Search

Nowadays, everyone requires a good quality software. The quality of software can't be assured due to lack of data availability for training, and testing. Thus, Transfer Learning (TL) plays an important role in the reusability of existing software for developing new software with a similar domain and task. TL focused on transferring knowledge from existing prediction models for the development of new prediction models. The developed models are used for unseen datasets based on the characteristics, and nature of the dataset. The sufficient amount of training data is unavailable. The data distribution and task of the source and target project must be checked before employing TL for software development. In this Systematic Review (SR), we have investigated 39 studies from January 1990 to March 2024 that used TL in the software engineering domain. The review focused on the identification of Machine Learning (ML) techniques used with TL techniques, types of TL explored, TL settings explored, experimental setting, dataset, quality attribute, validation methods, threats to validity, strengths and weakness of TL techniques, and hybrid techniques with TL. According to the experimental comparison, the performance of TL techniques is encouraging. The findings of this SR paper will serve as guidelines for academicians, software industry experts, software developers, software testers, and researchers. This SR is also helpful in the selection of appropriate types of TL and TL settings for the development of efficient software in the future based on the type of problem and TL setting. Thus, this study showed that 30.67% of the studies are focused on defect prediction, that used 15% open-source dataset. Further, 35% of studies used SVM as a base classifier for TL, and different independent variables of the used dataset are considered as prediction model input. Further, the K-fold Cross-Validation (CV) method is used in 15 studies. [ABSTRACT FROM AUTHOR]

In software engineering, cross-project defect prediction is an important area in the field of defect prediction and change prediction. Software defect prediction aims for the identification of defects in the early stages of the software development life cycle. Defect prediction helps in optimizing the resources in terms of testing and reduction of maintenance efforts. Defect prediction works appropriately if a huge amount of data is available for training a prediction model. Nowadays, a sufficient amount of data is not available due to which we have to use different data for training and testing of a project. The cross-project defect prediction idea emerged when different projects are used as training and testing dataset for the identification of defects. To design a defect prediction model we have to consider only significant features out of all the features set. Feature selection techniques are categorized based on unsupervised and supervised learning. The major limitation of cross-project defect prediction is handling different data distributions of source and target projects. The experiment was conducted using AEEEM and ReLink software defect dataset. Moreover, five projects of AEEEM and three projects of ReLink with a maximum count of files in the selected projects are 1862 and 399. In this study, we have analyzed the effect of feature selection techniques in cross-project defect prediction. The results were analyzed using AUC. The significance of filter, wrapper, and swarm search-based methods for feature selection techniques was analyzed separately. There is a trade-off between computational complexity and the performance of feature selection techniques. Swarm search-based methods performed better than filter and wrapper methods in terms of computational cost and overall performance of the prediction modes. The results were statistically validated using Friedman test and Wilcoxon signed rank test. [ABSTRACT FROM AUTHOR]

Software defect prediction (SDP) plays an important role in new research areas of software engineering. Cross-project defect prediction (CPDP) technique achieved success for prediction of defects in innovating projects having lack of data. In this study, we have developed predictive models using different machine learning algorithms. Two approaches for DP were discussed in this study. One approach is heterogeneous defect prediction (HDP), and another approach is within-project defect prediction (WPDP). We have used different projects in HDP. The metrics matching analyzer is used for identifying matching metrics across different projects. The concept of transfer learning helps to improve DP by using HDP. HDP takes one project for training and testing against another project. We have shown the best machine learning algorithm for the DP model. The DP helps us in identifying defects before its delivery to the customer. To achieve our goal, we have taken twelve datasets from NASA and PROMISE repository. The machine learning algorithms are compared using statistical tests. The performance of the developed prediction model has been evaluated using area under curve performance measure [ABSTRACT FROM AUTHOR]

The optimization of device dimensions along with high-k gate dielectric is investigated in this work for improving RF/analog performance of double gate (DG) TFET device. Through simulation, it is seen that RF/analog performance of DG-TFET deteriorate by the use of high-k gate materials in place of SiO2 as gate oxide. It has been anticipated that this degradation (ΔFoM = FoM(k = 3.9) – FoM(k = 40)) can be abated by optimizing the device dimensions particularly the channel thickness (Tc) and length (Lc). It is seen that deprivation in intrinsic dc gain (ΔAV), cut –off frequency (ΔfT), maximum oscillation frequency (ΔfMAX), gain-frequency product (ΔGFP), transconductance-frequency product (ΔTFP), and gain-transconductance-frequency product (ΔGTFP) is 11.3%, 11.84%, 11.17%, 52.16%, 19.69% & 56.42% respectively for Tc = 12 nm and 32.6%, 10.96%, 12.8%, 59.58%, 14.61% & 61.24% respectively for Tc = 6 nm. Furthermore, it is seen that deprivation in RF/analog parameters as a result of high-k gate dielectrics varies with channel length (Lc). The ΔAV, ΔfT, ΔfMAX, ΔGFP, ΔTFP, and ΔGTFP is observed to be 32.6%, 10.96%, 12.8%, 59.58%, 14.61% & 61.24% respectively for Lc = 30 nm and 17.2%, 7.41%, 10.36%, 58.67%, 8.95% & 59.36% respectively for Lc = 15 nm. Consequently, high-k gate dielectrics-based DG-TFET should be designed at thicker Tc and lower Lc for improved analog/RF performance of TFET device. [ABSTRACT FROM AUTHOR]

First-principles investigations are performed on magnetic tunnel junctions (MTJs) constructed with chromium trihalide CrCl3 as barrier and half metallic ferromagnet (HMF) CrO2 as electrodes using density function theory (DFT) calculations, and the results are compared with CrBr3 and CrI3. On comparing the results obtained after performing the first-principles investigation on CrCl3, CrBr3, and CrI3, it is understood that CrCl3 offers very high tunnel magnetoresistance (TMR) of ~ 100% and spin injection efficiency (η) of ~ 100% for different bias voltages, and spin transport occurs due to tunneling which is evident from the transmission spectrum. Although CrBr3 has TMR greater than 95% and spin injection efficiency (η) of ~ 100%, transmission states are present at the Fermi level, hence, tunneling is not the phenomenon in the case of CrBr3. CrI3 suffers from fluctuations in both TMR and spin injection efficiency (η) over the applied bias voltage range and has more spin-up current and spin-down current in both parallel configuration and anti-parallel configuration. In CrI3, as there are states in transmission spectrum at the Fermi level, the spin transport occurs because of band-to-band hopping instead of tunneling. From the results obtained, CrCl3 is identified as the perfect material to be used as the barrier in MTJs. [ABSTRACT FROM AUTHOR]

Using first-principles spin-polarized density functional theory computations, the effect of functionalization (fluorination and hydrogenation) on spin-polarized transport of gallium nitride (GaN) nanosheet–based magnetic tunnel junction (MTJ) with CrO2 as electrodes is investigated. The results show fluorinated GaN–based structure exhibits better spin filtration and high TMR (maximum ~ 99%), as compared with hydrogenated GaN (maximum TMR ~ 93%)– and pristine GaN (maximum TMR ~ 83%)–based structures. In addition, fluorinated GaN nanosheet exhibits a ferromagnetic behavior with a magnetic movement of 1.0 μb per fluorine atom. The magnetic movements for pristine GaN and hydrogenated GaN sheets are reported to be 0 μb and 0.9 μb per hydrogen atom, respectively. Higher TMR, better spin filtration, and ferromagnetic behavior for fluorinated GaN–based structure open up its possibility as spin filter (injector) in MTJs and other spin-based devices. [ABSTRACT FROM AUTHOR]

Density functional theory (DFT) is used to investigate the spin-dependent quantum transport through bended graphene. Bending results in reduced bandgap in graphene and affects the spin transport by increasing current in parallel configuration (PC) resulting in an increase in magnetoresistance (MR). In antiparallel configuration (APC), bending limits the spin-down current, which results in higher magnetoresistance at all biases. In bended graphene, the magnetoresistance obtained is higher than the MR obtained in pristine and twisted graphene-based structure. High spin filtration for PC and APC is observed in the case of bended graphene as compared with pristine and twisted graphene. However, pristine graphene gives better spin filtration compared with twisted graphene at low voltages. [ABSTRACT FROM AUTHOR]