Search for: All records

Sapphire has various applications in photonics due to its broadband transparency, high-contrast index, and chemical and physical stability. Photonics integration on the sapphire platform has been proposed, along with potentially high-performance lasers made of group III–V materials. In parallel with developing active devices for photonics integration applications, in this work, silicon nitride optical waveguides on a sapphire substrate were analyzed using the commercial software Comsol Multiphysics in a spectral window of 800~2400 nm, covering the operating wavelengths of III–V lasers, which could be monolithically or hybridly integrated on the same substrate. A high confinement factor of ~90% near the single-mode limit was obtained, and a low bending loss of ~0.01 dB was effectively achieved with the bending radius reaching 90 μm, 70 μm, and 40 μm for wavelengths of 2000 nm, 1550 nm, and 850 nm, respectively. Furthermore, the use of a pedestal structure or a SiO2 bottom cladding layer has shown potential to further reduce bending losses. The introduction of a SiO2 bottom cladding layer effectively eliminates the influence of the substrate’s larger refractive index, resulting in further improvement in waveguide performance. The platform enables tightly built waveguides and small bending radii with high field confinement and low propagation losses, showcasing silicon nitride waveguides on sapphire as promising passive components for the development of high-performance and cost-effective PICs. 

Abstract Haze in Beijing is linked to atmospherically formed secondary organic aerosol, which has been shown to be particularly harmful to human health. However, the sources and formation pathways of these secondary aerosols remain largely unknown, hindering effective pollution mitigation. Here we have quantified the sources of organic aerosol via direct near-molecular observations in central Beijing. In winter, organic aerosol pollution arises mainly from fresh solid-fuel emissions and secondary organic aerosols originating from both solid-fuel combustion and aqueous processes, probably involving multiphase chemistry with aromatic compounds. The most severe haze is linked to secondary organic aerosols originating from solid-fuel combustion, transported from the Beijing–Tianjing–Hebei Plain and rural mountainous areas west of Beijing. In summer, the increased fraction of secondary organic aerosol is dominated by aromatic emissions from the Xi’an–Shanghai–Beijing region, while the contribution of biogenic emissions remains relatively small. Overall, we identify the main sources of secondary organic aerosol affecting Beijing, which clearly extend beyond the local emissions in Beijing. Our results suggest that targeting key organic precursor emission sectors regionally may be needed to effectively mitigate organic aerosol pollution. 

To address the sample selection bias between the training and test data, previous research works focus on reweighing biased training data to match the test data and then building classification models on there weighed raining data. However, how to achieve fairness in the built classification models is under-explored. In this paper, we propose a framework for robust and fair learning under sample selection bias. Our framework adopts there weighing estimation approach for bias correction and the minimax robust estimation approach for achieving robustness on prediction accuracy. Moreover, during the minimax optimization, the fairness is achieved under the worst case, which guarantees the model’s fairness on test data. We further develop two algorithms to handle sample selection bias when test data is both available and unavailable. 

In differentially private stochastic gradient descent (DPSGD), gradient clipping and random noise addition disproportionately affect underrepresented and complex classes and subgroups. As a consequence, DPSGD has disparate impact: the accuracy of a model trained using DPSGD tends to decrease more on these classes and subgroups vs. the original, non-private model. If the original model is unfair in the sense that its accuracy is not the same across all subgroups, DPSGD exacerbates this unfairness. In this work, we study the inequality in utility loss due to differential privacy, which compares the changes in prediction accuracy w.r.t. each group between the private model and the non-private model. We analyze the cost of privacy w.r.t. each group and explain how the group sample size along with other factors is related to the privacy impact on group accuracy. Furthermore, we propose a modified DPSGD algorithm, called DPSGD-F, to achieve differential privacy, equal costs of differential privacy, and good utility. DPSGD-F adaptively adjusts the contribution of samples in a group depending on the group clipping bias such that differential privacy has no disparate impact on group accuracy. Our experimental evaluation shows the effectiveness of our removal algorithm on achieving equal costs of differential privacy with satisfactory utility.