To improve the ability of particle identification of the RIBLL2 separator at the HIRFL-CSR complex, a new high-performance detector for measuring fragment starting time and position at the F1 dispersive plane has been constructed and installed, and a
method for achieving precise Br{ho} determination has been developed using the experimentally derived ion-optical transfer matrix elements from the measured position and ToF information. Using the high-performance detectors and the precise Br{ho} determination method, the fragments produced by the fragmentation of 78Kr at 300 MeV/nucleon were identified clearly at the RIBLL2-ETF under full momentum acceptance. The atomic number Z resolution of {sigma}Z~0.19 and the mass-to-charge ratio A/Q resolution of {sigma}A/Q~5.8e-3 were obtained for the 75As33+ fragment. This great improvement will increase the collection efficiency of exotic nuclei, extend the range of nuclei of interest from the A<40 mass region up to the A~80 mass region, and promote the development of radioactive nuclear beam experiments at the RIBLL2 separator.
Model extraction increasingly attracts research attentions as keeping commercial AI models private can retain a competitive advantage. In some scenarios, AI models are trained proprietarily, where neither pre-trained models nor sufficient in-distribu
tion data is publicly available. Model extraction attacks against these models are typically more devastating. Therefore, in this paper, we empirically investigate the behaviors of model extraction under such scenarios. We find the effectiveness of existing techniques significantly affected by the absence of pre-trained models. In addition, the impacts of the attackers hyperparameters, e.g. model architecture and optimizer, as well as the utilities of information retrieved from queries, are counterintuitive. We provide some insights on explaining the possible causes of these phenomena. With these observations, we formulate model extraction attacks into an adaptive framework that captures these factors with deep reinforcement learning. Experiments show that the proposed framework can be used to improve existing techniques, and show that model extraction is still possible in such strict scenarios. Our research can help system designers to construct better defense strategies based on their scenarios.
Gradient-based attribution methods can aid in the understanding of convolutional neural networks (CNNs). However, the redundancy of attribution features and the gradient saturation problem, which weaken the ability to identify significant features an
d cause an explanation focus shift, are challenges that attribution methods still face. In this work, we propose: 1) an essential characteristic, Strong Relevance, when selecting attribution features; 2) a new concept, feature map importance (FMI), to refine the contribution of each feature map, which is faithful to the CNN model; and 3) a novel attribution method via FMI, termed A-FMI, to address the gradient saturation problem, which couples the target image with a reference image, and assigns the FMI to the difference-from-reference at the granularity of feature map. Through visual inspections and qualitative evaluations on the ImageNet dataset, we show the compelling advantages of A-FMI on its faithfulness, insensitivity to the choice of reference, class discriminability, and superior explanation performance compared with popular attribution methods across varying CNN architectures.
Multi-access edge computing (MEC) and non-orthogonal multiple access (NOMA) have been regarded as promising technologies to improve computation capability and offloading efficiency of the mobile devices in the sixth generation (6G) mobile system. Thi
s paper mainly focuses on the hybrid NOMA-MEC system, where multiple users are first grouped into pairs, and users in each pair offload their tasks simultaneously by NOMA, and then a dedicated time duration is scheduled to the more delay-tolerable user for uploading the remaining data by orthogonal multiple access (OMA). For the conventional NOMA uplink transmission, successive interference cancellation (SIC) is applied to decode the superposed signals successively according to the channel state information (CSI) or the quality of service (QoS) requirement. In this work, we integrate the hybrid SIC scheme which dynamically adapts the SIC decoding order among all NOMA groups. To solve the user grouping problem, a deep reinforcement learning (DRL) based algorithm is proposed to obtain a close-to-optimal user grouping policy. Moreover, we optimally minimize the offloading energy consumption by obtaining the closed-form solution to the resource allocation problem. Simulation results show that the proposed algorithm converges fast, and the NOMA-MEC scheme outperforms the existing orthogonal multiple access (OMA) scheme.
In cells, cytoskeletal filament networks are responsible for cell movement, growth, and division. Filaments in the cytoskeleton are driven and organized by crosslinking molecular motors. In reconstituted cytoskeletal systems, motor activity is respon
sible for far-from-equilibrium phenomena such as active stress, self-organized flow, and spontaneous nematic defect generation. How microscopic interactions between motors and filaments lead to larger-scale dynamics remains incompletely understood. To build from motor-filament interactions to predict bulk behavior of cytoskeletal systems, more computationally efficient techniques for modeling motor-filament interactions are needed. Here we derive a coarse-graining hierarchy of explicit and continuum models for crosslinking motors that bind to and walk on filament pairs. We compare the steady-state motor distribution and motor-induced filament motion for the different models and analyze their computational cost. All three models agree well in the limit of fast motor binding kinetics. Evolving a truncated moment expansion of motor density speeds the computation by $10^3$--$10^6$ compared to the explicit or continuous-density simulations, suggesting an approach for more efficient simulation of large networks. These tools facilitate further study of motor-filament networks on micrometer to millimeter length scales.
This paper aims to provide a thorough study on the effectiveness of the transformation-based ensemble defence for image classification and its reasons. It has been empirically shown that they can enhance the robustness against evasion attacks, while
there is little analysis on the reasons. In particular, it is not clear whether the robustness improvement is a result of transformation or ensemble. In this paper, we design two adaptive attacks to better evaluate the transformation-based ensemble defence. We conduct experiments to show that 1) the transferability of adversarial examples exists among the models trained on data records after different reversible transformations; 2) the robustness gained through transformation-based ensemble is limited; 3) this limited robustness is mainly from the irreversible transformations rather than the ensemble of a number of models; and 4) blindly increasing the number of sub-models in a transformation-based ensemble does not bring extra robustness gain.
Massive multiple-input multiple-output (MIMO) and non-orthogonal multiple access (NOMA) are two key techniques for enabling massive connectivity in future wireless networks. A massive MIMO-NOMA system can deliver remarkable spectral improvements and
low communication latency. Nevertheless, the uncontrollable stochastic behavior of the wireless channels can still degrade its performance. In this context, intelligent reflecting surface (IRS) has arisen as a promising technology for smartly overcoming the harmful effects of the wireless environment. The disruptive IRS concept of controlling the propagation channels via software can provide attractive performance gains to the communication networks, including higher data rates, improved user fairness, and, possibly, higher energy efficiency. In this article, in contrast to the existing literature, we demonstrate the main roles of IRSs in MIMO-NOMA systems. Specifically, we identify and perform a comprehensive discussion of the main performance gains that can be achieved in IRS-assisted massive MIMO-NOMA (IRS-NOMA) networks. We outline exciting futuristic use case scenarios for IRS-NOMA and expose the main related challenges and future research directions. Furthermore, throughout the article, we support our in-depth discussions with representative numerical results.
The combination of non-orthogonal multiple access (NOMA) and mobile edge computing (MEC) can significantly improve the spectrum efficiency beyond the fifth-generation network. In this paper, we mainly focus on energy-efficient resource allocation for
a multi-user, multi-BS NOMA assisted MEC network with imperfect channel state information (CSI), in which each user can upload its tasks to multiple base stations (BSs) for remote executions. To minimize the energy consumption, we consider jointly optimizing the task assignment, power allocation and user association. As the main contribution, with imperfect CSI, the optimal closed-form expressions of task assignment and power allocation are analytically derived for the two-BS case. Specifically, the original formulated problem is nonconvex. We first transform the probabilistic problem into a non-probabilistic one. Subsequently, a bilevel programming method is proposed to derive the optimal solution. In addition, by incorporating the matching algorithm with the optimal task and power allocation, we propose a low complexity algorithm to efficiently optimize user association for the multi-user and multi-BS case. Simulations demonstrate that the proposed algorithm can yield much better performance than the conventional OMA scheme but also the identical results with lower complexity from the exhaustive search with the small number of BSs.
The combination of non-orthogonal multiple access (NOMA) and intelligent reflecting surface (IRS) is an efficient solution to significantly enhance the energy efficiency of the wireless communication system. In this paper, we focus on a downlink mult
i-cluster NOMA network, where each cluster is supported by one IRS. We aim to minimize the transmit power by jointly optimizing the beamforming, the power allocation and the phase shift of each IRS. The formulated problem is non-convex and challenging to solve due to the coupled variables, i.e., the beamforming vector, the power allocation coefficient and the phase shift matrix. To address this non-convex problem, we propose an alternating optimization based algorithm. Specifically, we divide the primal problem into the two subproblems for beamforming optimization and phase shifting feasiblity, where the two subproblems are solved iteratively. Moreover, to guarantee the feasibility of the beamforming optimization problem, an iterative algorithm is proposed to search the feasible initial points. To reduce the complexity, we also propose a simplified algorithm based on partial exhaustive search for this system model. Simulation results demonstrate that the proposed alternating algorithm can yield a better performance gain than the partial exhaustive search algorithm, OMA-IRS, and NOMA with random IRS phase shift.
Combining intelligent reflecting surface (IRS) and non-orthogonal multiple access (NOMA) is an effective solution to enhance communication coverage and energy efficiency. In this paper, we focus on an IRS-assisted NOMA network and propose an energy-e
fficient algorithm to yield a good tradeoff between the sum-rate maximization and total power consumption minimization. We aim to maximize the system energy efficiency by jointly optimizing the transmit beamforming at the BS and the reflecting beamforming at the IRS. Specifically, the transmit beamforming and the phases of the low-cost passive elements on the IRS are alternatively optimized until the convergence. Simulation results demonstrate that the proposed algorithm in IRS-NOMA can yield superior performance compared with the conventional OMA-IRS and NOMA with a random phase IRS.
