This dissertation develops the probably approximately correct (PAC) learnable fuzzy system to predict clinical outcomes from a small number of survey questions (short form). There are five layers in the system: input, fuzzification, inference, defuzzification, and production. The major product in this dissertation is to derive the PAC learnable knowledge-driven machine learning algorithm by growing sample using Bootstrap samples with Gaussian distributed noise. The input layer is the procedure for preparing data input. In the fuzzification layer, sample size is significantly increased using bootstrap re-sampling with replacement. The fuzzy set with proposed membership function is generated by introducing Gaussian distributed noise to survey responses of the bootstrap samples to reflect uncertainty. This is a natural language extension from the point option in survey questions to region input with probabilities from survey design space. The inference layer includes both classification and prediction. Here we use machine learning techniques to derive the algorithms in this layer, e.g. Naive Bayesian method and eXtreme Gradient Boosting (XGBoost). The final predicted values require a defuzzification process in the next layer to remove noise in prediction. There are four types of input after fuzzification, original input, fuzzy input, input required interpolation and input required extrapolation. The defuzzification process is based on weighted means of related information. The last step of the system is the output layer with algorithms, final prediction and validation internally and externally. Lastly, we apply this fuzzy system to derive PAC learnable algorithms to predict oral health clinical outcomes. The input predictors include short forms and demographic information. The short forms, developed from Graded Response Models in Item Response Theory, have two versions (children and their parents). The clinical outcomes are referral for treatment needs (categorical) and children’s oral health status index score (continuous). The prediction is evaluated internally and externally by sensitivity and specificity of a binary variable, correlation (between original value and predicted value) and root mean square error (RMSE) of a continuous variable. Both internal and external validation show the improvement of prediction when new information is added and generalizability as well as the stability of the algorithm. The best prediction (high sensitivity and relatively high specificity for categorical variables, low RMSE and high correlation) is reached when using child's self-reported short form, plus parent's proxy-reported short form, and demographic characteristics.

Quantum computation has aroused many attentions toward the superconducting devices. One of the essential elements is the Josephson junction. This dissertation presents a novel technique for fabricating Josephson junction applying to several superconducting materials using a focused helium ion beam. The fabrication procedure and transport properties of the devices made with different superconducting materials are reported individually.

One of the most well-know high-T$_\text{C}$ superconducting cuprate material, YBa$_2$Cu$_3$O$_{7-\delta}$, is first implemented with the helium ion technique. Numerous Josephson junctions have been made with YBa$_2$Cu$_3$O$_{7-\delta}$ thin films. A deeper analysis for junction characteristics is provided to point out the critical parameters for describing junction behaviors.

The other famous high-T$_\text{C}$ cuprate superconductor investigated in this dissertation is Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ single crystals. I fabricated the first planar Josephson junction on the \textit{a-b} plane of Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ with helium ion irradiation. The junction orientations are confirmed by the magnetic field measurements, and the nearly perfect Fraunhofer patterns is revealed. This is a breakthrough for development of multidimensional superconducting circuits.

The mechanisms for cuprate high-T$_\text{C}$ superconductivity hasn't been fully understood yet. This work compares the tunneling spectra of the cuprate thin films (YBa$_2$Cu$_3$O$_{7-\delta}$) and the cuprate single crystals (Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$). I further study the direct tunneling from a high-T$_\text{C}$ cuprate superconductor to another homogeneous high-T$_\text{C}$ cuprate superconductor. The spectra reveals tunneling from superconducting carriers and quasi-particles for both materialss. The temperature dependence of the gap-like structures in the spectrum can be depicted with the Bardeen-Cooper-Schrieffer theory, which implies a potential \textit{s}-wave behavior. In contrast, the gapless spectra demonstrate the possible existence of \textit{d}-wave component.

Last, the same technique of high-energy helium ions was employed to irradiate a conventional superconductor, niobium nitride, and a iron-based superconductor. The effects of ion irradiation to these superconducting materials are also presented to assist the potential extension of the technique.

In conclusion, this work provides another perspective to the fundamental mechanism for cuprate high-T$_\text{C}$ superconductivity through helium ion disorder barriers with YBa$_2$Cu$_3$O$_{7-\delta}$ thin films and Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ single crystlas. I also established the fabrication procedure for ion disordered \textit{a-b} plane Josephson junctions on the fresh cleaved Bi$_2$Sr$_2$CaCu$_2$O$_{8+\delta}$ crystals at any desired angles, which can facilitate the advancement for superconducting circuits. The technique of using helium ion irradiation is further extended to other non-cuprate superconductors and set the footing stone for future applications.

Human motion monitoring and activity classification, specifically in the free-living environment, are becoming increasingly important as preventative, diagnostic and rehabilitative measures in health and wellness applications.

In contrast to gait analysis, wearable sensor-based evaluation of upper body activities is not well studied. The work in this thesis tends to explore a novel system for upper limb activity monitoring and classification. The system focuses specifically on the application of motion classification to a complex task of automating rehabilitation evaluation, such as the Wolf Motor Function Test. The presented system consists of a novel wearable motion sensor platform that integrates accelerometers, gyroscopes and flex-sensors, and classification algorithms that convert motion data into an alphabet representation and form a string of primitives. String expressions are then derived for each test item and a regular expression based searching method is developed. We present results from the successful application of the proposed system to upper limb activity characterization in the context of the Wolf Motor Function Test.

Optical coherence tomography (OCT) is an optical imaging method based on low-coherence interferometry that is capable of high-resolution cross-sectional imaging of internal microstructure by measuring light backscattered from the sample. OCT has the capacity to perform non-contact in vivo imaging and has been applied in medical and scientific fields such as ophthalmology, dermatology, developmental biology, and cardiology. Extensions of OCT such as Doppler OCT and polarization-sensitive OCT (PS-OCT) provide additional information about biological tissues. However, the heavy computational load required to process the acquired data stream creates a limit in realization of real-time OCT. Performing multi-functional OCT imaging demands additional processing for reconstruction of functional images, further increasing the total processing time. Graphics processing unit (GPU) processing has been implemented into MRI, CT and ultrasound as well as intensity-only OCT systems to accelerate image processing using its inherent parallel computation architecture. In this thesis, the development and applications of a GPU accelerated real-time 4D multi-functional SD-OCT system was presented.

Chapter 1 describes the construction and characterization of a multi-functional SD-OCT system at 1300nm. The axial resolution, lateral resolution, imaging depth, signal sensitivity drop-off, spectrometer efficiency, phase noise, system noise, polarization noise, computation of phase retardation, optic axis and diattenuation were characterized. Multiple samples were imaged to demonstrate each imaging facet (intensity, polarization and flow) of the system.

In Chapter 2, GPU was included into the imaging system. CUDA C++ was implemented into the real-time data acquisition and processing program to realize real-time data processing and display. The hybrid GPU-CPU program can process all intensity, polarization and flow images 100 times faster than a comparable previous CPU program. The efficient line processing rate realized was 379kHz for all three image types simultaneously, and represents a computational speed 8 times faster than camera line acquisition rate.

In Chapter 3, volume ray casting using GPU was incorporated into the real-time data acquisition and processing program for real-time volume rendering of all intensity, polarization and flow volume images. Arterial pulsation flow was identified and differentiated from venal flow in flow volume images when imaging mouse femoral artery and vein in vivo. The real-time volume images allowed visualization of burn progression in a volume as well as identification of burn injury boundaries in chicken muscle tissue.

The application of PS-OCT for the study of peripheral nerve degeneration and regeneration post crush injury is described in Chapter 4. Results from a preliminary non-longitudinal study showed that axonal birefringence decreased post crush injury and then increased with nerve repair. A longitudinal study was initiated in order to reduce the variation caused by different animals. The facilitated visualization of the real-time 3D volume rendering program was used to rapidly identify the micro-suture markers used to demarcate the wound area in order to minimize operational time while ensuring the imaging region was consistent over repeated imaging sessions.

In Chapter 5, PS-OCT was applied on a study of human skin photo aging as a factor of age, gender. Thirty-seven volunteers were recruited in accordance with an approved protocol for skin imaging on sun exposed area (face) and sun protected area (inner upper arm). The results showed a significantly lower birefringence value in the older age group (>=55 years old) than that of younger age group (18-35 years old). There was no significant difference found between young male and young female as well as old male and old female groups.

Debugging is a tedious and time-consuming process for software developers. Therefore, providing effective and efficient debugging tools is essential for improving programmer productivity. Existing tools for debugging suffer from various drawbacks -- general-purpose debuggers provide little guidance for the programmers in locating the

bug source while specialized debuggers require knowledge of the type of bug encountered. This dissertation makes several advances in debugging leading to effective, efficient, and extensible framework for interactive debugging of singlethreaded programs and deterministic debugging of multithreaded programs.

This dissertation presents the Qzdb debugger for singlethreaded programs that raises the abstraction level of debugging by introducing high-level and powerful state alteration and state inspection capabilities. Case studies on 5 real reported bugs in 5 popular real programs demonstrate its effectiveness. To support integration of specialized debugging algorithms into Qzdb, a

new approach for constructing debuggers is developed that employs declarative specification of bug conditions and their root causes, and automatic generation of debugger code. Experiments show that about 3,300 lines of C code are generated automatically from only 8 lines of specification for 6 memory bugs. Thanks to the effective generated bug locators, for the 8 real-worlds bugs we have applied our approach to, users have to examine just 1 to 16 instructions. To reduce the runtime overhead of dynamic analysis used during debugging, relevant input analysis is developed and employed to carry out input simplification and execution simplification which reduce the length of analyzed execution by reducing the input size and limiting the analysis to subset of the execution. Experiments show that relevant input analysis based input simplification algorithm is both efficient and effective -- it only requires 11% to 21% test runs of that needed by standard delta debugging algorithm and generates even smaller inputs.

Finally, to demonstrate that the above approach can also be used for debugging multithreaded programs, this dissertation presents DrDebug, a deterministic and cyclic debugging framework. DrDebug allows efficient debugging by tailoring the scope of replay to a buggy execution region and an execution slice of a buggy region. Case studies of real reported concurrency bugs show that the buggy execution region size is less than 1 million instructions and the lengths of buggy execution region and execution slice are less than

15% and 7% of the total execution respectively.

Human activity analysis is becoming increasingly important to enable preventative, diagnostic and rehabilitative measures in health and wellness applications. Wearable sensors are now taking a leading role in this area. However, daily activity profiling is still challenging due to a number of complications: 1) Method to quantify human activities are not precisely defined, especially in the domain of upper body activities. 2) Body joints have multiple degrees of freedom that are required to establish the flexibility of activity performance. Therefore, even if an agreement is achieved in activity recognition, performance of a specific activity can still present numerous variations for one person in different performance bouts; 3) Performance evaluation is also challenging without a complete set of training data since the difference between pathological movements and special personal behaviors may not be accurately classified by a general activity analysis system.

Therefore, a new solution is provided to address the above three challenges. In this dissertation, a wearable sensor based location driven high-level activity classification and 3D human motion tracking system will be presented. This system consists of three sub-systems, location detection, activity classification and motion tracking. The location detection system determines the user’s location and gives this information to the activity classification system without reliance on any other localization instrumentation or infrastructure. In addition to being a fundamental advance, this adds an additional critical benefit, the location detection system also provides navigation functionality to the end users. The activity classification system classifies the general activities the end user is performing. Since the activity set is limited by a user’s location, the classification performance will be largely improved both in terms of accuracy and efficiency. The activity classification system also indicates the high-level activity classification result. On one hand, it can be used for compliance assessment to examine whether the user is following his prescribed exercise activity according to proper protocols. On the other hand, it enables an activity specific motion tracking protocol to be used in the motion tracking system resulting in reduced analytics computation demand. The motion tracking system supports reconstruction of motions for remote and post-event visualization as well as performance evaluation based on the idea of range of motions. In the dissertation, the algorithms at the foundation of each of the three subsystems will be described as well as complete system integration.