Pixelated High-Q Metasurfaces for in Situ Biospectroscopy and Artificial Intelligence-Enabled Classification of Lipid Membrane Photoswitching Dynamics.

Nanophotonic devices excel at confining light into intense hot spots of electromagnetic near fields, creating exceptional opportunities for light-matter coupling and surface-enhanced sensing. Recently, all-dielectric metasurfaces with ultrasharp resonances enabled by photonic bound states in the continuum (BICs) have unlocked additional functionalities for surface-enhanced biospectroscopy by precisely targeting and reading out the molecular absorption signatures of diverse molecular systems. However, BIC-driven molecular spectroscopy has so far focused on end point measurements in dry conditions, neglecting the crucial interaction dynamics of biological systems. Here, we combine the advantages of pixelated all-dielectric metasurfaces with deep learning-enabled feature extraction and prediction to realize an integrated optofluidic platform for time-resolved in situ biospectroscopy. Our approach harnesses high-Q metasurfaces specifically designed for operation in a lossy aqueous environment together with advanced spectral sampling techniques to temporally resolve the dynamic behavior of photoswitchable lipid membranes. Enabled by a software convolutional neural network, we further demonstrate the real-time classification of the characteristic cis and trans membrane conformations with 98% accuracy. Our synergistic sensing platform incorporating metasurfaces, optofluidics, and deep learning reveals exciting possibilities for studying multimolecular biological systems, ranging from the behavior of transmembrane proteins to the dynamic processes associated with cellular communication.

Dynamic response characteristics and damage rule of graphite ore rock under different strain rates.

In the process of mining graphite mine, rock mass is often subjected to dynamic loads such as blasting or mechanical crushing, which involves dynamic responses of different strain rates, and blasting and crushing effect are affected by the rock dynamic properties and damage specials. The dynamic response characteristics and damage rule of graphite ore rock under different strain rates are very important but rarely studied in the past. To study these issues and provide support for graphite ore rock mining, the dynamic compression tests of graphite ore rock under five kinds of impact pressures were designed and carried out by using the Split Hopkinson Pressure Bar (SHPB) test system, combining with the high-speed photography system and crushing screening tests. The dynamic characteristics, crushing process, crushing mode, crushing form and fragmentation distribution of the graphite ore rock under different strain rates were analyzed. The results show that the dynamic characteristics of the graphite ore rock have obvious strain rate effect. The hardening coefficient (DIF) is positively correlated with the cubic root of strain rate, and the softening factor (K) is negatively correlated with the cubic root of strain rate. Shear failure mainly occurs in the graphite ore rock under impact load, and the crushing process can be divided into five stages, they are compaction, crack initiation, crack expansion and penetration, fragmentation collision and fragmentation fall. In addition, the crushed blocks are mainly triangular pyramid (or cone-like) fine granular and powder. The broken fragments of the graphite ore rock are in accord with the fractal geometry characteristics. That is, the average broken particle size (dS) decreases linearly with the increase of strain rate, and the fractal dimension (Da) increases weakly with the increase of strain rate. Based on D-P fracture criterion and Weibull distribution model, the dynamic damage constitutive model of the graphite ore rock was established, and the correlation between strain rate and Weibull distribution parameters (m and F0) was used to reasonably modify the damage constitutive model. The modified damage constitutive model curve is in good agreement with the experimental curve, which can basically reflect the strain rate effect of the dynamic characteristics of the graphite ore rock and the evolution characteristics of the dynamic stress-strain curve at different stage.

Biaxial tensile test and meso damage numerical simulation of HTPB propellant.

Aiming at the shortcomings of the current research on the mechanical properties of solid propellants under complex stress conditions, an effective cross-shaped test piece configuration and variable-scale biaxial tensile test method are designed in this paper, and the meso-simulation model of propellant is constructed by Micro-CT test and random filling algorithm. Then, based on the Hook-Jeeves method and the cohesive force model, the mechanical performance parameters of each mesoscopic component were obtained, and finally the damage evolution process of the propellant was numerically simulated. The results show that the stress-strain curve of the propellant under biaxial loading is similar to that of uniaxial stretching, and has obvious rate dependence and stress state dependence. The mechanical properties of the propellant under biaxial tensile loading are significantly lower than those in uniaxial stretching, and the maximum elongation is only 45-85% of that in uniaxial stretching. The fracture process of propellant can be divided into initial linear stage, damage evolution stage and fracture stage. The dewetting phenomenon generally occurs at the interface between the large-sized AP particles and the matrix. With the loading of the load, the pores formed by the dewetting and matrix tearing continue to converge into cracks and expand in the direction perpendicular to the resultant force, and finally fracture. The propellant dehumidifies more easily under high strain rate loading, but the degree of dewetting is lower when the same strain is reached.

Instance-Dependent Positive and Unlabeled Learning with Labeling Bias Estimation.

This paper studies instance-dependent Positive and Unlabeled (PU) classification, where whether a positive example will be labeled (indicated by s) is not only related to the class label y, but also depends on the observation x. Therefore, the labeling probability on positive examples is not uniform as previous works assumed, but is biased to some simple or critical data points. To depict the above dependency relationship, a graphical model is built in this paper which further leads to a maximization problem on the induced likelihood function regarding P(s,y|x). By utilizing the well-known EM and Adam optimization techniques, the labeling probability of any positive example P(s=1|y=1,x) as well as the classifier induced by P(y|x) can be acquired. Theoretically, we prove that the critical solution always exists, and is locally unique for linear model if some sufficient conditions are met. Moreover, we upper bound the generalization error for both linear logistic and non-linear network instantiations of our algorithm. Empirically, we compare our method with state-of-the-art instance-independent and instance-dependent PU algorithms on a wide range of synthetic, benchmark and real-world datasets, and the experimental results firmly demonstrate the advantage of the proposed method over the existing PU approaches.