Adversarial learning is a widely used paradigm in machine learning, often formulated as a min-max optimization problem where the inner maximization imposes adversarial constraints to guide the outer learner toward more robust solutions. This framework underlies methods such as Sharpness-Aware Minimization (SAM) and Generative Adversarial Networks (GANs). However, traditional gradient-based approaches to such problems often face challenges in balancing accuracy and efficiency due to second-order complexities. In this paper, we propose a bilevel optimization framework that reformulates these adversarial learning problems by leveraging the tractability of the lower-level problem. The bilevel framework introduces no additional complexity and enables the use of advanced bilevel tools. We further develop a provably convergent single-loop stochastic algorithm that effectively balances learning accuracy and computational cost. Extensive experiments show that our method improves generation quality in terms of FID and JS scores for GANs, and consistently achieves higher accuracy for SAM under label noise and across various backbones, while promoting flatter loss landscapes. Overall, this work provides a practical and theoretically grounded framework for solving adversarial learning tasks through bilevel optimization.

By integrating memristors into a Hopfield neural network (HNN), a diverse range of dynamical behavior can be generated, which has significant implications for modeling and biomimetic applications of artificial neurons. However, research on the firing dynamics of HNNs remains relatively limited. In response, a memristive tri-neurons Hopfield neural network (MTN-HNN) was constructed, with the synapse of the second neuron replaced by the proposed memristor. A theoretical and experimental investigation of the dynamics of this neural network was conducted using general analytical tools, such as phase diagrams, Lyapunov exponents, bifurcation diagrams, and others. Experimental results indicate that the dynamics of the MTN-HNN is influenced by the internal parameters of the memristor, enabling the network to extend attractors in up to two directions and thereby form grid multi-scrolls. Notably, the MTN-HNN exhibits various firing modes, including periodic and chaotic bursting. Finally, an encryption scheme was proposed to demonstrate the potential of the MTN-HNN, and both the custom digital circuits and the encryption scheme were successfully implemented on a Field-Programmable Gate Array (FPGA).

This paper presents EFTD, a novel floating-point error detection method using EFT and stochastic progressive random sampling for multi-input programs. Compared to the state-of-the-art FPGen, EFTD detects higher errors on 13 out of 21 programs with an average detection time of 5 minutes, significantly outperforming FPGen's 2-hour average.

This work focuses on addressing two major challenges in the context of large-scale nonconvex Bi-Level Optimization (BLO) problems, which are increasingly applied in machine learning due to their ability to model nested structures. These challenges involve ensuring computational efficiency and providing theoretical guarantees. While recent advances in scalable BLO algorithms have primarily relied on lower-level convexity simplification, our work specifically tackles large-scale BLO problems involving nonconvexity in both the upper and lower levels. We simultaneously address computational and theoretical challenges by introducing an innovative single-loop gradient-based algorithm, utilizing the Moreau envelope-based reformulation, and providing non-asymptotic convergence analysis for general nonconvex BLO problems. Notably, our algorithm relies solely on first-order gradient information, enhancing its practicality and efficiency, especially for large-scale BLO learning tasks. We validate our approach’s effectiveness through experiments on various synthetic problems, two typical hyper-parameter learning tasks, and a real-world neural architecture search application, collectively demonstrating its superior performance.

Objective assessment of taste sensation is essential for medical diagnosis, food development, and multisensory interaction. Human taste sensation can be characterized through biosignals such as electroencephalography (EEG) and electromyography (EMG). However, taste sensation recognition on multiple-subject datasets remains challenging due to the low signal-to-noise ratio and substantial individual variability of biosignals. To address these problems, we propose SwinTaste for accurate and generalized taste sensation recognition from bimodal biosignals. The Transformer is introduced to extract hierarchical features. A two-stage patch partition module is optimized for the characteristics of biosignals. Moreover, a multi-task learning strategy is adopted to improve the generalization and subject adaptation abilities. The SwinTaste is evaluated on a multiple-subject taste sensation dataset. Comparison experiments and ablation studies demonstrate the superior performance of SwinTaste, indicating the potential for generalized application in biosignal recognition.

Gradient methods have become mainstream techniques for Bi-Level Optimization (BLO) in learning fields. The validity of existing works heavily rely on either a restrictive Lower- Level Strong Convexity (LLSC) condition or on solving a series of approximation subproblems with high accuracy or both. In this work, by averaging the upper and lower level objectives, we propose a single loop Bi-level Averaged Method of Multipliers (sl-BAMM) for BLO that is simple yet efficient for large-scale BLO and gets rid of the limited LLSC restriction. We further provide non-asymptotic convergence analysis of sl-BAMM towards KKT stationary points, and the comparative advantage of our analysis lies in the absence of strong gradient boundedness assumption, which is always required by others. Thus our theory safely captures a wider variety of applications in deep learning, especially where the upper-level objective is quadratic w.r.t. the lower-level variable. Experimental results demonstrate the superiority of our method.

Currently, some works on studying complete synchronization of dynamical systems are usually restricted to its two special cases: 1) power-rate synchronization and 2) exponential synchronization. Therefore, how to give a generalization of these types of complete synchronization by the mathematical expression is an open question that needs to be urgently solved. To begin with, this article proposes multimode function synchronization by the mathematical expression for the first time, which is a generalization of exponential synchronization, power-rate synchronization, logarithmical synchronization, and so on. Moreover, two adaptive controllers are designed to achieve robust multimode function synchronization of memristive neural networks (MNNs) with mismatched parameters and uncertain parameters. Each adaptive controller includes function$r(t)$and update gain$\sigma $. By choosing different types of$r(t)$, multiple types of complete synchronization, including power-rate synchronization and exponential synchronization can be obtained. And update gain$\sigma $can be used to adjust the speed of synchronization. Therefore, our results enlarge and strengthen the existing results. Two examples are put forward to verify the validity of our results.

It is well known that weighted sum of node states plays an essential role in function implementation of neural networks. Therefore, this paper proposes a new weighted sum synchronization model for memristive neural networks. Unlike the existing synchronization models of memristive neural networks which control each network node to reach synchronization, the proposed model treats the networks as dynamic entireties by weighted sum of node states and makes the entireties instead of each node reach expected synchronization. In this paper, weighted sum complete synchronization and quasi-synchronization are both investigated by designing feedback controller and aperiodically intermittent controller, respectively. Meanwhile, a flexible control scheme is designed for the proposed model by utilizing some switching parameters and can improve anti-interference ability of control system. By applying Lyapunov method and some differential inequalities, some effective criteria are derived to ensure the synchronizations of memristive neural networks. Moreover, the error level of the quasi-synchronization is given. Finally, numerical simulation examples are used to certify the effectiveness of the derived results.