Abstract: Methods and systems for training a neural combinatorial optimization (NCO) model having a processor and memory for performing a task having a target distribution. The NCO model is meta-trained to learn an efficient heuristic on a set of distributions. The meta-trained NCO model is then fine-tuned to specialize a learned heuristic for the target distribution.

Abstract: A training system includes: a neural network configured to, using trained parameters, generate a first encoding based on an input query and second encodings based on candidate responses for the input query; and a training module configured to: train the trained parameters using hyperparameters; and jointly optimize the hyperparameters using coordinate descent and line searching, the hyperparameters including: a first hyperparameter indicative of a first weight value to apply based on positive interactions of entries of a distance matrix based on encodings; a second hyperparameter indicative of a second weight value to apply based on negative interactions of entries of the distance matrix generated based on the first and second encodings; and a third hyperparameter corresponding to a dimension of the distance matrix generated based on the first and second encodings.

Abstract: A method is disclosed for estimating a trajectory of an object on a map given a sequence of traces for the moving object. Each trace of the object including information defining a position measured at a given time for the object, as well as information as to an area of accuracy around the measured position. The method processes pairs of successive traces, corresponding to two positions successive in time in the sequence of measured positions for the moving object. For each trace of a pair of successive traces, the method defines road segments on the map within the area of accuracy of the trace. For each road segment within the area of accuracy of a first trace of a pair of traces and each road segment within the area of accuracy of the second trace of the pair, the method determines at least one candidate path between the two road segments. A neural network and a neural graph model are used to compute the most probable sequence of candidate paths to estimate the trajectory of the object on the map.

Abstract: In a method for generating a normalized sequential model using a processor, a sequential energy-based model computed by a parameterized neural network is provided. The sequential energy-based model defines an unnormalized probability distribution over a target sequence for a context source. The normalized sequential model is generated by projecting the sequential energy-based model onto a target autoregressive model that approximates a normalized distribution associated with the sequential energy-based model.

Abstract: A method is disclosed for estimating a trajectory of an object on a map given a sequence of traces for the moving object. Each trace of the object including information defining a position measured at a given time for the object, as well as information as to an area of accuracy around the measured position. The method processes pairs of successive traces, corresponding to two positions successive in time in the sequence of measured positions for the moving object. For each trace of a pair of successive traces, the method defines road segments on the map within the area of accuracy of the trace. For each road segment within the area of accuracy of a first trace of a pair of traces and each road segment within the area of accuracy of the second trace of the pair, the method determines at least one candidate path between the two road segments. A neural network and a neural graph model are used to compute the most probable sequence of candidate paths to estimate the trajectory of the object on the map.

Abstract: A power meter generates a time series of observations of power drawn by a device configured to operate in an active mode and to transition through a sequence of inactive modes comprising one or more intermediate inactive modes each having a bounded time duration terminating in a final inactive mode of unbounded time duration, and further configured to transition from any of the inactive modes to the active mode in response to a device activation signal. An electronic data processing device is programmed to determine statistical characteristics of the power draw of the device in each of the inactive modes by fitting a state model to the time series of observations. The state model represents each intermediate inactive mode including its bounded time duration and further includes a probabilistic representation of the transition from any of the inactive modes to the active mode in response to a device activation signal.

Abstract: A method for inferring user activity statistics includes receiving job logs of a device infrastructure. Each job log including job information for a job performed for one of a set of users by one of a plurality of shared devices. A job descriptor is generated for each job based on the job log. The job is assigned to one of a set of defined job types based on the job descriptor. An activity matrix is composed where activity of each user for each of a plurality of time periods is represented as a fixed-length vector representing at least some of the job types. Each index of the vector is derived from a count of one job type for the user in that time period. The activity matrix is decomposed into first and second factor matrices to minimize an overall reconstruction error. User activity statistics are output based on the decomposition.

Abstract: A system and method predict the translation quality of a translated input document. The method includes receiving an input document pair composed of a plurality of sentence pairs, each sentence pair including a source sentence in a source language and a machine translation of the source language sentence to a target language sentence. For each of the sentence pairs, a representation of the sentence pair is generated, based on a set of features extracted for the sentence pair. Using a generative model, a representation of the input document pair is generated, based on the sentence pair representations. A translation quality of the translated input document is computed, based on the representation of the input document pair.

Abstract: A system and method predict the translation quality of a translated input document. The method includes receiving an input document pair composed of a plurality of sentence pairs, each sentence pair including a source sentence in a source language and a machine translation of the source language sentence to a target language sentence. For each of the sentence pairs, a representation of the sentence pair is generated, based on a set of features extracted for the sentence pair. Using a generative model, a representation of the input document pair is generated, based on the sentence pair representations. A translation quality of the translated input document is computed, based on the representation of the input document pair.

Abstract: Methods and systems input an energy consumption profile for each of a plurality of different sleep modes available for a device, and input a probability distribution of interjob times for the device. The methods and systems then compute the optimal time-out period for each sleep mode based on the energy consumption profile of each sleep mode and the probability distribution of interjob times. Further, such methods and systems monitor the usage of the device to determine the current interjob time, and switch between sleep modes to relatively lower power sleep modes as the current interjob time becomes larger.

Abstract: A power meter generates a time series of observations of power drawn by a device configured to operate in an active mode and to transition through a sequence of inactive modes comprising one or more intermediate inactive modes each having a bounded time duration terminating in a final inactive mode of unbounded time duration, and further configured to transition from any of the inactive modes to the active mode in response to a device activation signal. An electronic data processing device is programmed to determine statistical characteristics of the power draw of the device in each of the inactive modes by fitting a state model to the time series of observations. The state model represents each intermediate inactive mode including its bounded time duration and further includes a probabilistic representation of the transition from any of the inactive modes to the active mode in response to a device activation signal.

Abstract: A control system reduces energy consumption in a multi-device system comprising a plurality of devices. The control system includes at least one processor. The processor is programmed to receive a job to be executed, as well as a selection of one of the plurality of devices for executing the job and a transfer cost for transferring the job from the selected device to each of the plurality of devices. A device to execute the job is determined through optimization of a first cost function. The first cost function is based on the device selection and the transfer costs. The job is assigned to the determined device and a time-out for each device in the multi-device system is determined through optimization of a second cost function. The second cost function is based on an expected energy consumption by the multi-device system. The devices are provided with the determined time-outs.

Abstract: Multi-task regression or classification includes optimizing parameters of a Bayesian model representing relationships between D features and P tasks, where D?1 and P?1, respective to training data comprising sets of values for the D features annotated with values for the P tasks. The Bayesian model includes a matrix-variate prior having features and tasks dimensions of dimensionality D and P respectively. The matrix-variate prior is partitioned into a plurality of blocks, and the optimizing of parameters of the Bayesian model includes inferring prior distributions for the blocks of the matrix-variate prior that induce sparseness of the plurality of blocks. Values of the P tasks are predicted for a set of input values for the D features using the optimized Bayesian model. The optimizing also includes decomposing the matrix-variate prior into a product of matrices including a matrix of reduced rank in the tasks dimension that encodes correlations between tasks.

Abstract: Methods and systems input an energy consumption profile for each of a plurality of different sleep modes available for a device, and input a probability distribution of interjob times for the device. The methods and systems then compute the optimal time-out period for each sleep mode based on the energy consumption profile of each sleep mode and the probability distribution of interjob times. Further, such methods and systems monitor the usage of the device to determine the current interjob time, and switch between sleep modes to relatively lower power sleep modes as the current interjob time becomes larger.

Abstract: A system and method of localizing elements (shared devices and/or their users) in a device infrastructure, such as a printing network, are provided. The method includes mapping a structure in which the elements of a device infrastructure are located, the elements comprising shared devices and users of the shared devices. Probable locations of fewer than all of the elements in the structure are mapped, with at least some of the elements being initially assigned to an unknown location. Usage logs for a plurality of the shared devices are acquired. The acquired usage log for each device includes a user identifier for each of a set of uses of the device, each of the uses being initiated from a respective location within the mapped structure by one of the users. Based on the acquired usage logs and the input probable locations of some of the elements, locations of at least some of the elements initially assigned to an unknown location are predicted.

Abstract: An apparatus operating on a time sequence of events includes an event handling module configured to generate a predicted label for a current observed event of the time sequence of events and a true label handling module configured to process a true label revealed for an observed event of the time sequence of events. The event handling module and the true label handling module cooperatively model stochastic dependence of a true label for the current observed event based on the time sequence of events and revealed true labels for the past observed events of the time sequence of events. The event handling module and the true label handing module operate asynchronously. The event handling module and the true response handling module suitably operate as one or more digital processors.

Abstract: Multi-task regression or classification includes optimizing parameters of a Bayesian model representing relationships between D features and P tasks, where D?1 and P?1, respective to training data comprising sets of values for the D features annotated with values for the P tasks. The Bayesian model includes a matrix-variate prior having features and tasks dimensions of dimensionality D and P respectively. The matrix-variate prior is partitioned into a plurality of blocks, and the optimizing of parameters of the Bayesian model includes inferring prior distributions for the blocks of the matrix-variate prior that induce sparseness of the plurality of blocks. Values of the P tasks are predicted for a set of input values for the D features using the optimized Bayesian model. The optimizing also includes decomposing the matrix-variate prior into a product of matrices including a matrix of reduced rank in the tasks dimension that encodes correlations between tasks.

Abstract: A control system reduces energy consumption in a multi-device system comprising a plurality of devices. The control system includes at least one processor. The processor is programmed to receive a job to be executed, as well as a selection of one of the plurality of devices for executing the job and a transfer cost for transferring the job from the selected device to each of the plurality of devices. A device to execute the job is determined through optimization of a first cost function. The first cost function is based on the device selection and the transfer costs. The job is assigned to the determined device and a time-out for each device in the multi-device system is determined through optimization of a second cost function. The second cost function is based on an expected energy consumption by the multi-device system. The devices are provided with the determined time-outs.

Abstract: In a monitoring method, a time sequence of information pertaining to a monitored device, network, or system is recorded, comprising observations of the monitored device, network, or system and known prior correct action recommendations for the monitored device, network, or system. A hidden Markov model (HMM) operating on the time sequence of information is maintained. The HMM comprises a hidden state of the monitored device, network, or system. A current state of the monitored device, network, or system is classified using a classification value comprising an emission of the HMM that depends on an estimate of the distribution of the hidden state and on a selected portion of the time sequence of information. An action recommendation is generated for the current state of the monitored device, network, or system based on the classification value.

Abstract: An events analysis method comprises: optimizing respective to a set of training data a set of branching transition likelihood parameters associating parent events of type k with child events of type k? in branching processes; inferring a most probable branching process for a set of input data comprising events based on the optimized set of branching transition likelihood parameters; and identifying rare or unusual events of the set of input data based on the inferred most probable branching process. An events analysis apparatus includes a probabilistic branching process learning engine configured to optimize the set of branching transition likelihood parameters, and a probabilistic branching process inference engine configured to infer the most probable branching process.