Search

This paper presents a cycle by cycle Frequency Shift Keying (FSK) demodulator, able to demodulate a FSK signal with 1% frequency modulation index (MI), in a single cycle. Based on the proposed demodulation scheme, a high rate data transmission link can be established through a high-Q inductive coupling link, breaking the basic tradeoff between the power transfer efficiency (PTE) and data rate in single carrier wireless power and data transfer systems. Designed and simulated with $0.18\mu m$ CMOS process, the proposed FSK demodulator, detects successfully a 5Mbps data with a carrier frequency of 5MHz. A test chip is fabricated in 180nm CMOS technology. Measurement results shows that the demodulator is able to demodulate the cycle-by-cycle modulated FSK signals with low modulation index of 2.5%, with a BER of $1.24\times 10^{-5}$. [ABSTRACT FROM AUTHOR]

Soft-input soft-output (SISO) detector for the iterative detection and decoding of coded spatial modulation (SM) signals needs to compute the extrinsic bit log-likelihood ratio (LLR) associated with each detected bit. The maximum logarithmic maximum a posteriori (max-log-MAP) detector is known to perform closely to the optimal logarithmic maximum a posteriori (log-MAP) detector with much lower computational complexity. In this paper, we first apply several mathematical tricks to derive a new and low-complexity algorithm for the max-log-MAP detector. The proposed algorithm features a new tree pruning technique to visit few lattice points and a smart list administration technique to retain a max-log-MAP detector. By following the proposed algorithm, we then design a hardware architecture for SISO detection of coded SM signals over the scenario of 64-QAM symbols, 8 transmit antennas, and 4 receive antennas. Under the TSMC 40 nm CMOS technology, the VLSI implementation results reveal that our architecture requires 46.4K gates and leads to detection throughput 450 Mbps, while operating at clock frequency 400 MHz and consuming power 46 mW. Being the first hardware architecture for SISO detection of coded SM signals, our proposed low-cost max-log-MAP detector is very attractive. [ABSTRACT FROM AUTHOR]

In this paper, a low complexity moving average nested generalized memory polynomial model (MAN-GMP) is proposed for digital predistortion (DPD) of broadband power amplifiers (PAs). As the signal bandwidth increases drastically, the strong nonlinear distortions, especially those induced by the memory effect, are generated from the highly efficient PAs. To compensate for the strong memory effect, a moving average nested envelope memory polynomial (MAN-EMP) model is derived from an accuracy-enhanced GMP model, which offers reduced complexity while suffering from degraded modeling accuracy. The MAN-GMP model is further proposed to improve the modeling accuracy by connecting several memory branches of the MAN-EMP model in parallel. An iterative algorithm is designed to extract the model coefficients efficiently through only one or two iterations. Experimental measurements are carried out on two sub-7 GHz broadband GaN Doherty PAs with up to 200 MHz bandwidth OFDM signals to benchmark the proposed MAN-GMP model against the GMP, the parallel-LUT-MP-EMP (PLUME), the augmented complexity-reduced GMP (ACR-GMP), the generalized twin-nonlinear two-box (GTNTB), and the enhanced Wiener models. The experimental results show that the MAN-GMP model can effectively compensate for the nonlinear distortion of broadband PAs with significant complexity reduction. [ABSTRACT FROM AUTHOR]

Sparse recovery algorithms are integral to compressed sensing (CS) as they facilitate the reconstruction of higher dimensional signals from sub-Nyquist measurements. Although Orthogonal Matching Pursuit (OMP) has been ubiquitously adopted in hardware implementations to curb the computational complexity of CS recovery, increasing sparsity levels incur higher reconstruction costs. Sparsity independent regularized pursuit (SIRP) is a recent algorithm that employs parallel index selection and regularization to curtail the number of iterations required to reconstruct the signal and thereby enhance the reconstruction speed. This paper proposes a novel reformulation of the SIRP algorithm from the hardware perspective, incorporating a cheaper regularization strategy and a modified Gram-Schmidt (MGS) based incremental QR decomposition (QRD) approach. The proposed design incorporates an iterative QRD architecture with feedback circuitry to exploit the parallelism of the triangularization step in MGS. Additionally, a fast inverse square block circumvents the need for parallel divider blocks giving considerable hardware and latency savings. The design reuses the iterative QRD block to implement the interdependent computations of the algorithm by sophisticated scheduling techniques. The proposed implementation on a Xilinx Virtex Ultrascale FPGA can recover 1024-dimensional signals from 25% measurements within $77~\mu \text{s}$ , which translates to a 37% reduction in processing cycles from state-of-art. [ABSTRACT FROM AUTHOR]

The stability criteria based on the Rollett parameter ($K$) plus one auxiliary condition is yet the most popular method used by designers to evaluate the stability of a two-port network although other criteria are available. In this paper, novel analytical proofs of the criteria utilizing $K$ and a new geometrical proof of the relationship between the stability of the ports of a circuit are proposed. They are obtained by making use of relationships between the reflection coefficients at the ports of the network and novel expressions that bind stability circles and unit mapping circles. The developed relations make the proposed proofs easier and more immediate than the traditional ones. Moreover, a new auxiliary condition having the same complexity and a greater geometrical insight with respect to the traditional ones is used. [ABSTRACT FROM AUTHOR]

In this paper, we presented a method of minimizing the number of overlapped partial products in the accumulation of four-term Karatsuba multiplication. This method reduced the summation of 13 overlapped partial products to 9 and the summation of 40 overlapped partial products to 24 in the case of four-term Karatsuba multiplication and eight-term Karatsuba multiplication, respectively. Moreover, to deal with the problem of negative multiplicands introduced by the choice in the evaluation of four-term Karatsuba multiplication, we further presented a method of converting negative multiplicands to non-negative multiplicands. In this way, our method would lead to the reduction of summation complexity and not using signed multipliers. Compared to the original Karatsuba multiplication designs, our proposed Karatsuba-like multiplication with optimal interpolation would lead to up to 7.7% reduction in ADP (area-delay-product) in ASIC implementations, and up to 20.2% reduction in ADP compared with built-in designs of Synopsys Design Compiler. [ABSTRACT FROM AUTHOR]

Approximate computing is an efficient approach for reducing computational costs. This method involves a trade-off between computational accuracy and the circuit’s power consumption, delay, and area. However, the accuracy requirements may differ for different applications. In some situations, precise results are required. Therefore, this paper proposes an accuracy-configurable radix-4 adder (ACRA), which uses the power gating technique to turn on or turn off the partial logic gates of an adder element dynamically to compute accurate or approximate results. When the ACRA operates in the approximate mode, the partial sum of one adder element is modified to reduce the error distance between the approximate and accurate results. A comparison of the ACRA with two state-of-the-art accuracy-configurable adders indicated that the ACRA achieved the best trade-off between the power–delay product and computational accuracy. In addition, in the image processing experiment conducted in this study, under half-approximate computation, the peak signal-to-noise ratio of the ACRA was less than 1 dB and its structural similarity index measure was maintained above 0.99. These results are superior to those obtained for other accuracy-configurable adders. [ABSTRACT FROM AUTHOR]

A small correction is made in the example of Section V (A representative application to networked systems) of the paper with the above title. [ABSTRACT FROM AUTHOR]

Quantization is a prominent approach to compress model sizes of deep neural networks (DNNs), which clusters high-precision weights into a smaller set of quantization levels and represents high-precision weights by low-precision indexes. To achieve the same accuracy, nonuniform quantized DNNs (NUQ-DNNs) with unequal quantization intervals need lower index precision than uniform quantized DNNs (UQ-DNNs) with equal intervals, achieving smaller model sizes. Hence, deploying NUQ-DNNs on accelerators costs less on- and off-chip memory accesses than UQ-DNNs, which are more valuable for edge devices. However, accelerating NUQ-DNNs is nontrivial, since weight indexes cannot be directly used for computations. Previous NUQ-DNN accelerators adopt standard convolutions by decoding weight indexes into actual-weights multiplied with activations, causing abundant look-up overhead and redundant computations. In this work, we propose a weight-repetition-aware activation aggregating (WPAA) convolution approach to accelerate inference of variable-precision NUQ- and UQ-DNNs. By merging convolutions of multiple kernels, WPAA requires no look-up operation and removes redundant computations. Based on WPAA, we design a generic quantized DNN accelerator (GQNA). Furthermore, we propose a layer-adaptive kernel-reordering merging scheme to off-line adjust merging order of kernels for minimizing energy consumption of GQNA. Implemented under TSMC 28-nm technology, GQNA achieves 31.9 and 32.6 TOPS/W energy efficiency for 1-b UQ- and NUQ-VGG-16, respectively. [ABSTRACT FROM AUTHOR]

Learning to perform perfect tracking tasks based on measurement data is desirable in the controller design of systems operating repetitively. This motivates this article to seek an optimization-based design and analysis approach for data-driven learning control systems by focusing on iterative learning control (ILC) of repetitive systems with unknown nonlinear time-varying dynamics. It is shown that perfect output tracking can be realized with updating inputs, where no explicit model knowledge but only measured input–output data are leveraged. In particular, adaptive updating strategies are proposed to obtain parameter estimations of nonlinearities. A double-dynamics analysis approach is applied to establish ILC convergence, together with boundedness of input, output, and estimated parameters, which benefits from employing properties of nonnegative matrices. Simulations are implemented to verify the validity of our optimization-based adaptive ILC. [ABSTRACT FROM AUTHOR]

Indivisible resource allocation motivates us to study the matrix completion concerning the class of (0,1)-matrices with prescribed row/column sums and preassigned zeros. We illustrate and generalize the (0,1)-matrix completion in the following two scenarios: a demand-response application involving nonnegative integer matrices with different bounds across rows and an edge caching matching problem allowing row and column sums to vary within separately designated bounds. The applications require analytic characterizations of the supply adequacy and cause large-scale matrix completion instances. Remarkably, we derive a structure tensor and use its nonnegativity to establish a necessary and sufficient condition under which the considered matrix class is nonempty. The tensor condition can characterize the adequacy of a supply for a prescribed demand and facilitate identifying the minimum supplement to the supply so that the augmented supply becomes adequate when the adequacy gap is nonzero. Notably, we design a tensor-based combinatorial algorithm to construct a required matrix, representing a feasible resource allocation. Numerical simulations justify the efficiency of our approach. [ABSTRACT FROM AUTHOR]

We present order reduction results for linear time invariant descriptor systems. Results are given for both forced and unforced systems as well methods for constructing the reduced order systems. Our results establish a precise connection between classical and new results on this topic, and lead to an elementary construction of quasi-Weierstrass forms for a descriptor system. Examples are given to illustrate the usefulness of our results. [ABSTRACT FROM AUTHOR]

Differential-algebraic equations (DAEs) are a widespread dynamical model that describes continuously evolving quantities defined with differential equations, subject to constraints expressed through algebraic relationships. As such, DAEs arise in many fields ranging from physics, chemistry, and engineering. In this article, we focus on linear DAEs, and develop a theory for their minimization up to an equivalence relation. We present differential equivalence, which relates DAE variables that have equal solutions at all time points (thus requiring them to start with equal initial conditions) and extends the line of research on bisimulations developed for Markov chains and ordinary differential equations. We apply our results to the electrical engineering domain, showing that differential equivalence can explain invariances in certain networks as well as analyze DAEs, which could not be originally treated due to their size. [ABSTRACT FROM AUTHOR]

We consider a scenario where a system experiences a disruption, and the states (representing health values) of its components continue to reduce over time, unless they are acted upon by a controller. Given this dynamical setting, we consider the problem of finding an optimal control (or switching) sequence to maximize the sum of the weights of the components, whose states are brought back to the maximum value. We first provide several characteristics of the optimal policy for the general (fully heterogeneous) version of this problem. We then show that under certain conditions on the rates of repair and deterioration, we can explicitly characterize the optimal control policy as a function of the states. When the deterioration rate (when not being repaired) is larger than or equal to the repair rate, and the deterioration and repair rates as well as the weights are homogeneous across all the components, we show that the policy that targets the component with the largest state value at each time step is optimal under certain conditions. On the other hand, if the repair rates are sufficiently larger than the deterioration rates, the optimal control policy is to target the component whose state minus the deterioration rate is least in a particular subset of components at each time step. [ABSTRACT FROM AUTHOR]

The Lyapunov equation for polytopic linear inclusions over the space of Minkowski functions of nonempty compact and convex sets that contain the origin as an interior point is studied. In particular, necessary and sufficient conditions for the characterization, existence, and uniqueness of its fundamental solution are derived. [ABSTRACT FROM AUTHOR]

This article investigates quantitative supervisory control with local mean payoff objectives on discrete event systems modeled as weighted automata. Weight flows are generated as new events occur, which are required to satisfy some quantitative conditions. We focus on mean weights (payoffs) over a finite number of events, which serve as a measure for the stability or robustness of weight flows. The range of events to evaluate the mean payoff is termed a window, which slides as new events occur. Qualitative requirements such as safety and liveness are also necessary along with quantitative requirements. Supervisory control is employed to manipulate the operation of the system so that the requirements are satisfied. We consider two different scenarios based on whether the window size is fixed or not. Correspondingly, we formulate two supervisory control problems, both of which are solved sequentially by first tackling the qualitative issues and then the quantitative ones. The automaton model is then transformed to a two-player game between the supervisor and the environment, where safety and liveness are enforced. Based on the intermediate results, several quantitative objectives are defined to formulate two games, which correspond to the two proposed supervisory control problems. Finally, we synthesize provably correct supervisors by solving the games and completely resolve both problems. [ABSTRACT FROM AUTHOR]

This article considers the design and analysis of a distributed sensor fault accommodation scheme for a class of nonlinear systems, modeled as interconnected subsystems. Each subsystem is subject to possible local sensor faults or may be impacted by a sensor fault in a neighboring subsystem due to the exchange of information between neighboring controllers. The proposed adaptive approximation-based distributed fault accommodation scheme is triggered in the event that a sensor fault is detected. The stability of the closed-loop system in the presence of sensor faults is rigorously analyzed both for the period prior to the detection of the fault, as well as after the detection of the fault when the fault accommodation scheme is activated. The effectiveness of the proposed scheme is illustrated by a simulation example. [ABSTRACT FROM AUTHOR]

Continuous-time linear systems with periodic coefficients of period $\omega$ are considered in this article. We study the question of the controllability interval length of periodic systems. It is known that a controllable $\omega$ -periodic system is controllable over intervals of the length $k\omega$ , where $k$ is the degree of the minimal polynomial of the monodromy matrix. It is proved that this boundary $k\omega$ is exact and cannot be decreased. It is proved that controllability of an $\omega$ -periodic system on some interval $[\tau,\tau +s\omega ]$ depends on the initial moment $\tau$ , in general. [ABSTRACT FROM AUTHOR]

We study the deadline scheduling problem for multiple deferrable jobs that arrive in a random manner and are to be processed before individual deadlines. The processing of the jobs is subject to a time-varying limit on the total processing rate at each stage. We formulate the scheduling problem as a restless multiarmed bandit problem. Relaxing the scheduling problem into multiple independent single-arm scheduling problems, we define the Lagrangian priority value as the greatest tax under which it is optimal to activate the arm. We propose a Lagrangian priority policy that processes jobs in the order of their Lagrangian priority values, and establish its asymptotic optimality as the system scales. Numerical results show that the proposed Lagrangian priority policy achieves 22%–49% higher average reward than the classical Whittle index policy (that does not take into account the processing rate limits). [ABSTRACT FROM AUTHOR]

The $\mathcal {H}_2$ norm of an exponentially stable system described by delay differential algebraic equations (DDAEs) might be infinite due to the existence of hidden feedthrough terms and, as shown in this article, it might become infinite as a result of infinitesimal changes to the delay parameters. We introduce the notion of strong $\mathcal {H}_2$ norm of semi-explicit DDAEs, a robustified measure that considers delay perturbations, and we analyze its properties. We derive necessary and sufficient finiteness criteria for the strong $\mathcal {H}_2$ norm in terms of a finite number of equalities involving multidimensional powers of a finite set of matrices. As the main contribution, we present a strengthened, sufficient, condition for finiteness of the strong $\mathcal {H}_2$ norm, along with an algorithm for checking it, which has significantly better scalability properties in terms of both the system dimension and the number of delays. We show that the satisfaction of the novel condition is equivalent to the existence of a simultaneous block triangularization of the matrices of the delay difference equation associated with the DDAE. The latter is instrumental to a novel regularization procedure that allows us to transform the DDAE to a neutral-type system with the same transfer matrix, without any need for differentiation of inputs or outputs. This transformation enables computing of the strong $\mathcal {H}_2$ norm using Lyapunov matrices. Finally, we show by a counterexample that the strengthened condition is, in general, not necessary, inducing open problems, but we also list several classes of DDAEs for which it is necessary and sufficient. [ABSTRACT FROM AUTHOR]

The issue of finite-time $H_{\infty }$ state estimation is studied for a class of discrete-time nonlinear two-time-scale complex networks (TTSCNs) whose measurement outputs are transmitted to a remote estimator via a bandwidth-limited communication network under the stochastic communication protocol (SCP). To reflect different time scales of state evolutions, a new discrete-time TTSCN model is devised by introducing a singular perturbation parameter (SPP). For the sake of avoiding/alleviating the undesirable data collisions, the SCP is adopted to schedule the data transmissions, where the transition probabilities involved are assumed to be partially unknown. By constructing a new Lyapunov function dependent on the information of the SCP and SPP, a sufficient condition is derived which ensures that the resulting error dynamics is stochastically finite-time bounded and satisfies a prescribed $H_{\infty }$ performance index. By resorting to the solutions of several matrix inequalities, the gain matrices of the state estimator are given and the admissible upper bound of the SPP can be evaluated simultaneously. The performance of the designed state estimator is demonstrated by two examples. [ABSTRACT FROM AUTHOR]

With fast developments in communication technologies, a large number of practical systems adopt the networked control structure. For this structure, the fading problem is an emerging issue among other network problems. It has not been extensively investigated how to guarantee superior control performance in the presence of unknown fading channels. This article presents a learning strategy for gradually improving the tracking performance. To this end, an iterative estimation mechanism is first introduced to provide necessary statistical information such that the biased signals after transmission can be corrected before being utilized. Then, learning control algorithms incorporating with a decreasing step-size sequence are designed for both output and input fading cases. The convergence in both mean-square and almost-sure senses of the proposed schemes is strictly proved under mild conditions. Illustrative simulations verify the effectiveness of the entire learning framework. [ABSTRACT FROM AUTHOR]

Equilibrium-independent passive-short (EIPS) systems are a class of systems that satisfy a passivity-like dissipation inequality with respect to any forced equilibria with nonpositive passivity indices. This article presents a geometric approach for finding a passivizing transformation for such systems, relying on their steady-state input–output relation and the notion of projective quadratic inequalities (PQIs). We show that PQIs arise naturally from passivity-shortage characteristics of an EIPS system, and the set of their solutions can be explicitly expressed. We leverage this connection to build an input–output mapping that transforms the steady-state input–output relation to a monotone relation, and show that the same mapping passivizes the EIPS system. We show that the proposed transformation can be implemented through a combination of feedback, feedthrough, post- and premultiplication gains. Furthermore, we consider an application of the presented passivation scheme for the analysis of networks comprised of EIPS systems. Numerous examples are provided to illustrate the theoretical findings. [ABSTRACT FROM AUTHOR]

We consider a game-theoretic model where individuals compete over a shared failure-prone system or resource. We investigate the effectiveness of a taxation mechanism in controlling the utilization of the resource at the Nash equilibrium when the decision-makers have behavioral risk preferences, captured by prospect theory. We first observe that heterogeneous prospect-theoretic risk preferences can lead to counter-intuitive outcomes. In particular, for resources that exhibit network effects, utilization can increase under taxation and there may not exist a tax rate that achieves the socially optimal level of utilization. We identify conditions under which utilization is monotone and continuous, and then characterize the range of utilizations that can be achieved by a suitable choice of tax rate. We further show that resource utilization is higher when players are charged differentiated tax rates compared to the case when all players are charged an identical tax rate, under suitable assumptions. [ABSTRACT FROM AUTHOR]

This article presents a small-gain theorem for networks composed of a countably infinite number of finite-dimensional subsystems. Assuming that each subsystem is exponentially input-to-state stable, we show that if the gain operator, collecting all the information about the internal Lyapunov gains, has a spectral radius less than one, the overall infinite network is exponentially input-to-state stable. The effectiveness of our result is illustrated through several examples including nonlinear spatially invariant systems with sector nonlinearities and a road traffic network. [ABSTRACT FROM AUTHOR]

We present mean-nonovershooting tracking control designs for stochastic strict-feedback nonlinear systems. Compared with the existing stochastic nonlinear tracking designs, the advantage of our design is that arbitrary reference trajectories can be tracked “from below.” We first consider a special class of stochastic nonlinear systems where the diffusion terms satisfy the linear growth condition and the drift terms are in special triangular forms. A new controller is designed to guarantee that the mean of the system output can asymptotically track a given trajectory without overshooting while keeping all of the states mean-square bounded. In this case, an appealing feature in our design is that strict mean-nonovershooting tracking is achieved while the existing results on nonlinear systems with deterministic disturbance can only achieve practical nonovershooting tracking. Then, we design a new controller for general stochastic strict-feedback nonlinear systems to ensure that the mean of overshoot can be tuned to be arbitrarily small while all the states of the closed-loop system remain bounded in probability. In this case, the new design ingredient is that our control gains are independent of the initial conditions of system states, and the initial conditions of derivatives of the reference trajectory. This is in contrast with the existing results on corresponding deterministic systems whose control gains heavily rely on these initial conditions. Finally, two simulation examples are given to illustrate the stochastic nonlinear nonovershooting tracking designs. [ABSTRACT FROM AUTHOR]

This article presents a collaborative neurodynamic optimization (CNO) approach to multivehicle task assignments (TAs). The original combinatorial quadratic optimization problem for TA is reformulated as a quadratic unconstrained binary optimization (QUBO) problem with a quadratic utility function and a penalty function for handling load capacity and cooperation constraints. In the framework of CNO with a population of discrete Hopfield networks (DHNs), a TA algorithm is proposed for solving the formulated QUBO problem. Superior experimental results in four typical multivehicle operation scenarios are reported to substantiate the efficacy of the proposed neurodynamics-based TA approach. [ABSTRACT FROM AUTHOR]

This article studies the controllability degree via analyzing the condition number of Gramian matrix. Our aim is to explore how the network characteristics affect the controllability degree. Specifically, we prove that a large time parameter would worsen the controllability degree. The time parameter could be understood as the network coupling strength. For directed path networks, we derive how edge weights and time parameter jointly determine the best controllability degree. Furthermore, we prove that either adding a new edge or enhancing an existing edge weight appropriately would worsen the controllability degree. Moreover, through the numerical simulation of external inputs deployment, we find a significant statistical relationship between the controllability index and the controllability degree. In this article, the Gramian matrix reveals the importance of network characteristics that cannot be captured by classic Kalman rank condition or Popov–Belevitch–Hautus (PBH) test. [ABSTRACT FROM AUTHOR]

With the rapid development of sensor technologies, multisensor signals are now readily available for health condition monitoring and remaining useful life (RUL) prediction. To fully utilize these signals for a better health condition assessment and RUL prediction, health indices are often constructed through various data fusion techniques. Nevertheless, most of the existing methods fuse signals linearly, which may not be sufficient to characterize the health status for RUL prediction. To address this issue and improve the predictability, this article proposes a novel nonlinear data fusion approach, namely, a shape-constrained neural data fusion network for health index construction. Especially, a neural network-based structure is employed, and a novel loss function is formulated by simultaneously considering the monotonicity and curvature of the constructed health index and its variability at the failure time. A tailored adaptive moment estimation algorithm (Adam) is proposed for model parameter estimation. The effectiveness of the proposed method is demonstrated and compared through a case study using the Commercial Modular Aero-Propulsion System Simulation (C-MAPSS) data set. [ABSTRACT FROM AUTHOR]

A solution is presented to a long-standing open problem of control theory, the decoupling of linear systems using static-state feedback while assuring stability. Both diagonal and block decoupling is considered. The earliest known investigation of system decoupling dates back to 1934 and all past stable solutions were obtained under restrictive assumptions on system, block structure, or on a particular form of the control law. Most often, the results were limited to regular state feedback. The present formulation avoids restrictive hypotheses. The system is stabilized first and then decoupled while preserving stability. The existence of decoupling feedback is conditioned by system invariants with respect to the group of stability-preserving decoupling transformations. Since nonregular state feedback is authorized, some well-known results of regular decoupling are not retained. In particular, the presence of interconnection zeros does not prevent a stable decoupling to be achievable. [ABSTRACT FROM AUTHOR]

This article applies a singular perturbation theory to solve an optimal linear quadratic tracker problem for a continuous-time two-time-scale process. Previously, singular perturbation was applied for system regulation. It is shown that the two-time-scale tracking problem can be separated into a linear–quadratic tracker (LQT) problem for the slow system and a linear–quadratic regulator (LQR) problem for the fast system. We prove that the solutions to these two reduced-order control problems can approximate the LQT solution of the original control problem. The reduced-order slow LQT and fast LQR control problems are solved by off-policy integral reinforcement learning (IRL) using only measured data from the system. To test the effectiveness of the proposed method, we use an industrial thickening process as a simulation example and compare our method to a method with the known system model and a method without time-scale separation. [ABSTRACT FROM AUTHOR]

In deep face recognition, the commonly used softmax loss and its newly proposed variations are not yet sufficiently effective to handle the class imbalance and softmax saturation issues during the training process while extracting discriminative features. In this brief, to address both issues, we propose a class-variant margin (CVM) normalized softmax loss, by introducing a true-class margin and a false-class margin into the cosine space of the angle between the feature vector and the class-weight vector. The true-class margin alleviates the class imbalance problem, and the false-class margin postpones the early individual saturation of softmax. With negligible computational complexity increment during training, the new loss function is easy to implement in the common deep learning frameworks. Comprehensive experiments on the LFW, YTF, and MegaFace protocols demonstrate the effectiveness of the proposed CVM loss function. [ABSTRACT FROM AUTHOR]

We face a binary multiple instance learning (MIL) problem, whose objective is to discriminate between two kinds of point sets: positive and negative. In the MIL terminology, such sets are called bags, and the points inside each bag are called instances. Considering the case with two classes of instances (positive and negative) and inspired by a well-established instance-space support vector machine (SVM) model, we propose to extend to MIL classification the proximal SVM (PSVM) technique that has revealed very effective for supervised learning, especially in terms of computational time. In particular, our approach is based on a new instance-space model that exploits the benefits coming from both SVM (better accuracy) and PSVM (computational efficiency) paradigms. Starting from the standard MIL assumption, such a model is aimed at generating a hyperplane placed in the middle between two parallel hyperplanes: the first one is a proximal hyperplane that clusters the instances of the positive bags, while the second one constitutes a supporting hyperplane for the instances of the negative bags. Numerical results are presented on a set of MIL test data sets drawn from the literature. [ABSTRACT FROM AUTHOR]