2nd Mini Workshop of Ideas!
Virtual Meeting - April 29th, 2020
Title: Mapping graph coloring to quantum annealingAbstract:
Adiabatic quantum computation (AQC) is an alternative to the gate model quantum computation and operates in a working graph such as the Chimera of the D-Wave Systems. These physical systems are suitable for quantum annealing (QA) which is a metaheuristic capable to find a global minimum of a given objective function from a set of candidate solutions. As such, problems like graph coloring can be mapped to quantum hardware resorting, for example, to pseudo-Boolean constraints. The SATyrus approach allows us to convert the constraints to an energy minimization problem which can be translated to a quadratic unconstrained binary optimization (QUBO) problem by applying polynomial reduction, a proper formulation for the D-Wave machine. The problem was implemented using three different approaches: (1) classical, QUBO using simulated annealing (SA) in a von neumann machine, (2) quantum 1, QUBO using the D-Wave quantum machine and reducing polynomial degree using a D-Wave library, and (3) quantum 2, QUBO using the D-Wave quantum machine and reducing polynomial degree using our own implementation. We conclude that (a) the variation of penalty terms did not have a significant impact in the number of solutions found, (b) QA produce better heuristics for this specific problem than SA since we consistently obtained always more solutions with the QA approach.
A Novel Ensemble Method For ILP and Classification Models
In Search of Cloud Computing Efficiencies for Data Science Workloads
|15:00||Andrť Rodrigues||Title: Scalable Bayesian Networks|
Title: MultimodalDermaCAD - Classification of multimodal dermatological data
Title: Artificial Neural Networks Comparative Performance Study
Title: A Web Application for the Assessment of Breast Cancer Conservative Treatment Cosmetic Results
Title: Early fault detection and mitigation in complex systems
Complex systems experience numerous faults daily, many of these faults are none systems impacting. However, some faults are system impacting and some faults experience a propagation beyond a given system and effect the entire complex system. Understanding these failures and attempting to predict through mitigation factors can greatly enhance total system stability. Through this work I walk through system fault analysis and propose a few mitigation techniques for reducing impact of complete system effecting faults.
|16:15||Josť Carvalho||Title: Structure Learning in Neural Networks|