Zygarde: Time-Sensitive On-Device Deep Intelligence on Intermittently-Powered Systems

In this paper, we propose a time-, energy-, and accuracy-aware scheduling algorithm for intermittently powered systems that execute compressed deep learning tasks that are suitable for MCUs and are powered solely by harvested energy. The sporadic nature of harvested energy, resource constraints of the embedded platform, and the computational demand of deep neural networks (even though compressed) pose a unique and challenging real-time scheduling problem for which no solutions have been proposed in the literature. We empirically study the problem and model the energy harvesting pattern as well as the trade-off between the accuracy and execution of a deep neural network. We develop an imprecise computing-based scheduling algorithm that improves the schedulability of deep learning tasks on intermittently powered systems. We also utilize the dependency of the computational need of data samples for deep learning models and propose early termination of deep neural networks. We further propose a semi-supervised machine learning model that exploits the deep features and contributes in determining the imprecise partition of a task. We implement our proposed algorithms on two different datasets and real-life scenarios and show that it increases the accuracy by 9.45% - 3.19%, decreases the execution time by 14\% and successfully schedules 33%-12% more tasks.

2.5D Thermometry Maps for MRI-guided Tumor Ablation

Fast and reliable monitoring of volumetric heat distribution during MRI-guided tumor ablation is an urgent clinical need. In this work, we introduce a method for generating 2.5D thermometry maps from uniformly distributed 2D MRI phase images rotated around the applicator's main axis. The images can be fetched directly from the MR device, reducing the delay between image acquisition and visualization. For reconstruction, we use a weighted interpolation on a cylindric coordinate representation to calculate the heat value of voxels in a region of interest. A pilot study on 13 ex vivo bio protein phantoms with flexible tubes to simulate a heat sink effect was conducted to evaluate our method. After thermal ablation, we compared the measured coagulation zone extracted from the post-treatment MR data set with the output of the 2.5D thermometry map. The results show a mean Dice score of 0.75+-0.07, a sensitivity of 0.77+-0.03, and a reconstruction time within 18.02ms+-5.91ms. Future steps should address improving temporal resolution and accuracy, e.g., incorporating advanced bioheat transfer simulations.

Hyperparameter Optimization: Foundations, Algorithms, Best Practices and Open Challenges

Most machine learning algorithms are configured by one or several hyperparameters that must be carefully chosen and often considerably impact performance. To avoid a time consuming and unreproducible manual trial-and-error process to find well-performing hyperparameter configurations, various automatic hyperparameter optimization (HPO) methods, e.g., based on resampling error estimation for supervised machine learning, can be employed. After introducing HPO from a general perspective, this paper reviews important HPO methods such as grid or random search, evolutionary algorithms, Bayesian optimization, Hyperband and racing. It gives practical recommendations regarding important choices to be made when conducting HPO, including the HPO algorithms themselves, performance evaluation, how to combine HPO with ML pipelines, runtime improvements, and parallelization.

Multiple radio frequency measurement with an improved frequency resolution based on stimulated Brillouin scattering with a reduced gain bandwidth

A photonic-assisted multiple radio frequency (RF) measurement approach based on stimulated Brillouin scattering (SBS) and frequency-to-time mapping with high accuracy and high-frequency resolution is reported. A two-tone signal is single-sideband (SSB) modulated on an optical carrier via a dual-parallel Mach-Zehnder modulator to construct one SBS gain and two SBS losses for SBS gain bandwidth reduction. The unknown RF signal is also SSB modulated on a carrier that has been modulated by a sweep signal, thus the unknown RF signal is converted to a sweep optical signal along with the sweep optical carrier. The bandwidth-reduced SBS gain spectrum is detected by the sweep optical signals at different specific time, mapping the RF frequencies to the time domain. An experiment is performed. RF frequencies from 0.3 to 7.6 GHz are simultaneously measured with a root mean square error of less than 1 MHz. In addition, the frequency resolution of the measurement can be much lower than 10 MHz, which is now the best result in the RF frequency measurement methods employing the SBS effect.

IRS-Aided Energy Efficient UAV Communication

Unmanned aerial vehicles (UAVs) have steadily gained attention to overcome the harsh propagation loss and blockage issue of millimeter-wave communication. However, UAV communication systems suffer from energy consumption, which limits the flying time of UAVs. In this paper, we propose several UAV energy consumption minimization techniques through the aid of multiple intelligent reflecting surfaces (IRSs). In specific, we introduce a tractable model to effectively capture the characteristics of multiple IRSs and multiple user equipments (UEs). Then, we derive a closed form expression for the UE achievable rate, resulting in tractable optimization problems. Accordingly, we effectively solve the optimization problems by adopting the successive convex approximation technique. To compensate for the high complexity of the optimization problems, we propose a low complexity algorithm that has marginal performance loss. In the numerical results, we show that the proposed algorithms can save UAV energy consumption significantly compared to the benchmark with no IRSs, justifying that exploiting the IRSs is indeed favorable to UAV energy consumption minimization.

Transmit Beamspace DDMA Based Automotive MIMO Radar

The time division multiple access (TDMA) technique has been applied in automotive multiple-input multiple-output (MIMO) radar. However, it suffers from the transmit energy loss, and as a result the parameter estimation performance degradation when the number of transmit elements increases. To tackle these problem, a transmit beamspace (TB) Doppler division multiple access (DDMA) approach is proposed. First, a phase modulation matrix with empty Doppler spectrum is introduced. By exploiting the empty Doppler spectrum, a test function based on sequential detection is developed to mitigate the Doppler ambiguity in DDMA waveform. Then, a discrete Fourier transform (DFT)-based TB in slow-time is formed.The proposed method can achieve waveform diversity in Doppler domain and generate a TB in slow-time that concentrates the transmitted power in a fixed spatial region to improve the transmit energy distribution for automotive MIMO radar, which is favored by medium/long range radar (MRR/LRR) applications. As compared to the conventional TDMA technique, the proposed TB DDMA approach can fully exploit the transmission capabilities of all transmit elements to ensure that the emitted power is efficiently used and inherits easy implementation. Moreover, the proposed TB DDMA method avoids the trade-off between the active time for each transmit antenna and the frame time. Simulation results verify the effectiveness of the proposed TB DDMA approach for automotive MIMO radar.

UniFaceGAN: A Unified Framework for Temporally Consistent Facial Video Editing

Recent research has witnessed advances in facial image editing tasks including face swapping and face reenactment. However, these methods are confined to dealing with one specific task at a time. In addition, for video facial editing, previous methods either simply apply transformations frame by frame or utilize multiple frames in a concatenated or iterative fashion, which leads to noticeable visual flickers. In this paper, we propose a unified temporally consistent facial video editing framework termed UniFaceGAN. Based on a 3D reconstruction model and a simple yet efficient dynamic training sample selection mechanism, our framework is designed to handle face swapping and face reenactment simultaneously. To enforce the temporal consistency, a novel 3D temporal loss constraint is introduced based on the barycentric coordinate interpolation. Besides, we propose a region-aware conditional normalization layer to replace the traditional AdaIN or SPADE to synthesize more context-harmonious results. Compared with the state-of-the-art facial image editing methods, our framework generates video portraits that are more photo-realistic and temporally smooth.

Unifying Probabilistic Models for Time-Frequency Analysis

In audio signal processing, probabilistic time-frequency models have many benefits over their non-probabilistic counterparts. They adapt to the incoming signal, quantify uncertainty, and measure correlation between the signal's amplitude and phase information, making time domain resynthesis straightforward. However, these models are still not widely used since they come at a high computational cost, and because they are formulated in such a way that it can be difficult to interpret all the modelling assumptions. By showing their equivalence to Spectral Mixture Gaussian processes, we illuminate the underlying model assumptions and provide a general framework for constructing more complex models that better approximate real-world signals. Our interpretation makes it intuitive to inspect, compare, and alter the models since all prior knowledge is encoded in the Gaussian process kernel functions. We utilise a state space representation to perform efficient inference via Kalman smoothing, and we demonstrate how our interpretation allows for efficient parameter learning in the frequency domain.

Accelerating Neural Network Training with Distributed Asynchronous and Selective Optimization (DASO)

With increasing data and model complexities, the time required to train neural networks has become prohibitively large. To address the exponential rise in training time, users are turning to data parallel neural networks (DPNN) to utilize large-scale distributed resources on computer clusters. Current DPNN approaches implement the network parameter updates by synchronizing and averaging gradients across all processes with blocking communication operations. This synchronization is the central algorithmic bottleneck. To combat this, we introduce the Distributed Asynchronous and Selective Optimization (DASO) method which leverages multi-GPU compute node architectures to accelerate network training. DASO uses a hierarchical and asynchronous communication scheme comprised of node-local and global networks while adjusting the global synchronization rate during the learning process. We show that DASO yields a reduction in training time of up to 34% on classical and state-of-the-art networks, as compared to other existing data parallel training methods.

Robust Interactive Semantic Segmentation of Pathology Images with Minimal User Input

From the simple measurement of tissue attributes in pathology workflow to designing an explainable diagnostic/prognostic AI tool, access to accurate semantic segmentation of tissue regions in histology images is a prerequisite. However, delineating different tissue regions manually is a laborious, time-consuming and costly task that requires expert knowledge. On the other hand, the state-of-the-art automatic deep learning models for semantic segmentation require lots of annotated training data and there are only a limited number of tissue region annotated images publicly available. To obviate this issue in computational pathology projects and collect large-scale region annotations efficiently, we propose an efficient interactive segmentation network that requires minimum input from the user to accurately annotate different tissue types in the histology image. The user is only required to draw a simple squiggle inside each region of interest so it will be used as the guiding signal for the model. To deal with the complex appearance and amorph geometry of different tissue regions we introduce several automatic and minimalistic guiding signal generation techniques that help the model to become robust against the variation in the user input. By experimenting on a dataset of breast cancer images, we show that not only does our proposed method speed up the interactive annotation process, it can also outperform the existing automatic and interactive region segmentation models.