Yue Wang

Perovskite synthesis out in the open Although methods have been developed that create the photoactive black perovskite phase of formamidinium lead iodide (α-FAPbI 3 ), these routes are temperature and humidity sensitive and less compatible with large-scale solar cell production. Hui et al. report an alternative route in which vertically aligned lead iodide thin films are grown from the ionic liquid methylamine formate. Nanoscale channels in the films lower the barrier to permeation of formamidinium iodide and enable transformation to α-FAPbI 3 , even at high humidity and room temperature. Solar cells made with these films have power conversion efficiencies as high as 24.1% that display high stability. Science , this issue p. 1359

Dropout is a powerful and widely used technique to regularize the training of deep neural networks. In this paper, we introduce a simple regularization strategy upon dropout in model training, namely R-Drop, which forces the output distributions of different sub models generated by dropout to be consistent with each other. Specifically, for each training sample, R-Drop minimizes the bidirectional KL-divergence between the output distributions of two sub models sampled by dropout. Theoretical analysis reveals that R-Drop reduces the freedom of the model parameters and complements dropout. Experiments on $\bf{5}$ widely used deep learning tasks ($\bf{18}$ datasets in total), including neural machine translation, abstractive summarization, language understanding, language modeling, and image classification, show that R-Drop is universally effective. In particular, it yields substantial improvements when applied to fine-tune large-scale pre-trained models, e.g., ViT, RoBERTa-large, and BART, and achieves state-of-the-art (SOTA) performances with the vanilla Transformer model on WMT14 English$\to$German translation ($\bf{30.91}$ BLEU) and WMT14 English$\to$French translation ($\bf{43.95}$ BLEU), even surpassing models trained with extra large-scale data and expert-designed advanced variants of Transformer models. Our code is available at GitHub{\url{https://github.com/dropreg/R-Drop}}.

We demonstrated a stable and spectrum-selective self-powered UV photodetector based on lead-free Cs₃Cu₂I₅ films with excellent photodetection performance.

Ligand-free rutile and anatase TiO₂nanocrystals were synthesized<italic>via</italic>a hydrolytic sol–gel method. The improved power conversion efficiency was achieved by using anatase TiO₂nanocrystals as electron extraction layer in the organic solar cells.

While halide perovskites (HPs) have achieved enormous success in the field of optoelectronic applications, much attention has been recently drawn to the unique polarization sensitivity of HPs, either intrinsic or extrinsic, which makes HPs a potential candidate for innovative applications in directly polarized luminescence and detection. Herein, the research status in the field of polarization-sensitive HPs, including linear polarization and circular polarization, is comprehensively summarized. To evaluate the effectiveness of HPs in generating and detecting linearly or circularly polarized light, the principles and characterization methods of polarized luminescence and detection are introduced. Sequentially, the state-of-the-art development of the strategies that induce the linear or circular polarization characteristics of HPs is systematically reviewed, based on which the application of polarization-sensitive HPs in the field of polarization luminescence and detection are summarized. Moreover, the current challenges and opportunities are discussed, and prospects of the future development in this promising field are outlined.