It is important to incorporate patient values in healthcare decision-making systems so that the system gives patient-centered decision solutions in the decision-making process. The extant studies show less integration of both. Patient values can be summarized in several medical community review data. The existing knowledge of values is distilled and then combined with existing knowledge of clinical guidelines and rehabilitation. The patient's personalized information is obtained during the consultation process and matched with the general medical data, resulting in personalized recommendations that incorporate the patient's values. An ontology model is constructed based on five domains: population features, medical treatments, personal values, side effects and rehabilitation. Clinical guideline data, rehabilitation data and patient review data are mapped and linked through the ontology to obtain a general medical knowledge graph. The patient's voice, emotion and gesture information during the consultation process are analyzed and recognized as weighted entities, which are passed into the general medical knowledge graph to form a personalized multi-modal knowledge graph of the patient. A prototype system in a man with breast cancer was designed and implemented to demonstrate the method's feasibility.

The physical layer (PHY) security technology based on radio frequency (RF) fingerprint can effectively solve the secure access problem of wireless devices. The hardware impairments of the devices can be used to generate the unique RF fingerprint to identify different wireless devices. Fingerprint extraction as a key step in the process of identification faces the challenges of ensuring the identification accuracy with reduced sample dimension and low testing and training time. To address the above problems, we propose a lightweight RF fingerprint extraction scheme to extract the physical layer attributes and effectively reduce the data dimension and time consumption. Based on the proposed RF fingerprint, the Bayesian classifier is used to identify the wireless devices. Furthermore, a joint judgment strategy is proposed to improve the identification accuracy by using multiple segments of one signal frame. The experimental result shows that, compared to the existing RF fingerprint identification schemes, the proposed RF fingerprint identification scheme obtains the best identification accuracy with lower time and data consumption.

Space-borne radar altimeters can provide global data of significant wave height (SWH) of total waves, including wind-sea and swells. But wind-seas and swells are often separately considered in the fields of marine science and marine engineering. In this study, a wind-sea-swell SWH separation model was developed using the simultaneous 10-m wind speed and SWH measurement, which are available from altimeter data. The model was based on a look up table extrapolated by an artificial neural network trained by the insitu data from National Data Buoy Center. After applying the model to altimeter data, the results showed that the accuracy of the model is fairly good with a root-mean-square error of wind-sea/swell SWH of only ~0.5 m. This method provides a new method to investigate global wind-sea and swell using remote sensing data.

This paper surveys how to use Convolutional Neural Networks (CNN) to hypothesize object location and categorization from images or videos in mobile heterogeneous SoCs. Recently a variety of CNN-based object detection frameworks have demonstrated both increasing accuracy and speed. Though they are making fast progress in high quality image recognition, state-of-the-art CNN-based detection frameworks seldom discuss their hardware-depended aspects and the cost-effectiveness of real-time image analysis in off-the-shelf low-power devices. As the focus of deep learning and convolutional neural nets is shifting to the embedded or mobile applications with limited power and computational resources, scaling down object detection framework and CNNs is becoming a new and important direction. In this work we conduct a comprehensive comparative study of state-of-the-art real-time object detection frameworks about their performance, cost-effectiveness/energy-efficiency (in the metric of mAP/Wh) in off-the-shelf mobile GPU devices. Based on the analysis results and observation in investigation, we propose to adjust the design parameters of such frameworks and employ a design space exploration procedure to maximize the energy-efficiency (mAP/Wh) of real-time object detection solution in mobile GPUs. As shown in the benchmarking result, we successfully boost the energy-efficiency of multiple popular CNN-based detection solutions by maximizing the utility of computation resources of SoC and trading-off between prediction accuracy and energy cost. In the second Low-Power Image Recognition Challenge (LPIRC), our system achieved the best result measured in mAP/Energy on the embedded Jetson TX1 CPU+GPU SoC.

Multilevel spin toque transfer RAM (STT-RAM) is a suitable storage device for energy-efficient neural network accelerators (NNAs), which relies on large-capacity on-chip memory to support brain-inspired large-scale learning models from conventional artificial neural networks to current popular deep convolutional neural networks. In this paper, we investigate the application of multilevel STT-RAM to general-purpose NNAs. First, the error-resilience feature of neural networks is leveraged to tolerate the read/write reliability issue in multilevel cell STT-RAM using approximate computing. The induced read/write failures at the expense of higher storage density can be effectively masked by a wide spectrum of NN applications with intrinsic forgiveness. Second, we present a precision-tunable STT-RAM buffer for the popular general-purpose NNA. The targeted STT-RAM memory design is able to transform between multiple working modes and adaptable to meet the varying quality constraint of approximate applications. Lastly, the reconfigurable STT-RAM buffer not only enables precision scaling in NNA but also provides adaptiveness to the demand for different learning models with distinct working-set sizes. Particularly, we demonstrate the concept of capacity/precision-tunable STT-RAM memory with the emerging reconfigurable deep NNA and elaborate on the data mapping and storage mode switching policy in STT-RAM memory to achieve the best energy efficiency of approximate computing.

Convolutional neural networks (CNN) accelerators have been proposed as an efficient hardware solution for deep learning based applications, which are known to be both compute-and-memory intensive. Although the most advanced CNN accelerators can deliver high computational throughput, the performance is highly unstable. Once changed to accommodate a new network with different parameters like layers and kernel size, the fixed hardware structure, may no longer well match the data flows. Consequently, the accelerator will fail to deliver high performance due to the underutilization of either logic resource or memory bandwidth. To overcome this problem, we proposed a novel deep learning accelerator, which offers multiple types of data-level parallelism: inter-kernel, intra-kernel and hybrid. Our design can adaptively switch among the three types of parallelism and the corresponding data tiling schemes to dynamically match different networks or even different layers of a single network. No matter how we change the hardware configurations or network types, the proposed network mapping strategy ensures the optimal performance and energy-efficiency. Compared with previous state-of-the-art NN accelerators, it is possible to achieve a speedup of 4.0x-8.3x for some layers of the well-known large scale CNNs. For the whole phase of network forward-propagation, our design achieves 28.04% PE energy saving, 90.3% on-chip memory energy saving on average.

Multi-level STT-RAM cell is able to boost the memory density at the expense of read/write reliability. However, the induced data integrity issue in STT-RAM memory can be effectively masked by a wide spectrum of applications with intrinsic forgiveness, which belong to the specific domain such as multimedia, synthesis and mining. In this work, we leverage the reconfigurable capability of MLC STT-RAM to provide variable-precision data storage for popular machine learning architectures. The targeted STT-RAM memory design is able to transform between multiple work modes and adaptable to meet the varying quality constraint of approximate applications. Particularly, we demonstrate the concept of precision-tunable STT-RAM memory with the emerging Convolution Neural Network accelerators and elaborate on the data mapping policy in STT-RAM memory to achieve the best energy-efficiency.