Photonic-aware neural networks

Neural Computing and Applications - Photonics-based neural networks promise to outperform electronic counterparts, accelerating neural network computations while reducing power consumption and...     

A decision support system for optimal deployment of sonobuoy networks based on sea current forecasts and multi-objective evolutionary optimization

A decision support system for the optimal deployment of drifting acoustic sensor networks for cooperative track detection in underwater surveillance applications is proposed and tested on a simulated scenario. The system integrates sea water current forecasts, sensor range models and simple drifting buoy kinematic models to predict sensor positions and temporal network performance. A multi-objective genetic optimization algorithm is used for searching a set of Pareto optimal deployment solutions (i.e. the initial position of drifting sonobuoys of the network) by simultaneously optimizing two quality of service metrics: the temporal mean of the network area coverage and the tracking coverage. The solutions found after optimization, which represent different efficient tradeoffs between the two metrics, can be conveniently evaluated by the mission planner in order to choose the solution with the desired compromise between the two conflicting objectives. Sensitivity analysis through the Unscented Transform is also performed in order to test the robustness of the solutions with respect to network parameters and environmental uncertainty. Results on a simulated scenario making use of real probabilistic sea water current forecasts are provided showing the effectiveness of the proposed approach. Future work is envisioned to make the tool fully operational and ready to use in real scenarios.     

Computer-aided detection of lung nodules based on decision fusion techniques

We adopted decision fusion techniques to develop a computer-aided detection (CAD) system for automatic detection of pulmonary nodules in low-dose CT images. Two distinct phases, aimed, respectively, at detecting volumes of interests (VOIs) within the CT scan, and at classifying VOIs into nodules and non-nodules, were considered. Three algorithms, namely thresholding, region growing and robust fuzzy clustering, were used as VOI detectors. For the classification phase, we built multi-classifier systems, which aggregate the decisions of three statistical classifiers, a neural network and a decision tree. Finally, the receiver operating characteristic convex hull method was used to build the final classifier, which results to be the aggregation of the best local behaviors of both classifiers and combiners. All the CAD modules were tested on CT scans analyzed by two expert radiologists. In the experiments, we achieved a sensitivity of 92.5% against a specificity of 83.5%.     

A Pareto-based multi-objective evolutionary approach to the identification of Mamdani fuzzy systems

In the last years, the numerous successful applications of fuzzy rule-based systems (FRBSs) to several different domains have produced a considerable interest in methods to generate FRBSs from data. Most of the methods proposed in the literature, however, focus on performance maximization and omit to consider FRBS comprehensibility. Only recently, the problem of finding the right trade-off between performance and comprehensibility, in spite of the original nature of fuzzy logic, has arisen a growing interest in methods which take both the aspects into account. In this paper, we propose a Pareto-based multi-objective evolutionary approach to generate a set of Mamdani fuzzy systems from numerical data. We adopt a variant of the well-known (2+2) Pareto Archived Evolutionary Strategy ((2+2)PAES), which adopts the one-point crossover and two appropriately defined mutation operators. (2+2)PAES determines an approximation of the optimal Pareto front by concurrently minimizing the root mean squared error and the complexity. Complexity is measured as sum of the conditions which compose the antecedents of the rules included in the FRBS. Thus, low values of complexity correspond to Mamdani fuzzy systems characterized by a low number of rules and a low number of input variables really used in each rule. This ensures a high comprehensibility of the systems. We tested our version of (2+2)PAES on three well-known regression benchmarks, namely the Box and Jenkins Gas Furnace, the Mackey-Glass chaotic time series and Lorenz attractor time series datasets. To show the good characteristics of our approach, we compare the Pareto fronts produced by the (2+2)PAES with the ones obtained by applying a heuristic approach based on SVD-QR decomposition and four different multi-objective evolutionary algorithms.     

Estimating the concentration of optically active constituents of sea water by Takagi-Sugeno models with quadratic rule consequents

Determining the concentrations of chlorophyll, suspended particulate matter and coloured dissolved organic matter in the sea water is basic to support the monitoring of upwelling phenomena, algae blooms, and changes in the marine ecosystem. Since these concentrations affect the spectral distribution of the solar light back-scattered by the water body, their estimation can be computed by using a set of remotely sensed multispectral measurements of the reflected sunlight. In this paper, the relation between the concentrations of interest and the average subsurface reflectances is modelled by means of a set of second-order Takagi-Sugeno (TS) fuzzy rules. Unlike first-order TS rules, which adopt linear functions as consequent, second-order TS rules exploit quadratic functions, thus improving the modelling capability of the rule in the subspace determined by the antecedent. First, we show how we can build a second-order TS model through a simple transformation, which allows estimating the consequent parameters using standard linear least-squares algorithms, and by adopting one of the most used methods proposed in the literature to generate first-order TS models. Then, we compare first-order and second-order TS models against mean square error and interpretability of rules. We highlight how second-order TS models allow us to achieve better approximation than first-order TS models though maintaining interpretability of the rules. Finally, we show how second-order TS models perform considerably better (the mean square error is lower by two orders of magnitude) than the specific implementations of radial basis function networks and multi-layer perceptron networks used in previous papers for the same application domain.     

On reducing computational overhead in multi-objective genetic Takagi–Sugeno fuzzy systems

The use of multi-objective evolutionary algorithms (MOEAs) to generate a set of fuzzy rule-based systems (FRBSs) with different trade-offs between complexity and accuracy has gained more and more interest in the scientific community. The evolutionary process requires, however, a large number of FRBS generations and evaluations. When we deal with high dimensional datasets, these tasks can be very time-consuming, especially when we generate Takagi–Sugeno FRBSs, thus making a satisfactory exploration of the search space very awkward. In this paper, we first analyze the time complexity for both the generation and the evaluation of Takagi–Sugeno FRBSs. Then we introduce a simple but effective technique for speeding up the identification of the rule consequent parameters, one of the most time-consuming phases in Takagi–Sugeno FRBS generation. Finally, we highlight how the application of this technique produces as a side-effect a decoupling of the rules. This decoupling allows us to avoid re-computing consequent parameters of rules which are not directly modified during the evolutionary process, thus saving a considerable amount of time.In the experimental part we first test the correctness of the predicted asymptotical time complexity. Then we show the benefits in terms of computing time saving and improved search space exploration through an example of multi-objective genetic learning of Takagi–Sugeno FRBSs in the regression domain.     

Approaching the ocean color problem using fuzzy rules

In this paper, we propose a fuzzy logic-based approach which exploits remotely sensed multispectral measurements of the reflected sunlight to estimate the concentration of optically active constituents of the sea water. The relation between the concentrations of interest and the subsurface reflectances is modeled by a set of fuzzy rules extracted automatically from the data through a two-step procedure. First, a compact initial rule base is generated by projecting onto the input variables the clusters produced by a fuzzy clustering algorithm. Then, a genetic algorithm is applied to optimize the rules. Appropriate constraints maintain the semantic properties of the initial model during the genetic evolution. Results of the application of the fuzzy model obtained from data simulated with an ocean color model over the channels of the Medium Resolution Imaging Spectrometer are shown and discussed.