Large language models (LLMs) have demonstrated impressive results on natural tasks, and security researchers are beginning to employ them in both offensive defensive systems. In cyber-security, there been multiple research efforts that utilize LLMs focusing the pre-breach stage of attacks like phishing malware generation. However, so far lacks a comprehensive study regarding whether LLM-based systems can be leveraged simulate post-breach typically human-operated, or "hands-on-keyboard"...

Quantifying the importance of each training point to a learning task is fundamental problem in machine and estimated scores have been leveraged guide range data workflows such as summarization domain adaption. One simple idea use leave-one-out error indicate its importance. Recent work has also proposed Shapley value, it defines unique value distribution scheme that satisfies set appealing properties. However, calculating values often expensive, which limits applicability real-world...

Pulse Doppler (PD) systems are widely used for moving target detection, especially in scenarios with clutter. Range ambiguity, which arises from fixed parameters waveforms, is an inherent drawback conventional systems. By using a diverse pulse train such as of coherent phase coded pulses, these ambiguous peaks can be suppressed effectively but at the cost sidelobe dispersions. In this work, novel efficient PD process proposed to suppress range ambiguity and detect targets under strong...

This paper presents a novel low complexity time-of arrival (TOA) estimation method for unknown intra-pulse modulated signal under SNR condition. The received is firstly pre-processed by delayed conjugate multiplication to reduce the modulation and get rough range of signal. Then correlation accumulation applied achieve precise rising falling edge. Simulations are conducted show feasibility evaluate performance estimation. Compared with existing methods, this can better result.

Emerging applications in Printed Circuit Board (PCB) routing impose new challenges on automatic length matching, including adaptability for any-direction traces with their original preserved interactiveness. The can be addressed through two orthogonal stages: assign non-overlapping regions to each trace and meander the within reach target length. In this paper, mainly focusing meandering stage, we propose an obstacle-aware detailed approach optimize utilization of available space achieve...

Reference placement is promising to handle the increasing complexity in PCB design. We model netlist into a graph and use subgraph matching algorithm find isomorphism of placed template component combination reuse placement. The state-of-the-art VF3 can achieve high accuracy while suffering from computation time large-scale instances. Thus, we propose D2BS guarantee quality efficiency. build filter candidate set (CS) according designed features construct CS structure. In optimization,...

A novel low complexity time-of arrival (TOA) estimation method is presented in this paper. This achieves a precise TOA under SNR condition for unknown intra-pulse modulated signal, especially frequency pulse. Firstly, delayed conjugate multiplication employed to reduce the modulation as well estimate rough range of received Then, part, correlation accumulation used by accumulating pre-processed signal with proper gain. Finally, simulation results are given examine estimating performance and...

Mismatched filter is an important technique to suppress the sidelobes of phase-coded pulses. Here, authors propose a mainlobe windowing procedure, by which much lower sidelobe level can be obtained at slightly signal-to-noise ratio loss. Based on this idea, they present two mismatched design procedures, namely integrated (ISL) optimal and peak (PSL) filter. The ISL formula in closed form PSL optimisation procedures. Simulation results are presented verify theory.

The worst-case training principle that minimizes the maximal adversarial loss, also known as (AT), has shown to be a state-of-the-art approach for enhancing robustness. Nevertheless, min-max optimization beyond purpose of AT not been rigorously explored in context. In this paper, we show how general framework over multiple domains can leveraged advance design different types attacks. particular, given set risk sources, minimizing attack loss reformulated problem by introducing domain weights...

Although the prevention of AI vulnerabilities is critical to preserve safety and privacy users businesses, educational tools for robust are still underdeveloped worldwide. We present design, implementation, assessment Maestro. Maestro an effective open-source game-based platform that contributes advancement education. provides goal-based scenarios where college students exposed challenging life-inspired assignments in a competitive programming environment. assessed Maestro's influence on...

Although the prevention of AI vulnerabilities is critical to preserve safety and privacy users businesses, educational tools for robust are still underdeveloped worldwide. We present design, implementation, assessment Maestro. Maestro an effective open-source game-based platform that contributes advancement education. provides "goal-based scenarios" where college students exposed challenging life-inspired assignments in a "competitive programming" environment. assessed Maestro's influence on...

Recent research efforts on 3D point cloud semantic segmentation (PCSS) have achieved outstanding performance by adopting neural networks. However, the robustness of these complex models not been systematically analyzed. Given that PCSS has applied in many safety-critical applications like autonomous driving, it is important to fill this knowledge gap, especially, how are affected under adversarial samples. As such, we present a comparative study robustness. First, formally define attacker's...

The joint design of mismatched filter and waveform is an important technique in suppressing sidelobes phase coded waveform. Conventionally, implied constraint that the mainlobe response same width as a chip or bit included. In fact, slightly broadened acceptable most cases. More importantly, relaxation will provide more freedom sidelobe suppression. this work, into fixed window shape. Under constraints shape constant modular, optimization problem to minimize integrated level (ISL). solved by...

Pulse edge estimation has always been an important issue in the field of electronic reconnaissance. Here, for unknown received pulse with non-uniform envelope, a novel method based on instantaneous frequency is proposed. By using sliding window detection standard deviation segmented signal, rough can be obtained. Then, double threshold smoothed signal segment performed to obtain accurate estimation. Moreover, parameter setting and application this are analysed. The simulation results show...

Quantifying the importance of each training point to a learning task is fundamental problem in machine and estimated scores have been leveraged guide range data workflows such as summarization domain adaption. One simple idea use leave-one-out error indicate its importance. Recent work has also proposed Shapley value, it defines unique value distribution scheme that satisfies set appealing properties. However, calculating values often expensive, which limits applicability real-world...

Recent research efforts on 3D point cloud semantic segmentation (PCSS) have achieved outstanding performance by adopting neural networks. However, the robustness of these complex models not been systematically analyzed. Given that PCSS has applied in many safety-critical applications like autonomous driving, it is important to fill this knowledge gap, especially, how are affected under adversarial samples. As such, we present a comparative study robustness. First, formally define attacker's...