The paper focuses on Cross-Version Defect Prediction (CVDP) where the classification model is trained on information of the prior version and then tested to predict defects in the components of the last release. To avoid the distribution differences which could negatively impact the performances of machine learning based model, we consider Dissimilarity-based Sparse Subset Selection (DS3) technique for selecting meaningful representatives to be included in the training set. Furthermore, we employ a Convolutional Neural Network (CNN) to generate structural and semantic features to be merged with the traditional software measures to obtain a more comprehensive list of predictors. To evaluate the usefulness of our proposal for the CVDP scenario, we perform an empirical study on a total of 20 cross-version pairs from 10 different software projects. To build prediction models we consider Logistic Regression (LR) and Random Forest (RF) and we adopt 3 evaluation criteria (i.e., F-measure, G-mean, Balance) to assess the prediction accuracy. Our results show that the use of CNN with both LR and RF models has a significant impact, with an improvement of ∼20% for each evaluation criteria. Differently, we notice that DS3does not impact significantly in improving prediction accuracy.

Nanoscale light sources are being intensively investigated for their potential to enable low-energy, high-density optical communication and sensing systems. Both nanolight-emitting diodes (nanoLEDs) and nanolasers have been considered, based on advanced nanophotonic concepts such as photonic crystals and plasmonic structures, with dimensions well into the submicrometer domain. With decreasing dimensions, light-matter interaction becomes stronger, potentially leading to efficient and ultrafast radiative emission, both in the spontaneous and stimulated regimes. These features have created wide expectations for the practical prospects of such nanoscale light sources, in particular for optical interconnects. In this paper, we examine the limits to the downscaling of LEDs and lasers and ask ourselves which type of source is most suited to ultralow-power optical communications. Based on simple physical considerations on the scaling of spontaneous and stimulated emission rates for semiconductor active regions at room temperature, we analyze the speed and energy limits for nanoLEDs and nanolasers as a function of their size. The role of spontaneous emission enhancement (Purcell effect) in practical nanophotonic sources is also revisited. The main conclusion is that nanoLEDs reach a fundamental energy/speed limit for data rates exceeding a few gigahertz per second, whereas nanolasers with active dimensions in the range of few hundred nanometers may enable direct modulation rates larger than 40 Gb/s at power levels adequate for short-distance and low-energy optical interconnects.