of a lane-splitting vehicle. The location of the lane-splitting vehicle and the simulated outputs of the sensors perceiving the scenario are input to a machine learning algorithm that trains a model to detect the location of a lane-splitting vehicle based on the sensor outputs. A vehicle controller then incorporates the machine learning model and estimates the presence and/or location of a lane-splitting vehicle based on actual sensor outputs input to the machine learning model.

Abstract: Various embodiments provide a coupling mechanism, method of activation and a square lattice. The coupling mechanism comprises two qubits and a tunable coupling qubit that activates an interaction between the two qubits by modulation of a frequency of the tunable coupling qubit. The tunable coupling qubit capacitively couples the two qubits. The tunable coupling qubit is modulated at a difference frequency of the two qubits. The difference frequency may be significantly larger than an anharmonicity of the two qubits. The tunable coupling qubit may be coupled to the two qubits by two electrodes separated by a superconducting quantum interference device (SQUID) loop having two Josephson junctions or by a single electrode with a SQUID loop coupling to ground. The SQUID loop is controlled by an inductively-coupled flux bias line positioned at the center of the tunable coupling qubit.

Abstract: Techniques are disclosed for automatically assigning weights to attributes of media content based in part on how many users actually viewed or listened to the content, as well as how many users “liked” or otherwise indicated a preference for the content. The content items can be any type of audio or visual media content, such as songs, videos, or movies, as well as written content, such as books, articles, journals, advertisements, or magazines. A first similarity score is determined based on a similarity between user preferences for content items. A second similarity score is determined based on a similarity between one or more common attributes of the content items. These attributes are assigned ratings that represent the number of users who consumed the corresponding content. Next, weights are assigned to each of the attributes based on the first and second similarity scores using, for example, linear equation regression techniques.

Abstract: A machine learning device, which learns shocks to a teaching device, includes a state observation unit which observes data based on an inclination of the teaching device or a present position of the teaching device; a label obtaining unit which obtains a label based on a shock received by the teaching device; and a learning unit which generates a learning model based on an output of the state observation unit and an output of the label obtaining unit.

Abstract: The disclosed embodiments illustrate methods and systems for predicting requirements of a user for resources. The method includes transforming a message posted by the user into a first message vector. The method further includes categorizing one or more first message vectors into one or more categories. The method further includes transforming each of the categorized first message vectors into one or more second message vectors using a wavelet transform technique. The method further includes determining, for each of the categorized first message vectors, a first score based on at least a probability distribution of one or more coefficients associated with each associated feature. The method further includes selecting a predefined number of features based on at least the first score. The method further includes training one or more classifiers on the selected predefined number of features to identify at least the one or more needs of the user.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating parse trees for input text segments. One of the methods includes obtaining an input text segment, processing the input text segment using a first long short term memory (LSTM) neural network to convert the input text segment into an alternative representation for the input text segment, and processing the alternative representation for the input text segment using a second LSTM neural network to generate a linearized representation of a parse tree for the input text segment.

Abstract: Systems and methods for training a neural network or an ensemble of neural networks are described. A hyper-parameter that controls the variance of the ensemble predictors is used to address overfitting. For larger values of the hyper-parameter, the predictions from the ensemble have more variance, so there is less overfitting. This technique can be applied to ensemble learning with various cost functions, structures and parameter sharing. A cost function is provided and a set of techniques for learning are described.