There is a need for more effective and efficient constrained-optimization-based operational load balancing. In one example, a method comprises determining constraint-satisfying operator-unit mapping arrangements that satisfy an operator unity constraint and an operator capacity constraint; for each constraint-satisfying operator-unit mapping arrangement, determining an arrangement utility measure; processing each arrangement utility measure using an optimization-based ensemble machine learning model that is configured to determine an optimal operator-unit mapping arrangement of the plurality of constraint-satisfying operator-unit mapping arrangements; and performing one or more operational load balancing operations based on the optimal operator-unit mapping arrangement.

Concepts and technologies disclosed herein are directed to interpretation workflows for machine learning-enabled event tree-based diagnostic and customer problem resolution. According to one aspect, a system can receive a workflow construction specification derived from a machine learning-enabled event tree (“MLET”). The MLET can be generated for use by a customer service agent to resolve a customer problem. The workflow construction specification can include a plurality of objects, each of which represents a navigation path through the MLET. The system can traverse the workflow construction specification and can create a set of workflow creation commands based upon at least one policy. The system can generate a workflow visualization interpretation file based upon the set of workflow creation commands. The workflow visualization interpretation file can identify how the MLET derived a root cause of the customer problem. The system can then present the workflow visualization interpretation file to the customer service agent.

Systems, methods, and non-transitory computer-readable media can construct a simulation framework for a ride sharing service. The simulation framework comprises a simulation environment and an agent comprising one or more algorithms including an order dispatching algorithm and a driver reposition algorithm. One or more states of the simulation environment include information about a plurality of drivers and a plurality of trip order requests, and can be provided to the agent. One or more actions from the agent can be obtained. The one or more actions comprises at least one of: a plurality of matches between the plurality of drivers and the plurality of trip order requests, or a plurality of reposition destinations for a subset of the plurality of drivers. The one or more states of the simulation environment can be updated based on the one or more actions.

A multi-objective scheduling advisor for generating a multi-stop visitation schedule includes generating, by a computer, a road network map corresponding to a predetermined area including a plurality of tasks locations. A task to be performed is assigned to each of the plurality of task locations. The computer calculates a business value for each task location using at least one of a calculation, a selected business rule applied to a delay duration, and an input value received from a client. A duration of a respective task is calculated using historical data on task durations associated with a staff of operators over different predetermined areas to determine an average task duration for every operator in the staff. Finally, using a metaheuristic binary optimization algorithm, the computer chooses different candidate tasks for the multi-stop visitation schedule to visit multiple assets in a single trip within the predetermined area.

Embodiments employ an inference method for neural networks that enforces deterministic constraints on outputs without performing post-processing or expensive discrete search over the feasible space. Instead, for each input, the continuous weights are nudged until the network's unconstrained inference procedure generates an output that satisfies the constraints. This is achieved by expressing the hard constraints as an optimization problem over the continuous weights and employing backpropagation to change the weights of the network. Embodiments optimize over the energy of the violating outputs; since the weights directly determine the output through the energy, embodiments are able to manipulate the unconstrained inference procedure to produce outputs that conform to global constraints.

An attentional neural machine translation model is provided for the task of machine translation that, according to some embodiments, leverages the hierarchical structure of language to perform natural language processing without a priori annotation. Other features are also provided.

Abnormalities in electromagnetic fields in the heart, brain, and stomach, among other organs and tissues of the human body, can be indicative of serious health conditions. Described herein are methods, software, systems and devices for detecting the presence of an abnormality in an organ or tissue of a subject by analysis of the electromagnetic fields generated by the organ or tissue.

A computer-implemented method for predicting a conformation of a ligand docked into a protein is disclosed. According to some embodiments, the method may include determining one or more poses of the ligand in the protein, the poses being representative conformations of the ligand. The method may also include determining, using a neural network, energy scores of the poses. The method may further include determining a proper conformation for the docked ligand based on the energy scores.

First, a dual basis B.sup.− of B of modulo N is obtained by classical computation. Next, quantum computation is performed using the periodicity of a point sequence included in a sum set of sets obtained by parallel translation of a lattice L(B) by integral multiples of t for a plurality of integers, and an n-dimensional r.sub.j=(r.sub.j1, . . . , r.sub.jn) and r.sub.j0 are obtained for j=1, . . . , m. Subsequently, by classical computation, the closest vector r.sub.j.sup.(c)=(r.sub.j1.sup.(c), . . . , r.sub.jn.sup.(c))∈L(B.sup.−) of the n-dimensional vector r.sub.j, and the difference vector r.sub.j.sup.(d)=r.sub.j−r.sub.j.sup.(c)=(r.sub.j1.sup.(d), . . . , r.sub.jn.sup.(d)) corresponding to r.sub.j.sup.(c) are obtained.

A method, system, and computer program product are provided for generating a predictive model. A processor(s) obtains a raw data set (peptide libraries) of patients designated as diagnosed/pre-diagnosed with a condition or not diagnosed with the condition. The processor(s) segments the raw data set into a pre-defined number of groups and separates out a holdout group. The processor(s) performs a principal component analysis on the remaining groups to identify, based on a frequency of features in the remaining groups, common features (principal components) in the remaining groups and weighs the common features based on frequency of occurrence. The processor(s) determines a smallest number of the principal components that yields a pre-defined level of validation accuracy. The processor(s) generates a predictive model, by utilizing the smallest number for a best fit in a logistic regression model. The predictive model provides binary outcomes.