Hyperparameter Optimization and Neural Architecture Search are powerful in attaining state-of-the-art machine learning models, with Bayesian Optimization (BO) standing out as a mainstream method. Extending BO into the multi-fidelity setting has been an emerging research topic in this field, but faces the challenge of determining an appropriate fidelity for each hyperparameter configuration to fit the surrogate model. To tackle the challenge, we propose a multi-fidelity BO method named FastBO, which excels in adaptively deciding the fidelity for each configuration and providing strong performance while ensuring efficient resource usage. These advantages are achieved through our proposed techniques based on the concepts of efficient point and saturation point for each configuration, which can be obtained from the empirical learning curve of the configuration, estimated from early observations. Extensive experiments demonstrate FastBO's superior anytime performance and efficiency in identifying high-quality configurations and architectures. We also show that our method provides a way to extend any single-fidelity method to the multi-fidelity setting, highlighting the wide applicability of our approach.

Bayesian networks (BNs) have recently attracted more attention, because they are interpretable machine learning models and enable a direct representation of causal relations between variables. However, exact inference on BNs is time-consuming, especially for complex problems, which hinders the widespread adoption of BNs. To improve the efficiency, we propose a fast BN exact inference named Faster-BNI on multi-core CPUs. Faster-BNI enhances the efficiency of a well-known BN exact inference algorithm, namely the junction tree algorithm, through hybrid parallelism that tightly integrates coarse- and fine-grained parallelism. Moreover, we identify that the bottleneck of BN exact inference methods lies in recursively updating the potential tables of the network. To reduce the table update cost, Faster-BNI employs novel optimizations, including the reduction of potential tables and re-organizing the potential table storage, to avoid unnecessary memory consumption and simplify potential table operations. Comprehensive experiments on real-world BNs show that the sequential version of Faster-BNI outperforms existing sequential implementation by 9 to 22 times, and the parallel version of Faster-BNI achieves up to 11 times faster inference than its parallel counterparts.

Bayesian networks (BNs) are attractive, because they are graphical and interpretable machine learning models. However, exact inference on BNs is time-consuming, especially for complex problems. To improve the efficiency, we propose a fast BN exact inference solution named Fast-BNI on multi-core CPUs. Fast-BNI enhances the efficiency of exact inference through hybrid parallelism that tightly integrates coarse- and fine-grained parallelism. We also propose techniques to further simplify the bottleneck operations of BN exact inference. Fast-BNI source code is freely available at https://github.com/jjiantong/FastBN.

Positron Emission Tomography (PET) and Single Photon Emission Computerized Tomography (SPECT) are essential medical imaging tools with inherent drawback of slow data acquisition process. With the knowledge that radionuclide images are sparse in transform domain, we have applied a novel concept of Compressive Sampling on them. The proposed approach aims to reconstruct images from fewer measurements, significantly reducing scan time and radiopharmaceutical doze, with benefits for patients and health care economics. The reconstruction of tomographic images is realized by compressed sensing the 2-D Fourier projections. These 2-D projections being sparse in transform domain are sensed with fewer samples in k-space and are reconstructed without loss of fidelity. These undersampled Fourier projections can then be backprojected by employing the iterative reconstruction approach for a complete 3-D volume. Our work focuses on the acquisition of 2-D SPECT/PET projections based on compressive sampling and their reconstruction using a non-linear recovery algorithm. Compressive sampling of a phantom image and PET bone scan scintigraph with radial Fourier samples are performed. The reconstructions of these images are compared to conventionally sampled images with MSE, PSNR and a new image quality measure, Structure SIMilarity (SSIM). The results show high quality image reconstruction using considerably few measurements.

Non Destructive Inspections (NDI) plays a vital role in aircraft industry as it determines the structural integrity of aircraft surface and material characterization. The existing NDI methods are time consuming, we propose a new NDI approach using Digital Image Processing that has the potential to substantially decrease the inspection time. The aircraft imagery is analyzed by two methods i.e Contourlet Transform (CT) and Discrete Cosine Transform (DCT). With the help of Contourlet Transform the two dimensional (2-D) spectrum is divided into fine slices, using iterated directional filter banks. Next, directional energy components for each block of the decomposed subband outputs are computed. These energy values are used to distinguish between the crack and scratch images using the Dot Product classifier. In next approach, the aircraft imagery is decomposed into high and low frequency components using DCT and the first order moment is determined to form feature vectors. A correlation based approach is then used for distinction between crack and scratch surfaces. A comparative examination between the two techniques on a database of crack and scratch images revealed that texture analysis using the combined transform based approach gave the best results by giving an accuracy of 96.6% for the identification of crack surfaces and 98.3% for scratch surfaces.

Palmprint based personal verification has gained preference over other biometric modalities due to its ease of acquisition, high user acceptance and reliability. This paper presents a novel palmprint based identification approach which uses the textural information available on the palmprint by employing the non subsampled contourlet transform (NSCT). After establishing the region of interest (ROI), the two dimensional (2-D) spectrum is divided into fine slices, using iterated directional filter banks. Next, directional energy component for each block from the decomposed subband outputs is computed. The proposed algorithm captures both local and global details in a palmprint as a compact fixed length palm code. Palmprint matching is then performed using normalized Euclidean distance classifier. The algorithm is tested on a total of 7752 palm images, acquired from the standard database of Polytechnic University of Hong Kong. The experimental results demonstrated the feasibility of the proposed system by exhibiting Decidability Index of 2.8125 and equal error rate of 0.1604%, better than the reported techniques in literature.