Search

Arguably the most representative application of artificial intelligence, autonomous driving systems usually rely on computer vision techniques to detect the situations of the external environment. Object detection underpins the ability of scene understanding in such systems. However, existing object detection algorithms often behave as a black box, so when a model fails, no information is available on When, Where and How the failure happened. In this paper, we propose a visual analytics approach to help model developers interpret the model failures. The system includes the micro- and macro-interpreting modules to address the interpretability problem of object detection in autonomous driving. The micro-interpreting module extracts and visualizes the features of a convolutional neural network (CNN) algorithm with density maps, while the macro-interpreting module provides spatial-temporal information of an autonomous driving vehicle and its environment. With the situation awareness of the spatial, temporal and neural network information, our system facilitates the understanding of the results of object detection algorithms, and helps the model developers better understand, tune and develop the models. We use real-world autonomous driving data to perform case studies by involving domain experts in computer vision and autonomous driving to evaluate our system. The results from our interviews with them show the effectiveness of our approach.

A strategy for the fast generation of hydroxyl radicals (HO·) via photo-electro-reduction of oxygen by rerouting the electron transfer pathway was proposed. The rate-determining step of HO· production is the formation of H2O2 and the simultaneous reduction of H2O2. Engineering of F-TiO2 with single atom Pd bonded with four F and two O atoms favored the electrocatalytic 2-electron oxygen reduction to H2O2 with as high as 99% selectivity, while the additional channel bond HO-O···Pd-F-TiO2 facilitates the photogenerated electron transfer from the conduction band to single atom Pd to reduce Pd···O-OH to HO·. The optimized HO· production rate is 9.18 μ mol L-1 min-1, which is 2.6-52.5 times higher than that in traditional advanced oxidation processes. In the application of wastewater treatment, this proposed photoelectrocatalytic oxygen reduction method, respectively, shows fast kinetics of 0.324 and 0.175 min-1 for removing bisphenol A and acetaminophen. Around 93.2% total organic carbon and 99.3% acute toxicity removal were achieved. Additionally, the degradation efficiency was less affected by the water source and pH value because of the evitable usage of metallic active sites. This work represents a fundamental investigation on the generation rate of HO·, which would pave the way for the future development of photoelectrocatalytic technologies for water purification.

Supracrustal rocks and garnet granite within the high-grade metamorphic complex in Daqingshan, at the northern margin of the Palaeoproterozoic Khondalite Belt, North China Craton (NCC), preserve various mineral assemblages that record three distinct stages of their metamorphic evolution. The mineral assemblages representative of prograde metamorphism (M1) comprise plagioclase (Pl) + quartz (Q) ± biotite (Bio) inclusions preserved in garnet porphyroblasts in gneiss and granite members. The metamorphic peak mineral assemblages (M2) are represented by garnet (Gt) + Pl + Bio + Q + sillimanite (Sil) + K-feldspar (Kfs) in each rock’s matrix. Retrograde mineral assemblages (M3) that formed during post-peak, near-isothermal decompression (ITD) occur between relict garnet porphyroblasts and matrix minerals. These M3 assemblages consist of Bio + cordierite (Crd) + Pl + Gt + Q + Sil + Kfs. Phase equilibria modelling shows that the supracrustal rocks record clockwise pressure–temperature (P–T) trajectories from 650 to 750 °C/5.3–8.9 kbar (M1) up to 800–830 °C/9.8–11.2 kbar (M2), to 760–830 °C/4.8–6.2 kbar (M3). Geochemical similarities between the Daqingshan supracrustal rocks (i.e. garnet-biotite gneisses) and garnet granites, such as relative enrichment in K, Rb, and Ba, and depletion in Th, U, Ta, Nb, P, and Ti, imply a petrological affinity. Sensitive high-resolution ion microprobe analysis (SHRIMP) U–Pb dating of zircons reveals that the protolith of the Daqingshan supracrustal rocks formed at ~2.5 Ga and subsequently metamorphosed at ~2.45–2.37 Ga. Anatexis likely occurred at ~2.43–2.40 Ga during the retrograde near-ITD stage. Combined with previous studies, the timing of metamorphism and anatexis in the Daqingshan area may be connected with the regional mafic magmatism.

DNA origami technology has rapidly developed into an ideal means to programmably crystallize nanoparticles. However, most existing DNA origami three-dimensional platforms normally used a single type of DNA origami unit, which greatly limits the types of nanoparticle superlattices that can be synthesized. Here, we report a universal strategy to vastly enrich the library of nanoparticle superlattices, based on multiple-unit (≥4 units) DNA origami platforms, which were constructed by programmably cocrystallizing three different DNA origami octahedral “homologs.” Through selectively inserting nanoparticles into DNA origami monomers, numerous nanoparticle superlattices can be synthesized on the basis of the same platform. In this work, we obtained 85 types of DOF/AuNP (DNA origami frame/gold nanoparticle) superlattices using three different DNA origami platforms as examples. We believe that our strategy can provide possible access to fabricate virtually endless types of nanoparticle superlattices and promote the construction of functional materials with special properties.

Administration of cisplatin, a common chemotherapeutic drug, has an inevitable side effect of sensorineural hearing loss. The main etiologies are stria vascularis injury, spiral ganglion degeneration, and hair cell death. Over several decades, the research scope of cisplatin-induced ototoxicity has expanded with the discovery of the molecular mechanism mediating inner ear cell death, highlighting the roles of reactive oxygen species and transport channels for cisplatin uptake into inner ear cells. Upon entering hair cells, cisplatin disrupts organelle metabolism, induces oxidative stress, and targets DNA to cause intracellular damage. Recent studies have also reported the role of inflammation in cisplatin-induced ototoxicity. In this article, we preform a narrative review of the latest reported molecular mechanisms of cisplatin-induced ototoxicity, from extracellular to intracellular. We build up a signaling network starting with cisplatin entering into the inner ear through the blood labyrinth barrier, disrupting cochlear endolymph homeostasis, and activating inflammatory responses of the outer hair cells. After entering the hair cells, cisplatin causes hair cell death via DNA damage, redox system imbalance, and mitochondrial and endoplasmic reticulum dysfunction, culminating in programmed cell death including apoptosis, necroptosis, autophagic death, pyroptosis, and ferroptosis. Based on the mentioned mechanisms, prominent therapeutic targets, such as channel-blocking drugs of cisplatin transporter, construction of cisplatin structural analogues, anti-inflammatory drugs, antioxidants, cell death inhibitors, and others, were collated. Considering the recent research efforts, we have analyzed the feasibility of the aforementioned therapeutic strategies and proposed our otoprotective approaches to overcome cisplatin-induced ototoxicity.

DNA nanotechnology has provided credible approaches for assembly of three-dimensional (3D) lattices with complex patterns. However, the symmetries are strictly dependent on their initial configurations and difficult to alter via non-thermal treatments. While switchable nucleic acid structures have been employed to construct deformable DNA motifs, it remains challenging to arrange them anisotropically in 3D lattices to trigger directed collective shape transition and dynamic symmetry conversion. In this work, we used octahedral DNA origami frames to synthesize four DNA origami lattices by placing the pH-reactive i-motif sequences in the desired dimensions. Thereinto, lattices with an anisotropic design can switch between simple cubic (SC) and simple tetragonal (ST) upon pH change. Small angle X-ray scattering (SAXS) results reveal the feasibility of obtaining 3D lattices with sensitive responses to external stimuli, expanding the way to obtain low-symmetry lattices.

The ongoing quest for understanding nonequilibrium dynamics of complex quantum systems underpins the foundation of statistical physics as well as the development of quantum technology. Quantum many-body scarring has recently opened a window into novel mechanisms for delaying the onset of thermalization, however its experimental realization remains limited to the $\mathbb{Z}_2$ state in a Rydberg atom system. Here we realize unconventional many-body scarring in a Bose--Hubbard quantum simulator with a previously unknown initial condition -- the unit-filling state. Our measurements of entanglement entropy illustrate that scarring traps the many-body system in a low-entropy subspace. Further, we develop a quantum interference protocol to probe out-of-time correlations, and demonstrate the system's return to the vicinity of the initial state by measuring single-site fidelity. Our work makes the resource of scarring accessible to a broad class of ultracold-atom experiments, and it allows to explore its relation to constrained dynamics in lattice gauge theories, Hilbert space fragmentation, and disorder-free localization.

Molybdenum carbide is a promising alternative of Pt/C in H2 evolution reaction (HER). However, its synthesis is time-consuming and energy-intensive. In this work, we propose a rapid one-pot strategy to fabricate C60-MoC nanocomposites in Li2CO3-K2CO3 molten salt using CO2 and ammonium molybdate as carbon and Mo sources, respectively. The as-obtained C60-MoC-600 sample at 600 °C shows high HER activity in both acid and alkaline electrolytes. Especially in 1 M KOH, the η 10 (overpotential at 10 mA cm−2) of C60-MoC-600 is 142 mV. Under an industrial current density of 220 mA cm−2, its activity with an overpotential of 250 mV is close to that of commercial Pt/C and exhibits excellent constant current stability during 10 h. This strategy not only implements the simple synthesis of MoC-based catalysts but also paves a highly efficient way for the rapid abatement and high-value-added utilization of CO2.

TiO2-mediated photoelectrocatalysis is emerging as a promising way to degrade refractory contaminates. Nevertheless, the concrete application of TiO2 is seriously limited because of its poor conductivity, unexpected recombination of photoinduced charges, and broad bandgap (3.2 eV). In this work, two-dimensional (2D) TiO2-g-C3N4 with both Ti N and C O bridges is successfully constructed and assembled with carbon fibers to realize efficient photoelectrochemical (PEC) pollutant degradation. Density functional theory (DFT) calculations suggest that the generated interface heterojunction of 2D TiO2-g-C3N4 can provide quick charge separation and transfer via both Ti N and C O bridges, resulting in a prepared catalyst that can facilitate the effective separation and transportation of photoinduced electron-hole pairs. In addition, according to electrochemical impedance spectroscopy, the 2D TiO2-g-C3N4 composition with the generated interface heterojunction reduces internal resistance and becomes more conducive to electrocatalysis compared with pure TiO2 or g-C3N4. Using bisphenol A (BPA) as a typical refractory contaminate, the 2D TiO2-g-C3N4/carbon fiber electrode exhibits a higher PEC activity, with reaction rates 1.7, 2.5, and 3 times faster than that of g-C3N4, TiO2, and commercial P25, respectively. Furthermore, the greatest BPA degradation of the PEC system is much higher than the sum of the photocatalytic (PC) and electrocatalytic (EC) systems. Additionally, the enhanced activities of electrocatalysis and photocatalysis for the degradation of BPA is attributed to the collaboration of the interfacial effect and excellent electrical conductivity derived from the 2D structure of the TiO2-g-C3N4 heterojunction. This study proposes a new tactic for the design and construction of photoelectrocatalysts via a synergistic interfacial effect to improve the photocatalytic and electrocatalytic degradation activities for refractory pollutants.

Biosensors are small but smart devices responding to the external stimulus, widely used in many fields including clinical diagnosis, healthcare and environment monitoring, etc. Moreover, there is still a pressing need to fabricate sensitive, stable, reliable sensors at present. DNA origami technology is able to not only construct arbitrary shapes in two/three dimension but also control the arrangement of molecules with different functionalities precisely. The functionalization of DNA origami nanostructure endows the sensing system potential of filling in weak spots in traditional DNA-based biosensor. Herein, we mainly review the construction and sensing mechanisms of sensing platforms based on DNA origami nanostructure according to different signal output strategies. It will offer guidance for the application of DNA origami structures functionalized by other materials. We also point out some promising directions for improving performance of biosensors.

Modular multilevel converter (MMC) owns high efficiency and its operation results in higher power quality. It has broad applications in the power system. However, unbalanced grid voltage leads to three-phase current imbalance and double frequency fluctuation power in MMC. In order to improve the performance of MMC under the unbalanced voltage conditions, this paper presents a control strategy for suppressing negative-sequence current based on the third harmonic voltage compensation method. The current output characteristics of MMC is analyzed under the situation of power grid voltage imbalance. For the purpose of maintaining the three-phase current balance, the mathematical model of the third harmonic voltage is established, which is used to Figure out the relationship between grid voltage unbalance degree and the third harmonic voltage compensation to keep the three phase current balance. The effectiveness and performance of the proposed control strategy are verified by the PSCAD/ EMTDC simulations.

Adversaries create phishing websites that spoof the visual appearances of frequently used legitimate websites in order to trick victims into providing their private information, such as bank accounts and login credentials. Phishing detection is an ongoing combat between the defenders and the attackers, where various defense mechanisms have been proposed, such as blacklists, heuristics, data mining, etc. In this paper, we present a new perspective on the identification of phishing websites. The proposed solution, namely PhishFencing, consists of three main steps: (1) filtering: a list of trusted and non-hosting websites is used to eliminate pages from legitimate hosts; (2) matching: a sub-graph matching mechanism is developed to determine if an unknown webpage contains logo images of whitelisted legitimate websites–once a match is detected, the unknown webpage is considered a suspicious page; (3) identification: host features are utilized to identify whether a suspicious webpage is hosted on the same cluster of servers as the corresponding legitimate pages–if not, the suspicious page is tagged as phishing. Compared with existing approaches in the literature, PhishFencing introduces an autonomous mechanism to replace the manual process of collecting and refreshing groundtruth data. As a in-network solution, PhishFencing could also partially detect phishing pages hosted on HTTPS servers, without requiring any support from clients. Through intensive experiments, we show that PhishFencing is very effective in comparing with the literature.

Hydrogen production can be promoted by replacing sluggish oxygen evolution reaction (OER) with a thermodynamically more favorable reaction, the primary oxidation reaction of benzyl alcohol to benzaldehyde. On a Bi2MoO6@TiO2NTA photocathode, the conversion of benzyl alcohol to benzaldehyde is realized with the selectivity of 100 %. This is originated from enhanced adsorption and activation of benzyl alcohol on this photoanode, as confirmed from tested by in situ FTIR techniques. The electrons generated during such a controllable and selective primary oxidation reaction is then utilized as the source for synergistical hydrogen production. The amount of generated hydrogen is then 5.5 times higher than that when OER is used. The efficiency for such hydrogen production is as high as 85 %. The proposed strategy combines solar energy and biomass for the efficient production of the valuable raw material - benzaldehyde as well as green energy source - hydrogen.

The degradation of pollutants with the simultaneous promotion of hydrogen production is an attractive and promising solution for the current environmental and energy problem, but few studies have been performed concerning the trends and mechanism of this process. Two classes of typical pollutants, phenols and alkylbenzenes, are treated using a self-biasing photoelectrochemical (PEC) cell to reveal the intrinsic relationships between the redox properties of the pollutants and hydrogen production. The results show that the TOC removal rates of these pollutants range from 63.5% to 93.6%. The highest hydrogen production efficiency (η H2 ), which is the ratio of experimental and theoretical hydrogen production amounts, reaches 81.5%. The degradation rate constant (k) decreases as the experimental photoelectrocatalytic oxidation potential (E PEC ) increases. This trend can be explained by the degradation pathway deduced from the detection of intermediate products using LC-MS and the associated reaction energetics from theoretical calculations. Neither the degradation rate constant (k) nor the TOC removal amounts (ΔTOC) determines the hydrogen production. The η H2 value positively depends on the average oxidation current efficiency (ACE), which is the ratio of experimental to theoretical TOC removal amounts. This study reveals a significant theory for the use of wastewater for hydrogen production.

It is of great practical significance to realize the efficient treatment of wastewater and the comprehensive utilization of chemical energy from wastewater. Therefore, a WO3 NFs–C/Cu2O nanowire arrays (NWAs) visible-light response dual-photoelectrodes solar-charged photoelectrochemical wastewater fuel cell (scPEWFC) was constructed for efficient hydrogen production based on the promotion of phenol oxidation at the anode. The hydrogen production reaches as high as 93.08 μmol cm−2 by the photoelectrocatalytic oxidation of phenol (TOC removal rate reached 82.12%) of WO3 NFs–C/Cu2O NWAs under visible light irradiation for 8 h without additional bias, which is 3.02 times higher than that of pure photocatalytic water splitting. The excellent photoelectrochemical performance can be attributed to the less endergonic process of phenol oxidation compared to water oxidation, which indicated that phenol oxidation can promote the hydrogen generation at the cathode. Isotopic labelling experiments show that protons were derived from the splitting of water rather than phenol. Interestingly, WO3 (with great electrons storage property) and photoelectrochemical cell (PEC) were combined to construct a scPEWFC, the hydrogen production (4 h to produce 7.50 μmol cm−2) was still carried out at the cathode under the dark owing to the stored electrons of WO3 under the light, which realized the ideal effect of full-day and high-efficiency hydrogen production. Besides, the mechanism of scPEWFC operation was investigated in detail. This paper provides a new research idea for scPEWFC wastewater treatment, energy recovery, electron storage, and simultaneous full-day hydrogen production.

Photoelectrocatalytic conversion of biomass and its derivatives is a very significant and promising strategy for energy production. The selective oxidation of biomass alcohol to the corresponding aldehydes is an important class of biomass conversion reactions. In this study, a Z-scheme type solar-driven dual photoelectrode photoelectrochemical cell was used for the simultaneous selective oxidation of benzyl alcohol, which was a model biomass alcohol to benzaldehyde reaction, for hydrogen production promotion using Au/TiO2 nanotube photonic crystals (Au/TiO2 NTPC) as the photoanode. The results show that the Au/TiO2 NTPC photoanode exhibits excellent photoelectrochemical performance and can efficiently conduct charge transport. The conversion of benzyl alcohol is as high 84.68%, whereas selectivity can reach more than 99%. The high conversion rate is attributed to the improvement in the photoelectrocatalytic performance due to the in situ doping of Au nanoparticles. In addition to the Au nanoparticles, O2 is an indispensable factor for the high selectivity. The simultaneous hydrogen production at the cathode can reach 143.83 μmol cm−2. The results show that the anodic oxidation reaction can promote hydrogen production and the hydrogen production efficiency increases with the oxidation rate. This model reaction can also be applied to a series of biomass alcohol selective oxidations. This study achieves simultaneous efficient selective access to hydrocarbon fuels and hydrogen and reveals a novel, sustainable and green energy production idea and direction for development.

TiO

Polymorphs of MnO2 have been studied extensively for their potential application in rechargeable batteries which provide attractive structures as the host for the cation insertion. In this study, the simplest form of δ-MnO2 has been taken as the representative of layered MnO2 and the ion diffusion performance is carefully examined. The potential energy surfaces of Li+, Na+, K+, NH4 +, Mg2+, Zn2+, Ca2+ and Al3+ ion diffusing in the 1T-δ-MnO2 are firstly obtained within the density functional theory framework, which give a comprehensive and detailed view of the diffusing process. We not only assess the 1T-δ-MnO2 as candidate electrode material quantitatively but also elucidate the influence of the cooperative Jahn–Teller effect which governs multivalent diffusion in the intuitive and clear view. Our findings are well matched with the reported results and lead to practical and implementable guidelines to find MnO2-based fast-diffusion multivalent ion conductors.

Wireless devices have been widely adopted across all domains. With the convenience brought by wireless communication technology, increasing number of conventional (wired) devices are evolving to become wireless. However, significant security issues arise with the popularity of wireless devices. To start an attack, the adversary usually performs a network reconnaissance to discover exposed devices, identify device manufacturers and types, and then scan for vulnerabilities. From the defense side, network administrators are expected to identify the potential vulnerabilities/risks and enforce Network Access Control (or Network Admission Control, NAC) on all the connecting devices. To do this, it is essential to accurately identify the make/model/type of each device that attempts to connect to the network, e.g., MacBooks, Samsung smart phones (Android), Amazon kindles, DLink surveillance cameras, TP-Link smart plugs, etc. In this paper, we present a novel approach, namely WDMTI, for the identification of wireless device manufacturer and type. We tackle the challenge from two aspects: the features and the classification model. First, we claim that it is critical to discover the device manufacturer and type as soon as the device requests to join the WLAN, and it is unrealistic to make other assumptions on the status of the device, e.g., assuming that the device is booting up or initializing a new connection to corresponding servers/clouds. We primarily depend on the features extracted from the network connection phase, while features from device booting are considered "bonus". In particular, we propose to utilize features from the raw HDCP packets, which is shown to be sufficient for device manufacturer and type recognition with high accuracy. Meanwhile, in the WDMTI system, we employ the Hierarchical Dirichlet Process (HDP), which is a nonparametric Bayesian model for grouped data. HDP allows new groups to be introduced with new data being added, i.e. previously unknown devices connect to the network and the extracted features receive new labels. The WDMTI mechanism is dynamically retrained on-line, instead of requiring a time-consuming off-line retraining process. Our experiments show that WDMTI identifies known types of devices with average accuracy of 0.89, and new types of devices with average accuracy of 0.96, both of which is higher than the state-of-art approaches. In summary, we present a wireless device manufacturer and type identification (WDMTI) system that is both scalable and accurate, and capable of adapting to unknown types of devices on-the-fly.

Ro-vibrational term values of the 3(1)Σg (+) state of (85,85)Rb2 and (85,87)Rb2 and resolved fluorescence spectra to the A(1)Σu (+) state are recorded following optical-optical double resonance excitation. The experimental data are heavily perturbed, and as a result, the standard analysis based on Dunham series representation of the energy levels fails. The analysis is done via modeling the adiabatic potential function with the Rydberg-Klein-Rees potential constructed from the generalized smoothing spline interpolation of the vibrational energies Gv and rotational constants Bv.