• Themes
• Newsmakers

# Fusion

08.04.2021
23:31 Google-backed nuclear energy firm TAE Technologies raises $280 mln (Reuters) ReutersGoogle-backed nuclear energy firm TAE Technologies raises$280 mln - TAE Technologies, a California-based firm building technology to generate power from nuclear fusion, said on Thursday it had raised 280 million from new and existing investors, including Google and New Enterprise Associates. ... 22:34 With ‘smoke ring’ technology, fusion startup marks steady progress Flush with venture capital, TAE Technologies says it sees a path to goal of net energy production 16:05 The Direct Fusion Drive That Could Get Us Past Neptune in 10 Years Scientists have outlined the wild way humans could travel past Neptune in under 10 years—with over 1.5 tons of cargo on board. ? You think space is badass. So do we. Let’s nerd out over it together. The secret is an in-the-works direct fusion drive (DFD), which will kick in once the spacecraft reaches orbit and propel it at up to 44 kilometers per second. From there, the spacecraft could conduct experiments on Neptune as well as trans-Neptunian objects (TNOs), or outer dwarf planets like Makemake, Eris, and Haumea. The DFD is a nuclear reactor being developed by Princeton Plasma Physics Laboratory and Princeton Satellite Systems that uses pure deuterium as the fuel to generate both thrust and electrical power for the spacecraft. In a new paper published to the preprint server arXiv, scientists from the U.S., Italy, and Russia explain how the DFD works: 08:56 The hot-tail runaway seed landscape during the thermal quench in tokamaks. (arXiv:2104.03272v1 [physics.plasm-ph]) Runaway electron populations seeded from the hot-tail generated by the rapid cooling in plasma-terminating disruptions are a serious concern for next-step tokamak devices such as ITER. Here, we present the first combined treatment of crucial terms in the runaway seed process: hot-tail generation and magnetic perturbation induced loss. We show simulations of the runaway generation in plasma disruptions mitigated by material injection, including the superthermal electron dynamics, heat and particle transport, atomic physics, and radial losses due to magnetic perturbations. We find limits on the impurity density and magnetic perturbation level for which the runaway seed current is acceptable without excessive thermal energy being lost to the wall via particle impact. The consistent modelling of generation and losses allows us to identify parameter regimes which lead to runaway beams forming near the edge of the plasma, where they could be deconfined via external perturbations. 06.04.2021 07:22 Comparative study on the uniform energy deposition achievable via optimized plasmonic nanoresonator distributions. (arXiv:2104.02027v1 [physics.optics]) Plasmonic nanoresonators of core-shell composition and nanorod shape were optimized to tune their absorption cross-section maximum to the central wavelength of a short pulse. Their distribution along a pulse-length scaled target was optimized to maximize the absorptance with the criterion of minimal absorption difference in between neighbouring layers. Successive approximation of layer distributions made it possible to ensure almost uniform deposited energy distribution up until the maximal overlap of two counter-propagating pulses. Based on the larger absorptance and smaller uncertainty in absorptance and energy distribution core-shell nanoresonators override the nanorods. However, optimization of both nanoresonator distributions has potential applications, where efficient and uniform energy deposition is crucial, including phase transitions and even fusion. 07:22 Late fusion of machine learning models using passively captured interpersonal social interactions and motion from smartphones predicts decompensation in heart failure. (arXiv:2104.01511v1 [eess.SP]) Objective: Worldwide, heart failure (HF) is a major cause of morbidity and mortality and one of the leading causes of hospitalization. Early detection of HF symptoms and pro-active management may reduce adverse events. Approach: Twenty-eight participants were monitored using a smartphone app after discharge from hospitals, and each clinical event during the enrollment (N=110 clinical events) was recorded. Motion, social, location, and clinical survey data collected via the smartphone-based monitoring system were used to develop and validate an algorithm for predicting or classifying HF decompensation events (hospitalizations or clinic visit) versus clinic monitoring visits in which they were determined to be compensated or stable. Models based on single modality as well as early and late fusion approaches combining patient-reported outcomes and passive smartphone data were evaluated. Results: The highest AUCPr for classifying decompensation with a late fusion approach was 0.80 using 05.04.2021 10:12 FT-BLAS: A High Performance BLAS Implementation With Online Fault Tolerance. (arXiv:2104.00897v1 [cs.DC]) Basic Linear Algebra Subprograms (BLAS) is a core library in scientific computing and machine learning. This paper presents FT-BLAS, a new implementation of BLAS routines that not only tolerates soft errors on the fly, but also provides comparable performance to modern state-of-the-art BLAS libraries on widely-used processors such as Intel Skylake and Cascade Lake. To accommodate the features of BLAS, which contains both memory-bound and computing-bound routines, we propose a hybrid strategy to incorporate fault tolerance into our brand-new BLAS implementation: duplicating computing instructions for memory-bound Level-1 and Level-2 BLAS routines and incorporating an Algorithm-Based Fault Tolerance mechanism for computing-bound Level-3 BLAS routines. Our high performance and low overhead are obtained from delicate assembly-level optimization and a kernel-fusion approach to the computing kernels. Experimental results demonstrate that FT-BLAS offers high reliability and high performance 02.04.2021 11:08 Optimizer Fusion: Efficient Training with Better Locality and Parallelism. (arXiv:2104.00237v1 [cs.LG]) Machine learning frameworks adopt iterative optimizers to train neural networks. Conventional eager execution separates the updating of trainable parameters from forward and backward computations. However, this approach introduces nontrivial training time overhead due to the lack of data locality and computation parallelism. In this work, we propose to fuse the optimizer with forward or backward computation to better leverage locality and parallelism during training. By reordering the forward computation, gradient calculation, and parameter updating, our proposed method improves the efficiency of iterative optimizers. Experimental results demonstrate that we can achieve an up to 20% training time reduction on various configurations. Since our methods do not alter the optimizer algorithm, they can be used as a general "plug-in" technique to the training process. 01.04.2021 12:03 UK nuclear fusion firms make progress towards providing clean power Two UK companies leading efforts to reproduce the way the sun makes energy are on the way to hitting significant milestones in their attempts to commercialise nuclear fusion by the 2030s 31.03.2021 08:25 DREAM: a fluid-kinetic framework for tokamak disruption runaway electron simulations. (arXiv:2103.16457v1 [physics.plasm-ph]) Avoidance of the harmful effects of runaway electrons (REs) in plasma-terminating disruptions is pivotal in the design of safety systems for magnetic fusion devices. Here, we describe a computationally efficient numerical tool, that allows for self-consistent simulations of plasma cooling and associated RE dynamics during disruptions. It solves flux-surface averaged transport equations for the plasma density, temperature and poloidal flux, using a bounce-averaged kinetic equation to self-consistently provide the electron current, heat, density and RE evolution, as well as the electron distribution function. As an example, we consider disruption scenarios with material injection and compare the electron dynamics resolved with different levels of complexity, from fully kinetic to fluid modes. 08:25 Magnetic Fluctuations in Gyrokinetic Simulations of Tokamak Scrape-Off Layer Turbulence. (arXiv:2103.16062v1 [physics.plasm-ph]) Understanding turbulent transport physics in the tokamak edge and scrape-off layer (SOL) is critical to developing a successful fusion reactor. The dynamics in these regions plays a key role in achieving high fusion performance by determining the edge pedestal that suppresses turbulence in H-mode. Additionally, the survivability of a reactor is set by the heat load to the vessel walls, making it important to understand turbulent spreading of heat as it flows along open magnetic field lines in the SOL. Large-amplitude fluctuations, magnetic X-point geometry, and plasma interactions with material walls make simulating turbulence in the edge/SOL more challenging than in the core region, necessitating specialized gyrokinetic codes. Further, the inclusion of electromagnetic effects in gyrokinetic simulations that can handle the unique challenges of the boundary plasma is critical to the understanding of phenomena such as the pedestal and ELMs, for which electromagnetic dynamics are expected 30.03.2021 11:05 ViralFP: A webserver of viral fusion proteins. (arXiv:2103.14754v1 [q-bio.QM]) Viral fusion proteins are attached to the membrane of enveloped viruses (a group that includes Coronaviruses, Dengue, HIV and Influenza) and catalyze fusion between the viral envelope and host membrane, enabling the virus to insert its genetic material into the host cell. Given the importance of these biomolecules, this work presents a centralized database containing the most relevant information on viral fusion proteins, available through a free-to-use web server accessible accessible by the URL https://viralfp.bio.di.uminho.pt/. This server also contains several bioinformatic tools, such as Clustal sequence alignment and Weblogo, and includes a machine learning-based tool capable of predicting the location of fusion peptides (the component of fusion proteins that inserts into the host's cell membrane) within the fusion protein sequence. Given the crucial role of these proteins in viral infection, their importance as natural targets of our immune system and their potential as 11:05 Influence of Micro-turbulence on Neoclassical Tearing Mode Onset. (arXiv:2103.15506v1 [physics.plasm-ph]) Direct evidence of micro-turbulence effect on the onset of neoclassical tearing mode (NTM) is reported for the first time in this letter. A puzzling positive correlation between critical width of seed island of NTM and normalized plasma pressure beta_p is first observed employing a novel method for clearly separating the processes of seed island and the onset of NTM in the EAST tokamak. Different from the methods developed before, the width of the seed island is well controlled by slowly ramping up the current in resonant magnetic perturbation coils. It is revealed that the positive correlation is mainly attributed to the enhancement of perpendicular transport by micro-turbulence, which overcomes the destabilizing effect of beta_p on the onset of NTM. Reduced magnetohydrodynamics (MHD) modeling well reproduced the two states of nonlinear bifurcations observed in this experiment by including the finite transport effect. This result provides a new route for understanding multi-scale 29.03.2021 06:31 Predictive Modeling of a Lithium Vapor Box Divertor in NSTX-U using SOLPS-ITER. (arXiv:2103.14135v1 [physics.plasm-ph]) The lithium vapor box divertor aims to detach the divertor plasma via evaporating and condensing lithium surfaces. By evaporating lithium near or at the divertor plate and condensing it closer to the main chamber, a lithium vapor density gradient can be created. This density gradient ties energy losses to poloidal distance between the target and the detachment point. The radiation zone is then prevented from reaching the X-point as the lithium ionization rate decreases when the detachment front moves away from the divertor target. Here we present Scrape Off Layer Plasma Simulator (SOLPS) simulations of a lithium vapor box divertor using an NSTX-U equilibrium and PFC geometry. The parameters for the core boundary conditions, gas puff intensity, and heat and particle transport coefficients are chosen to match experimental values. Acceptable agreement with experimental Scrape-Off Layer (SOL) widths is found, indicating a reasonable choice of transport coefficients. In predictive 26.03.2021 16:55 Asdex Upgrade experimental facility generates its first plasma For 30 years, the Asdex Upgrade has been paving the way for a fusion power plant that generates climate-neutral energy. The tokamak fusion plant was repeatedly expanded and improved during this time. Not least for this reason, it provides numerous insights that are incorporated into the design and operation of other fusion plants. For example, the Asdex Upgrade team has developed scenarios for the operation of the Jet test plant in the UK and the Iter test plant in France as well as forecasts for a planned demonstration power plant. A conversion planned for mid-2022 is intended to prepare the plant for the future. 25.03.2021 10:39 Minimization of magnetic forces on Stellarator coils. (arXiv:2103.13195v1 [math.OC]) Magnetic confinement devices for nuclear fusion can be large and expensive. Compact stellarators are promising candidates for costreduction, but introduce new difficulties: confinement in smaller volumes requires higher magnetic field, which calls for higher coil-currents and ultimately causes higher Laplace forces on the coils-if everything else remains the same. This motivates the inclusion of force reduction in stellarator coil optimization. In the present paper we consider a coil winding surface, we prove that there is a natural and rigorous way to define the Laplace force (despite the magnetic field discontinuity across the current-sheet), we provide examples of cost associated (peak force, surface-integral of the force squared) and discuss easy generalizations to parallel and normal force-components, as these will be subject to different engineering constraints. Such costs can then be easily added to the figure of merit in any multi-objective stellarator coil optimization code. 10:39 Effects of anisotropic energetic particles on zonal flow residual level. (arXiv:2103.13377v1 [physics.plasm-ph]) In tokamak plasmas, the interaction among the micro-turbulence, zonal flows (ZFs) and energetic particles (EPs) can affect the turbulence saturation level and the consequent confinement quality and thus, is important for future burning plasmas. In this work, the EP anisotropy effects on the ZF residual level are studied by using anisotropic EP distributions with dependence of pitch. Significant effects on the long wavelength ZFs have been found when small to moderate width around the dominant pitch in the EP distribution function is assumed. In addition, it is found that ZF residual level is enhanced by barely passing/trapped and/or deeply trapped EPs, but it is suppressed by well passing and/or intermediate trapped EPs. Numerical calculation shows that for ASDEX Upgrade plasmas, typical EP distribution functions can bring in -5%~+10% mitigation/enhancement in ZF residual level, depending on the EP distribution functions. 10:39 Quantum effects on plasma screening for thermonuclear reactions in laser-generated plasmas. (arXiv:2103.13311v1 [physics.plasm-ph]) A quantum plasma screening model based on the density matrix formalism is used to investigate theoretically the thermonuclear reactions^{13}$C($\alpha$,$n$)$^{16}$O and$^2$H($d$,$n$)$^3$He in laser-generated plasmas over a large range of densities and temperatures. We find that for cold and dense (solid-state density) plasmas, the quantum model predicts plasma screening enhancement factors up to one order of magnitude larger than the ones from classical plasma models. Our results indicate that quantum effects can enhance the plasma screening for thermonuclear reactions, with potential also for industrial fusion energy gain. We put forward a possible experimental test of the screening theory in laser-generated plasmas which could also confirm predictions from nuclear astrophysics. 10:39 Minimization of magnetic forces on Stellarator coils. (arXiv:2103.13195v1 [math.OC]) Magnetic confinement devices for nuclear fusion can be large and expensive. Compact stellarators are promising candidates for costreduction, but introduce new difficulties: confinement in smaller volumes requires higher magnetic field, which calls for higher coil-currents and ultimately causes higher Laplace forces on the coils-if everything else remains the same. This motivates the inclusion of force reduction in stellarator coil optimization. In the present paper we consider a coil winding surface, we prove that there is a natural and rigorous way to define the Laplace force (despite the magnetic field discontinuity across the current-sheet), we provide examples of cost associated (peak force, surface-integral of the force squared) and discuss easy generalizations to parallel and normal force-components, as these will be subject to different engineering constraints. Such costs can then be easily added to the figure of merit in any multi-objective stellarator coil optimization code. 24.03.2021 06:16 On the effect of beating during nonlinear frequency chirping. (arXiv:2103.12561v1 [physics.plasm-ph]) Spectral analyses of energetic particle (EP) driven bursts of MHD fluctuations in magnetically confined plasmas often exhibit multiple simultaneous chirps. While the superposition of oscillations at multiple frequencies necessarily causes beating in the signal acquired by a localized external probe, self-consistent hybrid simulations of chirping EP modes in a JT-60U tokamak plasma have demonstrated the possibility of global beating, where the mode's electromagnetic field vanishes globally between beats and reappears with opposite phase. This implies that there can be a single field mode that oscillates at multiple frequencies simultaneously when it is resonantly driven by multiple density waves in EP phase space. Conversely, this means that the EP density waves are mutually coupled and interfere with each other via the jointly driven field, a mechanism ignored in some theories of chirping. In this thesis-style treatise, we study the role of field pulsations in general and beating in 06:16 Plant efficiency: a sensitivity analysis of the capacity factor for fusion power plants with high recirculated power. (arXiv:2103.12451v1 [physics.soc-ph]) The plant efficiency of a nuclear fusion power plant is considered. During nominal operation, the plant efficiency is determined by the thermodynamic efficiency and the recirculated power fraction. However, on average the reactor operates below the nominal power, even when the long shutdown periods for large maintenance are left outside the averaging. Hence, next to the recirculated power fraction, the capacity factor must be factored in. An expression for the plant efficiency which incorporates both factors is given. It is shown that the combination of high recirculated power fraction and a low capacity factor results in poor plant efficiency. This is due to the fact that in a fusion reactor the recirculated power remains high if it runs at reduced output power. It is argued that, at least for a first generation of power plants, this combination is likely to occur. Worked out example calculations are given for the models of the Power Plant Conceptual Study. Finally, the impact on the 06:16 Multimodal Data Fusion for Power-On-and-GoRobotic Systems in Retail. (arXiv:2103.12241v1 [cs.RO]) Robotic systems for retail have gained a lot of attention due to the labor-intensive nature of such business environments. Many tasks have the potential to be automated via intelligent robotic systems that have manipulation capabilities. For example, empty shelves can be replenished, stray products can be picked up or new items can be delivered. However, many challenges make the realization of this vision a challenge. In particular, robots are still too expensive and do not work out of the box. In this paper, we discuss a work-in-progress approach for enabling power-on-and-go robots in retail environments through a combination of active, physical sensors and passive, artificial sensors. In particular, we use low-cost hardware sensors in conjunction with machine learning techniques in order to generate high-quality environmental information. More specifically, we present a setup in which a standard monocular camera and Bluetooth low-energy yield a reliable robot system that can 06:16 Machine learning based in situ quality estimation by molten pool condition-quality relations modeling using experimental data. (arXiv:2103.12066v1 [cond-mat.mtrl-sci]) The advancement of machine learning promises the ability to accelerate the adoption of new processes and property designs for metal additive manufacturing. The molten pool geometry and molten pool temperature are the significant indicators for the final part's geometric shape and microstructural properties for the Wire-feed laser direct energy deposition process. Thus, the molten pool condition-property relations are of preliminary importance for in situ quality assurance. To enable in situ quality monitoring of bead geometry and characterization properties, we need to continuously monitor the sensor's data for molten pool dimensions and temperature for the Wire-feed laser additive manufacturing (WLAM) system. We first develop a machine learning convolutional neural network (CNN) model for establishing the correlations from the measurable molten pool image and temperature data directly to the geometric shape and microstructural properties. The multi-modality network receives both the 22.03.2021 16:29 Faster fusion reactor calculations thanks to machine learning Fusion reactor technologies are well-positioned to contribute to our future power needs in a safe and sustainable manner. Numerical models can provide researchers with information on the behavior of the fusion plasma, as well as valuable insight on the effectiveness of reactor design and operation. However, to model the large number of plasma interactions requires a number of specialized models that are not fast enough to provide data on reactor design and operation. Aaron Ho from the Science and Technology of Nuclear Fusion group in the department of Applied Physics has explored the use of machine learning approaches to speed up the numerical simulation of core plasma turbulent transport. Ho defended his thesis on March 17. 04:42 The effect of rigid electron rotation on the Grad-Shafranov equilibria of a class of FRC devices. (arXiv:2103.10794v1 [physics.plasm-ph]) Rigid electron rotation of a fully penetrated Rotamak-FRC produces a pressure flux function that is more peaked than the Solov'ev flux function. This paper explores the implications of this peaked pressure flux function, including the isothermal case, which appear when the temperature profile is broader than the density profile, creating both benefits and challenges to a Rotamak-FRC based fusion reactor. In this regime, the density distribution becomes very peaked, enhancing the fusion power. The separatrix has a tendency to become oblate, which can be mitigated by flux conserving current loops. Plasma extends outside the separatrix, notably in the open field line region. This model does not apply to very kinetic FRCs or FRCs in which there are significant ion flows, but it may have some applicability to their outer layers. 04:42 Cognitive simulation models for inertial confinement fusion: Combining simulation and experimental data. (arXiv:2103.10590v1 [cs.LG]) The design space for inertial confinement fusion (ICF) experiments is vast and experiments are extremely expensive. Researchers rely heavily on computer simulations to explore the design space in search of high-performing implosions. However, ICF multiphysics codes must make simplifying assumptions, and thus deviate from experimental measurements for complex implosions. For more effective design and investigation, simulations require input from past experimental data to better predict future performance. In this work, we describe a cognitive simulation method for combining simulation and experimental data into a common, predictive model. This method leverages a machine learning technique called transfer learning, the process of taking a model trained to solve one task, and partially retraining it on a sparse dataset to solve a different, but related task. In the context of ICF design, neural network models trained on large simulation databases and partially retrained on experimental 04:42 USTC-NELSLIP System Description for DIHARD-III Challenge. (arXiv:2103.10661v1 [cs.SD]) This system description describes our submission system to the Third DIHARD Speech Diarization Challenge. Besides the traditional clustering based system, the innovation of our system lies in the combination of various front-end techniques to solve the diarization problem, including speech separation and target-speaker based voice activity detection (TS-VAD), combined with iterative data purification. We also adopted audio domain classification to design domain-dependent processing. Finally, we performed post processing to do system fusion and selection. Our best system achieved DERs of 11.30% in track 1 and 16.78% in track 2 on evaluation set, respectively. 04:42 Fusion-FlowNet: Energy-Efficient Optical Flow Estimation using Sensor Fusion and Deep Fused Spiking-Analog Network Architectures. (arXiv:2103.10592v1 [cs.CV]) Standard frame-based cameras that sample light intensity frames are heavily impacted by motion blur for high-speed motion and fail to perceive scene accurately when the dynamic range is high. Event-based cameras, on the other hand, overcome these limitations by asynchronously detecting the variation in individual pixel intensities. However, event cameras only provide information about pixels in motion, leading to sparse data. Hence, estimating the overall dense behavior of pixels is difficult. To address such issues associated with the sensors, we present Fusion-FlowNet, a sensor fusion framework for energy-efficient optical flow estimation using both frame- and event-based sensors, leveraging their complementary characteristics. Our proposed network architecture is also a fusion of Spiking Neural Networks (SNNs) and Analog Neural Networks (ANNs) where each network is designed to simultaneously process asynchronous event streams and regular frame-based images, respectively. Our network 04:42 Cognitive simulation models for inertial confinement fusion: Combining simulation and experimental data. (arXiv:2103.10590v1 [cs.LG]) The design space for inertial confinement fusion (ICF) experiments is vast and experiments are extremely expensive. Researchers rely heavily on computer simulations to explore the design space in search of high-performing implosions. However, ICF multiphysics codes must make simplifying assumptions, and thus deviate from experimental measurements for complex implosions. For more effective design and investigation, simulations require input from past experimental data to better predict future performance. In this work, we describe a cognitive simulation method for combining simulation and experimental data into a common, predictive model. This method leverages a machine learning technique called transfer learning, the process of taking a model trained to solve one task, and partially retraining it on a sparse dataset to solve a different, but related task. In the context of ICF design, neural network models trained on large simulation databases and partially retrained on experimental 19.03.2021 05:24 Correlations of primary fields in the critical Ising model. (arXiv:2103.10263v1 [math-ph]) We prove convergence of renormalized correlations of primary fields, i. e., spins, disorders, fermions and energy densities, in the scaling limit of the critical Ising model in arbitrary finitely connected domains, with fixed (plus or minus) or free boundary conditions, or mixture thereof. We describe the limits of correlations in terms of solutions of Riemann boundary value problems, and prove their conformal covariance. Moreover, we prove fusion rules, or operator product expansions, which describe asymptotics of the scaling limits of the correlations as some of the points collide together. We give explicit formulae for correlations in the case of simply-connected and doubly-connected domains. Our presentation is self-contained, and the proofs are simplified as compared to the previous work where particular cases are treated. 05:24 Transition from fishbone mode to$\beta$-induced Alfv\'en eigenmode on HL-2A tokamak. (arXiv:2103.10170v1 [physics.plasm-ph]) In the presence of energetic particles (EPs) from auxiliary heating and burning plasmas, fishbone instability and Alfv\'en modes can be excited and their transition can take place in certain overlapping regimes. Using the hybrid kinetic-magnetohydrodynamic model in the NIMROD code, we have identified such a transition between the fishbone instability and the$\beta$-induced Alfv\'en Eigenmode (BAE) for the NBI heated plasmas on HL-2A. When the safety factor at magnetic axis is well below one, typical kink-fishbone transition occurs as the EP fraction increases. When$q_0$is raised to approaching one, the fishbone mode is replaced with BAE for sufficient amount of EPs. When$q_0$is slightly above one, the toroidicity-induced Alfv\'en eigenmode (TAE) dominates at lower EP pressure, whereas BAE dominates at higher EP pressure. 05:24 A Framework for Energy and Carbon Footprint Analysis of Distributed and Federated Edge Learning. (arXiv:2103.10346v1 [cs.LG]) Recent advances in distributed learning raise environmental concerns due to the large energy needed to train and move data to/from data centers. Novel paradigms, such as federated learning (FL), are suitable for decentralized model training across devices or silos that simultaneously act as both data producers and learners. Unlike centralized learning (CL) techniques, relying on big-data fusion and analytics located in energy hungry data centers, in FL scenarios devices collaboratively train their models without sharing their private data. This article breaks down and analyzes the main factors that influence the environmental footprint of FL policies compared with classical CL/Big-Data algorithms running in data centers. The proposed analytical framework takes into account both learning and communication energy costs, as well as the carbon equivalent emissions; in addition, it models both vanilla and decentralized FL policies driven by consensus. The framework is evaluated in an 05:24 Finite-word-length FPGA implementation of model predictive control for ITER resistive wall mode control. (arXiv:2103.10146v1 [eess.SY]) In advanced tokamak scenarios, active feedback control of unstable resistive wall modes (RWM) may be required. A RWM is an instability due to plasma kink at higher plasma pressure, moderated by the presence of a resistive wall surrounding the plasma. We address the dominant kink instability associated with the main nonaxisymmetric (n = 1) RWM, described by the CarMa model. Model predictive control (MPC) is used, with the aim of enlarging the domain of attraction of the unstable RWM modes subject to power-supply voltage constraints. The implementation of MPC is challenging, because the related quadratic programming (QP) on-line optimization problems must be solved at a sub-ms sampling rate. Using complexity-reduction pre-processing techniques and a primal fast gradient method (FGM) QP solver, sufficiently short computation times for ITER are reachable using a standard personal computer (PC). In this work we explore even faster finite-word-length (FWL) implementation using 18.03.2021 08:19 Proposal of A Linked Mirror Configuration for Magnetic Confinement Experiment. (arXiv:2103.09457v1 [physics.plasm-ph]) A new linked mirror device for magnetic confinement experiment is proposed. The new linked mirror device consists of two straight magnetic mirrors connected by two half-torus. The structure of the configuration as a whole is three dimensional because the two linear mirror sections are not parallel. The angle between the two mirror sections generates rotational transform which results in good magnetic confinement of toroidally processing passing particles. In this way the usual loss cone of the traditional linear mirror machines is eliminated. The single particle confinement is similar to that of tokamaks with most of particles well confined. The calculated neoclassical confinement is very good and is even better than that of an equivalent tokamak. Potentially the proposed linked mirror configuration can allow MHD stable high beta equilibria with good plasma confinement suitable for neutron sources and magnetic fusion reactors. 08:19 Nonlinear coupling of reversed shear Alfven eigenmode and toroidal Alfven eigenmode during current ramp. (arXiv:2103.09446v1 [physics.plasm-ph]) Two novel nonlinear mode coupling processes for reversed shear Alfven eigenmode (RSAE) nonlinear saturation are proposed and investigated. In the first process, RSAE nonlinearly couples to a co-propagating toroidal Alfven eigenmode (TAE) with the same toroidal and poloidal mode numbers, and generates a geodesic acoustic mode (GAM). In the second process, RSAE couples to a counter-propagating TAE and generates an ion acoustic wave quasi-mode (IAW). The condition for the two processes to occur is favored during current ramp. Both processes contribute to effectively saturate the Alfvenic instabilities, as well as nonlinearly transfer of energy from energetic fusion alpha particles to fuel ions in burning plasmas. 17.03.2021 04:06 Multiple-isotope pellet cycles captured by turbulent transport modelling in the JET tokamak. (arXiv:2103.09222v1 [physics.plasm-ph]) For the first time the pellet cycle of a multiple-isotope plasma is successfully reproduced with reduced turbulent transport modelling, within an integrated simulation framework. Future nuclear fusion reactors are likely to be fuelled by cryogenic pellet injection, due to higher penetration and faster response times. Accurate pellet cycle modelling is crucial to assess fuelling efficiency and burn control. In recent JET tokamak experiments, deuterium pellets with reactor-relevant deposition characteristics were injected into a pure hydrogen plasma. Measurements of the isotope ratio profile inferred a Deuterium penetration time comparable to the energy confinement time. The modelling successfully reproduces the plasma thermodynamic profiles and the fast deuterium penetration timescale. The predictions of the reduced turbulence model QuaLiKiz in the presence of a negative density gradient following pellet deposition are compared with GENE linear and nonlinear higher fidelity modelling. 04:06 Gumbel-Attention for Multi-modal Machine Translation. (arXiv:2103.08862v1 [cs.CL]) Multi-modal machine translation (MMT) improves translation quality by introducing visual information. However, the existing MMT model ignores the problem that the image will bring information irrelevant to the text, causing much noise to the model and affecting the translation quality. In this paper, we propose a novel Gumbel-Attention for multi-modal machine translation, which selects the text-related parts of the image features. Specifically, different from the previous attention-based method, we first use a differentiable method to select the image information and automatically remove the useless parts of the image features. Through the score matrix of Gumbel-Attention and image features, the image-aware text representation is generated. And then, we independently encode the text representation and the image-aware text representation with the multi-modal encoder. Finally, the final output of the encoder is obtained through multi-modal gated fusion. Experiments and case analysis 16.03.2021 06:45 Fusion yield of plasma with velocity-space anisotropy at constant energy. (arXiv:2103.07834v1 [physics.plasm-ph]) Velocity-space anisotropy can significantly modify fusion reactivity. The nature and magnitude of this modification depends on the plasma temperature, as well as the details of how the anisotropy is introduced. For plasmas that are sufficiently cold compared to the peak of the fusion cross-section, anisotropic distributions tend to have higher yields than isotropic distributions with the same thermal energy. At higher temperatures, it is instead isotropic distributions that have the highest yields. However, the details of this behavior depend on exactly how the distribution differs from an isotropic Maxwellian. This paper describes the effects of anisotropy on fusion yield for the class of anisotropic distribution functions with the same energy distribution as a 3D isotropic Maxwellian, and compares those results with the yields from bi-Maxwellian distributions. In many cases, especially for plasmas somewhat below reactor-regime temperatures, the effects of anisotropy can be 06:45 Effects of Helium massive gas injection level on disruption mitigation on EAST. (arXiv:2103.07645v1 [physics.plasm-ph]) In this study, NIMROD simulations are performed to investigate the effects of massive helium gas injection level on the induced disruption on EAST tokamak. It is demonstrated in simulations that two different scenarios of plasma cooling (complete cooling and partial cooling) take place for different amounts of injected impurities. For the impurity injection above a critical level, a single MHD activity is able to induce a complete core temperature collapse. For impurity injection below the critical level, a series of multiple minor disruptions occur before the complete thermal quench (TQ). 06:45 Bayesian modelling of multiple plasma diagnostics at Wendelstein 7-X. (arXiv:2103.07582v1 [physics.plasm-ph]) Consistent inference of the electron density and temperature has been carried out with multiple heterogeneous plasma diagnostic data sets at Wendelstein 7-X. The predictive models of the interferometer, Thomson scattering and helium beam emission spectroscopy systems have been developed in the Minerva framework and combined to a single joint model. The electron density and temperature profiles are modelled by Gaussian processes with their hyperparameters. The model parameters such as the calibration factor of the Thomson scattering system and the model predictive uncertainties are regarded as additional unknown parameters. The joint posterior probability distribution of the electron density and temperature profiles, hyperparameters of the Gaussian processes and model parameters is explored by Markov chain Monte Carlo algorithms. The posterior samples drawn from the joint posterior distribution are numerically marginalised over the hyperparameters and model parameters to obtain the 06:45 Damage to Relativistic Interstellar Spacecraft by ISM Impact Gas Accumulation. (arXiv:2103.07517v1 [physics.space-ph]) As part of the NASA Starlight collaboration, we look at the implications of radiation effects from impacts with the interstellar medium (ISM) on a directed energy driven relativistic spacecraft. The spacecraft experiences a stream of MeV/nucleon impacts along the forward edge primarily from hydrogen and helium nuclei. The accumulation of implanted slowly diffusing gas atoms in solids drives damage through the meso-scale processes of bubble formation, blistering, and exfoliation. This results in macroscopic changes to material properties and, in the cases of blistering and exfoliation, material erosion via blister rupture and delamination. Relativistic hydrogen and helium at constant velocity will stop in the material at a similar depth, as predicted by Bethe-Bloch stopping and subsequent simulations of the implantation distribution, leading to a mixed hydrogen and helium system similar to that observed in fusion plasma-facing components (PFC's). However, the difference in location of 06:45 Radar Camera Fusion via Representation Learning in Autonomous Driving. (arXiv:2103.07825v1 [cs.CV]) Radars and cameras are mature, cost-effective, and robust sensors and have been widely used in the perception stack of mass-produced autonomous driving systems. Due to their complementary properties, outputs from radar detection (radar pins) and camera perception (2D bounding boxes) are usually fused to generate the best perception results. The key to successful radar-camera fusion is accurate data association. The challenges in radar-camera association can be attributed to the complexity of driving scenes, the noisy and sparse nature of radar measurements, and the depth ambiguity from 2D bounding boxes. Traditional rule-based association methods are susceptible to performance degradation in challenging scenarios and failure in corner cases. In this study, we propose to address rad-cam association via deep representation learning, to explore feature-level interaction and global reasoning. Concretely, we design a loss sampling mechanism and an innovative ordinal loss to overcome the 15.03.2021 09:40 Juggling With Representations: On the Information Transfer Between Imagery, Point Clouds, and Meshes for Multi-Modal Semantics. (arXiv:2103.07348v1 [cs.CV]) The automatic semantic segmentation of the huge amount of acquired remote sensing data has become an important task in the last decade. Images and Point Clouds (PCs) are fundamental data representations, particularly in urban mapping applications. Textured 3D meshes integrate both data representations geometrically by wiring the PC and texturing the surface elements with available imagery. We present a mesh-centered holistic geometry-driven methodology that explicitly integrates entities of imagery, PC and mesh. Due to its integrative character, we choose the mesh as the core representation that also helps to solve the visibility problem for points in imagery. Utilizing the proposed multi-modal fusion as the backbone and considering the established entity relationships, we enable the sharing of information across the modalities imagery, PC and mesh in a two-fold manner: (i) feature transfer and (ii) label transfer. By these means, we achieve to enrich feature vectors to multi-modal 09:40 Semantic Segmentation for Real Point Cloud Scenes via Bilateral Augmentation and Adaptive Fusion. (arXiv:2103.07074v1 [cs.CV]) Given the prominence of current 3D sensors, a fine-grained analysis on the basic point cloud data is worthy of further investigation. Particularly, real point cloud scenes can intuitively capture complex surroundings in the real world, but due to 3D data's raw nature, it is very challenging for machine perception. In this work, we concentrate on the essential visual task, semantic segmentation, for large-scale point cloud data collected in reality. On the one hand, to reduce the ambiguity in nearby points, we augment their local context by fully utilizing both geometric and semantic features in a bilateral structure. On the other hand, we comprehensively interpret the distinctness of the points from multiple resolutions and represent the feature map following an adaptive fusion method at point-level for accurate semantic segmentation. Further, we provide specific ablation studies and intuitive visualizations to validate our key modules. By comparing with state-of-the-art networks on 12.03.2021 10:54 Integrated Age Estimation Mechanism. (arXiv:2103.06546v1 [cs.CV]) Machine-learning-based age estimation has received lots of attention. Traditional age estimation mechanism focuses estimation age error, but ignores that there is a deviation between the estimated age and real age due to disease. Pathological age estimation mechanism the author proposed before introduces age deviation to solve the above problem and improves classification capability of the estimated age significantly. However,it does not consider the age estimation error of the normal control (NC) group and results in a larger error between the estimated age and real age of NC group. Therefore, an integrated age estimation mechanism based on Decision-Level fusion of error and deviation orientation model is proposed to solve the problem.Firstly, the traditional age estimation and pathological age estimation mechanisms are weighted together.Secondly, their optimal weights are obtained by minimizing mean absolute error (MAE) between the estimated age and real age of normal people. In the 11.03.2021 19:11 Not so fast, supernova: Highest-energy cosmic rays detected in star clusters For decades, researchers assumed the cosmic rays that regularly bombard Earth from the far reaches of the galaxy are born when stars go supernova—when they grow too massive to support the fusion occurring at their cores and explode. 06:31 The m/n=1/1 Instabilities in An EAST NBI Discharge. (arXiv:2103.05845v1 [physics.plasm-ph]) An analysis of precession fishbone, diamagnetic fishbone and internal kink mode in Tokamak plasmas are presented via solving the fishbone dispersion relation. Applying the dispersion relation to EAST discharge experiment, excitation of precession fishbone due to Neutral Beam Injection is successfully explained. Real frequency and growth rate of diamagnetic fishbone and internal kink mode are calculated. 10.03.2021 09:12 HemCNN: Deep Learning enables decoding of fNIRS cortical signals in hand grip motor tasks. (arXiv:2103.05338v1 [cs.LG]) We solve the fNIRS left/right hand force decoding problem using a data-driven approach by using a convolutional neural network architecture, the HemCNN. We test HemCNN's decoding capabilities to decode in a streaming way the hand, left or right, from fNIRS data. HemCNN learned to detect which hand executed a grasp at a naturalistic hand action speed of$~1\,$Hz, outperforming standard methods. Since HemCNN does not require baseline correction and the convolution operation is invariant to time translations, our method can help to unlock fNIRS for a variety of real-time tasks. Mobile brain imaging and mobile brain machine interfacing can benefit from this to develop real-world neuroscience and practical human neural interfacing based on BOLD-like signals for the evaluation, assistance and rehabilitation of force generation, such as fusion of fNIRS with EEG signals. 09:12 DISC: A Dynamic Shape Compiler for Machine Learning Workloads. (arXiv:2103.05288v1 [cs.DC]) Many recent machine learning models show dynamic shape characteristics. However, existing AI compiler optimization systems suffer a lot from problems brought by dynamic shape models, including compilation overhead, memory usage, optimization pipeline and deployment complexity. This paper provides a compiler system to natively support optimization for dynamic shape workloads, named DISC. DISC enriches a set of IR to form a fully dynamic shape representation. It generates the runtime flow at compile time to support processing dynamic shape based logic, which avoids the interpretation overhead at runtime and enlarges the opportunity of host-device co-optimization. It addresses the kernel fusion problem of dynamic shapes with shape propagation and constraints collecting methods. This is the first work to demonstrate how to build an end-to-end dynamic shape compiler based on MLIR infrastructure. Experiments show that DISC achieves up to 3.3x speedup than TensorFlow/PyTorch, and 1.8x than 09:12 Memory-Efficient, Limb Position-Aware Hand Gesture Recognition using Hyperdimensional Computing. (arXiv:2103.05267v1 [cs.LG]) Electromyogram (EMG) pattern recognition can be used to classify hand gestures and movements for human-machine interface and prosthetics applications, but it often faces reliability issues resulting from limb position change. One method to address this is dual-stage classification, in which the limb position is first determined using additional sensors to select between multiple position-specific gesture classifiers. While improving performance, this also increases model complexity and memory footprint, making a dual-stage classifier difficult to implement in a wearable device with limited resources. In this paper, we present sensor fusion of accelerometer and EMG signals using a hyperdimensional computing model to emulate dual-stage classification in a memory-efficient way. We demonstrate two methods of encoding accelerometer features to act as keys for retrieval of position-specific parameters from multiple models stored in superposition. Through validation on a dataset of 13 09.03.2021 11:58 RFN-Nest: An end-to-end residual fusion network for infrared and visible images. (arXiv:2103.04286v1 [cs.CV]) In the image fusion field, the design of deep learning-based fusion methods is far from routine. It is invariably fusion-task specific and requires a careful consideration. The most difficult part of the design is to choose an appropriate strategy to generate the fused image for a specific task in hand. Thus, devising learnable fusion strategy is a very challenging problem in the community of image fusion. To address this problem, a novel end-to-end fusion network architecture (RFN-Nest) is developed for infrared and visible image fusion. We propose a residual fusion network (RFN) which is based on a residual architecture to replace the traditional fusion approach. A novel detail-preserving loss function, and a feature enhancing loss function are proposed to train RFN. The fusion model learning is accomplished by a novel two-stage training strategy. In the first stage, we train an auto-encoder based on an innovative nest connection (Nest) concept. Next, the RFN is trained using the 05.03.2021 08:30 New plasma regimes with small ELMs and high confinement at the Joint European Torus. (arXiv:2103.02679v1 [physics.plasm-ph]) New plasma regimes with high confinement, low core impurity accumulation and small Edge localized mode (ELMs) perturbations have been obtained close to ITER conditions in magnetically confined plasmas from the Joint European torus (JET) tokamak. Such regimes are achieved by means of optimized particle fuelling conditions which trigger a self-organize state with a strong increase in rotation and ion temperature and a decrease of the edge density. An interplay between core and edge plasma regions leads to reduced turbulence levels and outward impurity convection. These results pave the way to an attractive alternative to the standard plasmas considered for fusion energy generation in a tokamak with metallic wall environment such as the ones expected in ITER 08:30 Model-based image adjustment for a successful pansharpening. (arXiv:2103.03062v1 [eess.IV]) A new model-based image adjustment for the enhancement of multi-resolution image fusion or pansharpening is proposed. Such image adjustment is needed for most pansharpening methods using panchromatic band and/or intensity image (calculated as a weighted sum of multispectral bands) as an input. Due various reasons, e.g. calibration inaccuracies, usage of different sensors, input images for pansharpening: low resolution multispectral image or more precisely the calculated intensity image and high resolution panchromatic image may differ in values of their physical properties, e.g. radiances or reflectances depending on the processing level. But the same objects/classes in both images should exhibit similar values or more generally similar statistics. Similarity definition will depend on a particular application. For a successful fusion of data from two sensors the energy balance between radiances/reflectances of both sensors should hold. A virtual band is introduced to compensate for 08:30 A Novel Context-Aware Multimodal Framework for Persian Sentiment Analysis. (arXiv:2103.02636v1 [cs.CL]) Most recent works on sentiment analysis have exploited the text modality. However, millions of hours of video recordings posted on social media platforms everyday hold vital unstructured information that can be exploited to more effectively gauge public perception. Multimodal sentiment analysis offers an innovative solution to computationally understand and harvest sentiments from videos by contextually exploiting audio, visual and textual cues. In this paper, we, firstly, present a first of its kind Persian multimodal dataset comprising more than 800 utterances, as a benchmark resource for researchers to evaluate multimodal sentiment analysis approaches in Persian language. Secondly, we present a novel context-aware multimodal sentiment analysis framework, that simultaneously exploits acoustic, visual and textual cues to more accurately determine the expressed sentiment. We employ both decision-level (late) and feature-level (early) fusion methods to integrate affective cross-modal 01:19 Nuclear engineering researchers develop new resilient oxide dispersion strengthened alloy Researchers have recently shown superior performance of a new oxide dispersion strengthened (ODS) alloy they developed for use in both fission and fusion reactors. 04.03.2021 23:00 Nuclear engineering researchers develop new resilient oxide dispersion strengthened alloy Texas A&M University researchers have recently shown superior performance of a new oxide dispersion strengthened (ODS) alloy they developed for use in both fission and fusion reactors. 19:41 Physicists trap ultracold plasma in a magnetic bottle for the 1st time The breakthrough technique supercools plasma with lasers before trapping them in a magnetic field; allowing physicists to study the northern lights, white dwarves and nuclear fusion in ever greater detail. 03.03.2021 17:02 Fusion startup plans reactor with small but powerful superconducting magnets Commonwealth Fusion Systems announces site for compact reactor 06:56 Quantifying experimental edge plasma evolution via multidimensional adaptive Gaussian process regression. (arXiv:2103.01305v1 [physics.plasm-ph]) The edge density and temperature of tokamak plasmas are strongly correlated with energy and particle confinement and their quantification is fundamental to understanding edge dynamics. These quantities exhibit behaviours ranging from sharp plasma gradients and fast transient phenomena (e.g. transitions between low and high confinement regimes) to nominal stationary phases. Analysis of experimental edge measurements therefore require robust fitting techniques to capture potentially stiff spatiotemporal evolution. Additionally, fusion plasma diagnostics inevitably involve measurement errors and data analysis requires a statistical framework to accurately quantify uncertainties. This paper outlines a generalized multidimensional adaptive Gaussian process routine capable of automatically handling noisy data and spatiotemporal correlations. We focus on the edge-pedestal region in order to underline advancements in quantifying time-dependent plasma profiles including transport barrier 02.03.2021 18:34 Nuclear fusion: Building a star on Earth is hard, which is why we need better materials Nuclear fusion is the process that powers the Sun and all other stars. During fusion, the nuclei of two atoms are brought close enough together that they fuse together, releasing huge amounts of energy. 09:29 Simulation of runaway electron generation in fusion grade tokamak and suppression by impurity injection. (arXiv:2103.00325v1 [physics.plasm-ph]) During disruptions in fusion-grade tokamaks like ITER, large electric fields are induced following the thermal quench (TQ) period which can generate a substantial amount of Runaway Electrons (REs) that can carry up to 10 MA current with energies as high as several tens of MeV [1-3] in current quench phase (CQ). These runaway electrons can cause significant damage to the plasma-facing-components due to their localized energy deposition. To mitigate these effects, impurity injections of high-Z atoms have been proposed [1-3]. In this paper, we use a self-consistent 0D tokamak disruption model as implemented in PREDICT code [6] which has been upgraded to take into account the effect of impurity injections on RE dynamics as suggested in [4-5]. Dominant RE generation mechanisms such as the secondary avalanche mechanism as well as primary RE-generation mechanisms namely Dreicer, hot-tail, tritium decay and Compton scattering (from {\gamma}-rays emitted from activated walls) have been taken 09:29 The Labeled Multiple Canonical Correlation Analysis for Information Fusion. (arXiv:2103.00359v1 [cs.CV]) The objective of multimodal information fusion is to mathematically analyze information carried in different sources and create a new representation which will be more effectively utilized in pattern recognition and other multimedia information processing tasks. In this paper, we introduce a new method for multimodal information fusion and representation based on the Labeled Multiple Canonical Correlation Analysis (LMCCA). By incorporating class label information of the training samples,the proposed LMCCA ensures that the fused features carry discriminative characteristics of the multimodal information representations, and are capable of providing superior recognition performance. We implement a prototype of LMCCA to demonstrate its effectiveness on handwritten digit recognition,face recognition and object recognition utilizing multiple features,bimodal human emotion recognition involving information from both audio and visual domains. The generic nature of LMCCA allows it to take as 01.03.2021 05:13 Zoetrope Genetic Programming for Regression. (arXiv:2102.13388v1 [stat.ML]) The Zoetrope Genetic Programming (ZGP) algorithm is based on an original representation for mathematical expressions, targeting evolutionary symbolic regression.The zoetropic representation uses repeated fusion operations between partial expressions, starting from the terminal set. Repeated fusions within an individual gradually generate more complex expressions, ending up in what can be viewed as new features. These features are then linearly combined to best fit the training data. ZGP individuals then undergo specific crossover and mutation operators, and selection takes place between parents and offspring. ZGP is validated using a large number of public domain regression datasets, and compared to other symbolic regression algorithms, as well as to traditional machine learning algorithms. ZGP reaches state-of-the-art performance with respect to both types of algorithms, and demonstrates a low computational time compared to other symbolic regression approaches. 05:13 Zoetrope Genetic Programming for Regression. (arXiv:2102.13388v1 [stat.ML]) The Zoetrope Genetic Programming (ZGP) algorithm is based on an original representation for mathematical expressions, targeting evolutionary symbolic regression.The zoetropic representation uses repeated fusion operations between partial expressions, starting from the terminal set. Repeated fusions within an individual gradually generate more complex expressions, ending up in what can be viewed as new features. These features are then linearly combined to best fit the training data. ZGP individuals then undergo specific crossover and mutation operators, and selection takes place between parents and offspring. ZGP is validated using a large number of public domain regression datasets, and compared to other symbolic regression algorithms, as well as to traditional machine learning algorithms. ZGP reaches state-of-the-art performance with respect to both types of algorithms, and demonstrates a low computational time compared to other symbolic regression approaches. 25.02.2021 16:28 Scientists use supercomputers to study reliable fusion reactor design, operation Nuclear fusion, the same kind of energy that fuels stars, could one day power our world with abundant, safe, and carbon-free energy. Aided by supercomputers Summit at the US Department of Energy's (DOE's) Oak Ridge National Laboratory (ORNL) and Theta at DOE's Argonne National Laboratory (ANL), a team of scientists strives toward making fusion energy a reality. 13:47 An aggressive market-driven model for US fusion power development Electricity generated by fusion power plants could play an important role in decarbonizing the U.S. energy sector by 08:20 Position Location for Futuristic Cellular Communications -- 5G and Beyond. (arXiv:2102.12074v1 [cs.IT]) With vast mmWave spectrum and narrow beam antenna technology, precise position location is now possible in 5G and future mobile communication systems. In this article, we describe how centimeterlevel localization accuracy can be achieved, particularly through the use of map-based techniques. We show how data fusion of parallel information streams, machine learning, and cooperative localization techniques further improve positioning accuracy. 08:20 DeepCervix: A Deep Learning-based Framework for the Classification of Cervical Cells Using Hybrid Deep Feature Fusion Techniques. (arXiv:2102.12191v1 [eess.IV]) Cervical cancer, one of the most common fatal cancers among women, can be prevented by regular screening to detect any precancerous lesions at early stages and treat them. Pap smear test is a widely performed screening technique for early detection of cervical cancer, whereas this manual screening method suffers from high false-positive results because of human errors. To improve the manual screening practice, machine learning (ML) and deep learning (DL) based computer-aided diagnostic (CAD) systems have been investigated widely to classify cervical pap cells. Most of the existing researches require pre-segmented images to obtain good classification results, whereas accurate cervical cell segmentation is challenging because of cell clustering. Some studies rely on handcrafted features, which cannot guarantee the classification stage's optimality. Moreover, DL provides poor performance for a multiclass classification task when there is an uneven distribution of data, which is prevalent 08:20 Position Location for Futuristic Cellular Communications -- 5G and Beyond. (arXiv:2102.12074v1 [cs.IT]) With vast mmWave spectrum and narrow beam antenna technology, precise position location is now possible in 5G and future mobile communication systems. In this article, we describe how centimeterlevel localization accuracy can be achieved, particularly through the use of map-based techniques. We show how data fusion of parallel information streams, machine learning, and cooperative localization techniques further improve positioning accuracy. 23.02.2021 15:01 Daily briefing: Fusion reactor set for fuel test run 06:29 Stochastic fluctuation and transport in the edge tokamak plasmas with the resonant magnetic perturbation field. (arXiv:2102.10733v1 [physics.plasm-ph]) The stochastic layer formation by the penetration of the resonant magnetic perturbation (RMP) field has been considered as a key mechanism in the RMP control of the edge localized mode (ELM) in tokamak plasmas. Here, we provide experimental observations that the fluctuation and transport in the edge plasmas become more stochastic with the more penetration of the RMP field into the plasma. The results support the importance of the stochastic layer formation in the RMP ELM control experiments. 22.02.2021 17:24 Heat loss control method in fusion reactors The core of a fusion reactor is incredibly hot. Hydrogen that inevitably escapes from it must be cooled on its way to the wall, as otherwise, the reactor wall would be damaged. Researchers from the Dutch institute DIFFER and EPFL's Swiss Plasma Center have developed a strict measurement and control method for the cooling of very hot particles escaping from fusion plasmas. 15:22 Fuel for world’s largest fusion reactor ITER is set for test run 19.02.2021 09:30 Federated Depression Detection from Multi-SourceMobile Health Data. (arXiv:2102.09342v1 [cs.CY]) Depression is one of the most common mental illness problems, and the symptoms shown by patients are not consistent, making it difficult to diagnose in the process of clinical practice and pathological research.Although researchers hope that artificial intelligence can contribute to the diagnosis and treatment of depression, the traditional centralized machine learning needs to aggregate patient data, and the data privacy of patients with mental illness needs to be strictly confidential, which hinders machine learning algorithms clinical application.To solve the problem of privacy of the medical history of patients with depression, we implement federated learning to analyze and diagnose depression. First, we propose a general multi-view federated learning framework using multi-source data,which can extend any traditional machine learning model to support federated learning across different institutions or parties.Secondly, we adopt late fusion methods to solve the problem of 18.02.2021 05:08 A Mutual Reference Shape for Segmentation Fusion and Evaluation. (arXiv:2102.08939v1 [eess.IV]) This paper proposes the estimation of a mutual shape from a set of different segmentation results using both active contours and information theory. The mutual shape is here defined as a consensus shape estimated from a set of different segmentations of the same object. In an original manner, such a shape is defined as the minimum of a criterion that benefits from both the mutual information and the joint entropy of the input segmentations. This energy criterion is justified using similarities between information theory quantities and area measures, and presented in a continuous variational framework. In order to solve this shape optimization problem, shape derivatives are computed for each term of the criterion and interpreted as an evolution equation of an active contour. A mutual shape is then estimated together with the sensitivity and specificity of each segmentation. Some synthetic examples allow us to cast the light on the difference between the mutual shape and an average 17.02.2021 07:43 An AutoML-based Approach to Multimodal Image Sentiment Analysis. (arXiv:2102.08092v1 [cs.LG]) Sentiment analysis is a research topic focused on analysing data to extract information related to the sentiment that it causes. Applications of sentiment analysis are wide, ranging from recommendation systems, and marketing to customer satisfaction. Recent approaches evaluate textual content using Machine Learning techniques that are trained over large corpora. However, as social media grown, other data types emerged in large quantities, such as images. Sentiment analysis in images has shown to be a valuable complement to textual data since it enables the inference of the underlying message polarity by creating context and connections. Multimodal sentiment analysis approaches intend to leverage information of both textual and image content to perform an evaluation. Despite recent advances, current solutions still flounder in combining both image and textual information to classify social media data, mainly due to subjectivity, inter-class homogeneity and fusion data differences. In 16.02.2021 10:18 INGRID: an interactive grid generator for 2D edge plasma modeling. (arXiv:2102.07040v1 [physics.plasm-ph]) A new grid generator for tokamak boundary plasma is presented. INGRID (Interactive grid generator) is a Python-based code that can generate grids for boundary plasma modeling for variety of configurations with one or two X-points in the domain. This includes single-null, double-null, and variety of topologically near-snowflake configurations. Users can generate grids through utilization of the INGRID Python package: either interactively via a parameter-file driven GUI, or by embedding in a non-interactive script-controlled workflow. The INGRID workflow consists of three phases: (i) import of magnetic field data and analysis of the magnetic configuration, (ii) creation of a skeleton grid ("Patch Map") via line tracing method, and (iii) refinement of the prototypical grid into a final grid. Performance of INGRID is analyzed by comparison with an existing grid generator. The results demonstrate INGRID's capability to produce grids for boundary plasma modeling in a robust and user-friendly 10:18 Affective State Recognition through EEG Signals Feature Level Fusion and Ensemble Classifier. (arXiv:2102.07127v1 [cs.HC]) Human affects are complex paradox and an active research domain in affective computing. Affects are traditionally determined through a self-report based psychometric questionnaire or through facial expression recognition. However, few state-of-the-arts pieces of research have shown the possibilities of recognizing human affects from psychophysiological and neurological signals. In this article, electroencephalogram (EEG) signals are used to recognize human affects. The electroencephalogram (EEG) of 100 participants are collected where they are given to watch one-minute video stimuli to induce different affective states. The videos with emotional tags have a variety range of affects including happy, sad, disgust, and peaceful. The experimental stimuli are collected and analyzed intensively. The interrelationship between the EEG signal frequencies and the ratings given by the participants are taken into consideration for classifying affective states. Advanced feature extraction 10:18 Fusion of convolution neural network, support vector machine and Sobel filter for accurate detection of COVID-19 patients using X-ray images. (arXiv:2102.06883v1 [eess.IV]) The coronavirus (COVID-19) is currently the most common contagious disease which is prevalent all over the world. The main challenge of this disease is the primary diagnosis to prevent secondary infections and its spread from one person to another. Therefore, it is essential to use an automatic diagnosis system along with clinical procedures for the rapid diagnosis of COVID-19 to prevent its spread. Artificial intelligence techniques using computed tomography (CT) images of the lungs and chest radiography have the potential to obtain high diagnostic performance for Covid-19 diagnosis. In this study, a fusion of convolutional neural network (CNN), support vector machine (SVM), and Sobel filter is proposed to detect COVID-19 using X-ray images. A new X-ray image dataset was collected and subjected to high pass filter using a Sobel filter to obtain the edges of the images. Then these images are fed to CNN deep learning model followed by SVM classifier with ten-fold cross validation 15.02.2021 05:59 Towards DeepSentinel: An extensible corpus of labelled Sentinel-1 and -2 imagery and a general-purpose sensor-fusion semantic embedding model. (arXiv:2102.06260v1 [cs.CV]) Earth observation offers new insight into anthropogenic changes to nature, and how these changes are effecting (and are effected by) the built environment and the real economy. With the global availability of medium-resolution (10-30m) synthetic aperture radar (SAR) Sentinel-1 and multispectral Sentinel-2 imagery, machine learning can be employed to offer these insights at scale, unbiased to the reporting of companies and countries. In this paper, I introduce DeepSentinel, a data pipeline and experimentation framework for producing general-purpose semantic embeddings of paired Sentinel-1 and Sentinel-2 imagery. I document the development of an extensible corpus of labelled and unlabelled imagery for the purposes of sensor fusion research. With this new dataset I develop a set of experiments applying popular self-supervision methods and encoder architectures to a land cover classification problem. Tile2vec spatial encoding with a self-attention enabled ResNet model outperforms deeper 13.02.2021 23:40 New machine learning theory raises questions about nature of science A novel computer algorithm, or set of rules, that accurately predicts the orbits of planets in the solar system could be adapted to better predict and control the behavior of the plasma that fuels fusion facilities designed to harvest on Earth the fusion energy that powers the sun and stars. 12.02.2021 18:35 New machine learning theory raises questions about nature of science A novel computer algorithm, or set of rules, that accurately predicts the orbits of planets in the solar system could be adapted to better predict and control the behavior of the plasma that fuels fusion facilities designed to harvest on Earth the fusion energy that powers the sun and stars. 06:35 Comparing spontaneous and pellet-triggered ELMs via non-linear extended MHD simulations. (arXiv:2102.05850v1 [physics.plasm-ph]) Injecting frozen deuterium pellets into an ELMy H-mode plasma is a well established scheme for triggering edge localized modes (ELMs) before they naturally occur. Based on an ASDEX Upgrade H-mode plasma, this article presents a comparison of extended MHD simulations of spontaneous type-I ELMs and pellet-triggered ELMs allowing to study their non-linear dynamics in detail. In particular, pellet-triggered ELMs are simulated by injecting deuterium pellets into different time points during the pedestal build-up described in [A. Cathey et al. Nuclear Fusion 60, 124007 (2020)]. Realistic ExB and diamagnetic background plasma flows as well as the time dependent bootstrap current evolution are included during the build-up to capture the balance between stabilising and destabilising terms for the edge instabilities accurately. Dependencies on the pellet size and injection times are studied. The spatio-temporal structures of the modes and the resulting divertor heat fluxes are compared in detail 10.02.2021 11:41 On the Universal Transformation of Data-Driven Models to Control Systems. (arXiv:2102.04722v1 [math.OC]) As in almost every other branch of science, the major advances in data science and machine learning have also resulted in significant improvements regarding the modeling and simulation of nonlinear dynamical systems. It is nowadays possible to make accurate medium to long-term predictions of highly complex systems such as the weather, the dynamics within a nuclear fusion reactor, of disease models or the stock market in a very efficient manner. In many cases, predictive methods are advertised to ultimately be useful for control, as the control of high-dimensional nonlinear systems is an engineering grand challenge with huge potential in areas such as clean and efficient energy production, or the development of advanced medical devices. However, the question of how to use a predictive model for control is often left unanswered due to the associated challenges, namely a significantly higher system complexity, the requirement of much larger data sets and an increased and often 11:41 Drift reduced Landau fluid model for magnetized plasma turbulence simulations in BOUT++ framework. (arXiv:2102.04976v1 [physics.plasm-ph]) Recently the drift-reduced Landau fluid six-field turbulence model within the BOUT++ framework has been upgraded. In particular, this new model employs a new normalization, adds a volumetric flux-driven source option, the Landau fluid closure for parallel heat flux and a Laplacian inversion solver which is able to capture n=0 axisymmetric mode evolution in realistic tokamak configurations. These improvements substantially extended model's capability to study a wider range of tokamak edge phenomena, and are essential to build a fully self-consistent edge turbulence model capable of both transient (e.g., ELM, disruption) and transport time-scale simulations. 11:41 On the Universal Transformation of Data-Driven Models to Control Systems. (arXiv:2102.04722v1 [math.OC]) As in almost every other branch of science, the major advances in data science and machine learning have also resulted in significant improvements regarding the modeling and simulation of nonlinear dynamical systems. It is nowadays possible to make accurate medium to long-term predictions of highly complex systems such as the weather, the dynamics within a nuclear fusion reactor, of disease models or the stock market in a very efficient manner. In many cases, predictive methods are advertised to ultimately be useful for control, as the control of high-dimensional nonlinear systems is an engineering grand challenge with huge potential in areas such as clean and efficient energy production, or the development of advanced medical devices. However, the question of how to use a predictive model for control is often left unanswered due to the associated challenges, namely a significantly higher system complexity, the requirement of much larger data sets and an increased and often 09.02.2021 21:36 US Navy Has Patents on Tech It Says Will Engineer the Fabric of Reality (Slashdot) SlashdotUS Navy Has Patents on Tech It Says Will Engineer the Fabric of Reality - The U.S. Navy has patents on weird and little understood technology. According to patents filed by the Navy, it is working on a compact fusion reactor that could power cities, an engine that works using inertial mass reduction, and a hybrid aerospace-underwater craft. From a report: Dubbed the UFO patents, ... 08.02.2021 21:41 Scientists propose lithium to cope with high-risk condition in future fusion facilities Perhaps the greatest technological challenge to harvesting on Earth the fusion energy that powers the sun and stars in future tokamak fusion reactors will be controlling the extreme heat that could strike the exhaust system inside the devices. Such heat flow, or flux, could seriously damage the walls of the divertor at the heart of the exhaust system and shut down fusion reactions in the doughnut-shaped facilities. 09:52 Sparse Normal Means Estimation with Sublinear Communication. (arXiv:2102.03060v1 [stat.ML]) We consider the problem of sparse normal means estimation in a distributed setting with communication constraints. We assume there are$M$machines, each holding a$d$-dimensional observation of a$K$-sparse vector$\mu$corrupted by additive Gaussian noise. A central fusion machine is connected to the$M$machines in a star topology, and its goal is to estimate the vector$\mu$with a low communication budget. Previous works have shown that to achieve the centralized minimax rate for the$\ell_2$risk, the total communication must be high - at least linear in the dimension$d$. This phenomenon occurs, however, at very weak signals. We show that once the signal-to-noise ratio (SNR) is slightly higher, the support of$\mu$can be correctly recovered with much less communication. Specifically, we present two algorithms for the distributed sparse normal means problem, and prove that above a certain SNR threshold, with high probability, they recover the correct support with total 09:52 Sparse Normal Means Estimation with Sublinear Communication. (arXiv:2102.03060v1 [stat.ML]) We consider the problem of sparse normal means estimation in a distributed setting with communication constraints. We assume there are$M$machines, each holding a$d$-dimensional observation of a$K$-sparse vector$\mu$corrupted by additive Gaussian noise. A central fusion machine is connected to the$M$machines in a star topology, and its goal is to estimate the vector$\mu$with a low communication budget. Previous works have shown that to achieve the centralized minimax rate for the$\ell_2$risk, the total communication must be high - at least linear in the dimension$d$. This phenomenon occurs, however, at very weak signals. We show that once the signal-to-noise ratio (SNR) is slightly higher, the support of$\mu$can be correctly recovered with much less communication. Specifically, we present two algorithms for the distributed sparse normal means problem, and prove that above a certain SNR threshold, with high probability, they recover the correct support with total 09:52 Nonlinear MHD simulations of external kinks in quasi-axisymmetric stellarators using an axisymmetric external rotational transform approximation. (arXiv:2102.03080v1 [physics.plasm-ph]) Reduced magnetohydrodynamic (MHD) equations are used to study the nonlinear dynamics of external kinks in quasi-axisymmetric (QA) stellarators with varying fractions of external rotational transform. The large bootstrap currents associated with high beta plasmas typically make QA configurations susceptible to low n external modes, limiting their operational space. The violence of the nonlinear dynamics, and, in particular, when these modes lead to a disruption, is not yet understood. In this paper, the nonlinear phase of external kinks in an unstable QA configuration with an edge safety factor below two is simulated. The nonlinear MHD code JOREK is used, validating the linear dynamics against CASTOR3D. The external rotational transform is approximated axisymmetrically, and varied from the tokamak limit to understand its influence on the nonlinear dynamics. The violence of the kink instability is quantified, and shown to reduce with the increasing external field. At the same time, 09:52 Sparse Normal Means Estimation with Sublinear Communication. (arXiv:2102.03060v1 [stat.ML]) We consider the problem of sparse normal means estimation in a distributed setting with communication constraints. We assume there are$M$machines, each holding a$d$-dimensional observation of a$K$-sparse vector$\mu$corrupted by additive Gaussian noise. A central fusion machine is connected to the$M$machines in a star topology, and its goal is to estimate the vector$\mu$with a low communication budget. Previous works have shown that to achieve the centralized minimax rate for the$\ell_2$risk, the total communication must be high - at least linear in the dimension$d$. This phenomenon occurs, however, at very weak signals. We show that once the signal-to-noise ratio (SNR) is slightly higher, the support of$\mu\$ can be correctly recovered with much less communication. Specifically, we present two algorithms for the distributed sparse normal means problem, and prove that above a certain SNR threshold, with high probability, they recover the correct support with total

07.02.2021
10:24 New fiber optic temperature sensing approach to keep fusion power plants running

The pursuit of fusion as a safe, carbon-free, always-on energy source has intensified in recent years, with a

05.02.2021
17:36 New fiber optic temperature sensing approach to keep fusion power plants running

The pursuit of fusion as a safe, carbon-free, always-on energy source has intensified in recent years, with a number of organizations pursuing aggressive timelines for technology demonstrations and power plant designs. New-generation superconducting magnets are a critical enabler for many of these programs, which creates growing need for sensors, controls, and other infrastructure that will allow the magnets to operate reliably in the harsh conditions of a commercial fusion power plant.

10:55 Big Data Analytics Applying the Fusion Approach of Multicriteria Decision Making with Deep Learning Algorithms. (arXiv:2102.02637v1 [cs.LG])

Data is evolving with the rapid progress of population and communication for various types of devices such as networks, cloud computing, Internet of Things (IoT), actuators, and sensors. The increment of data and communication content goes with the equivalence of velocity, speed, size, and value to provide the useful and meaningful knowledge that helps to solve the future challenging tasks and latest issues. Besides, multicriteria based decision making is one of the key issues to solve for various issues related to the alternative effects in big data analysis. It tends to find a solution based on the latest machine learning techniques that include algorithms like decision making and deep learning mechanism based on multicriteria in providing insights to big data. On the other hand, the derivations are made for it to go with the approximations to increase the duality of runtime and improve the entire system's potentiality and efficacy. In essence, several fields, including business,

04.02.2021
06:26 Unbox the Black-box for the Medical Explainable AI via Multi-modal and Multi-centre Data Fusion: A Mini-Review, Two Showcases and Beyond. (arXiv:2102.01998v1 [cs.AI])

Explainable Artificial Intelligence (XAI) is an emerging research topic of machine learning aimed at unboxing how AI systems' black-box choices are made. This research field inspects the measures and models involved in decision-making and seeks solutions to explain them explicitly. Many of the machine learning algorithms can not manifest how and why a decision has been cast. This is particularly true of the most popular deep neural network approaches currently in use. Consequently, our confidence in AI systems can be hindered by the lack of explainability in these black-box models. The XAI becomes more and more crucial for deep learning powered applications, especially for medical and healthcare studies, although in general these deep neural networks can return an arresting dividend in performance. The insufficient explainability and transparency in most existing AI systems can be one of the major reasons that successful implementation and integration of AI tools into routine clinical

06:26 Unbox the Black-box for the Medical Explainable AI via Multi-modal and Multi-centre Data Fusion: A Mini-Review, Two Showcases and Beyond. (arXiv:2102.01998v1 [cs.AI])

Explainable Artificial Intelligence (XAI) is an emerging research topic of machine learning aimed at unboxing how AI systems' black-box choices are made. This research field inspects the measures and models involved in decision-making and seeks solutions to explain them explicitly. Many of the machine learning algorithms can not manifest how and why a decision has been cast. This is particularly true of the most popular deep neural network approaches currently in use. Consequently, our confidence in AI systems can be hindered by the lack of explainability in these black-box models. The XAI becomes more and more crucial for deep learning powered applications, especially for medical and healthcare studies, although in general these deep neural networks can return an arresting dividend in performance. The insufficient explainability and transparency in most existing AI systems can be one of the major reasons that successful implementation and integration of AI tools into routine clinical

02.02.2021
11:52 Information fusion between knowledge and data in Bayesian network structure learning. (arXiv:2102.00473v1 [cs.AI])

Bayesian Networks (BNs) have become a powerful technology for reasoning under uncertainty, particularly in areas that require causal assumptions that enable us to simulate the effect of intervention. The graphical structure of these models can be determined by causal knowledge, learnt from data, or a combination of both. While it seems plausible that the best approach in constructing a causal graph involves combining knowledge with machine learning, this approach remains underused in practice. This paper describes and evaluates a set of information fusion methods that have been implemented in the open-source Bayesys structure learning system. The methods enable users to specify pre-existing knowledge and rule-based information that can be obtained from heterogeneous sources, to constrain or guide structure learning. Each method is assessed in terms of structure learning impact, including graphical accuracy, model fitting, complexity and runtime. The results are illustrated both with

01.02.2021
18:12 America on a new fast track to fusion energy

The giant appropriations bill signed by US President Donald Trump on December 27 contains a little-noticed section that sets the goal of creating a full-fledged fusion industry, a new industrial sector centered on the commercialization of nuclear fusion as an energy source. The envisioned model is the successful public-private partnership that built up America’s commercial space industry. In an exclusive interview with Asia Times, Fusion Industry Association director Andrew Holland tells the inside story to correspondent Jonathan Tenennbaum.

07:31 Development of New Tracking Detector with Fine-grained Nuclear Emulsion for sub-MeV Neutron Measurement. (arXiv:2101.12424v1 [physics.ins-det])

In this study, we have developed a new sub-MeV neutron detector that has a high position resolution, energy resolution, directional sensitivity, and low background. The detector is based on a super-fine-grained nuclear emulsion, called the Nano Imaging Tracker (NIT), and it is capable of detecting neutron induced proton recoils as tracks through topological analysis with sub-micrometric accuracy. We used a type of NIT with AgBr:I crystals of (98 +- 10) nm size dispersed in the gelatin. First, we calibrated the performance of NIT device for detecting monochromatic neutrons with sub-MeV energy generated by nuclear fusion reactions, and the detection efficiency for recoil proton tracks of more than 2 um range was consistently 100\% (the 1 sigma lower limit was 83%) in accordance with expectations by manual based analysis. In addition, recoil energy and angle distribution obtained good agreement with kinematical expectation. The primary neutron energy was reconstructed by using them, and