Voice user interface (VUI) has become increasingly popular in recent years. Speaker recognition system, as one of the most common VUIs, has emerged as an important technique to facilitate security-required applications and services. In this paper, we propose to design, for the first time, a real-time, robust, and adaptive universal adversarial attack against the state-of-the-art deep neural network (DNN) based speaker recognition systems in the white-box scenario. By developing an audio-agnostic universal perturbation, we can make the DNN-based speaker recognition systems to misidentify the speaker as the adversary-desired target label, with using a single perturbation that can be applied on arbitrary enrolled speaker’s voice. In addition, we improve the robustness of our attack by modeling the sound distortions caused by the physical over-the-air propagation through estimating room impulse response (RIR). Moreover, we propose to adaptively adjust the magnitude of perturbations according to each individual utterance via spectral gating. This can further improve the imperceptibility of the adversarial perturbations with minor increase of attack generation time. Experiments on a public dataset of 109 English speakers demonstrate the effectiveness and robustness of the proposed attack. Our attack method achieves average 90% attack success rate on both X-vector and d-vector speaker recognition systems. Meanwhile, our method achieves 100 × speedup on attack launching time, as compared to the conventional non-universal attacks.
Anchoring-based self-assembly (ASA) has emerged as a material-saving and highly scalable strategy to fabricate charge-transporting monolayers for perovskite solar cells (PSCs). However, the interfacial hole-extraction and electron-blocking performances are highly dependent on the compactness of the ASA monolayers, which has been largely ignored though it is very crucial to the efficiency and stability of PSCs. Here, strategically designed hole-transporting molecules with different anchoring groups are incorporated to investigate the effect of bonding strength on monolayer quality and correlate these with the performance of p-i-n structured PSCs. It is unraveled that the anchoring groups with a stronger bonding strength are advantageous for improving the assembly rate, density, and compactness of ASA monolayer, thus enhancing charge collection and suppressing interfacial recombination. The prototypical PSCs based on optimal ASA monolayer achieve a high power conversion efficiency (PCE) of 21.43% (0.09 cm2). More encouragingly, when enlarging the device area by tenfold, a comparable PCE of 20.09% (1.0 cm2) can be obtained, suggesting that the ASA strategy is practically useful for scaling-up. The robust anchoring of the ASA monolayer also enhances devices stability, retaining 90% of initial PCE after three months. This study provides important insights into the ASA charge-transporting monolayers for efficient and stable PSCs. 相似文献
