With recent progress in graphics, it has become more tractable to train models on synthetic images, potentially avoiding the need for expensive annotations. However, learning from synthetic images may not achieve the desired performance due to a gap between synthetic and real image distributions. To reduce this gap, we propose Simulated+Unsupervised (S+U) learning, where the task is to learn a model to improve the realism of a simulator's output using unlabeled real data, while preserving the annotation information from the simulator. We develop a method for S+U learning that uses an adversarial network similar to Generative Adversarial Networks (GANs), but with synthetic images as inputs instead of random vectors. We make several key modifications to the standard GAN algorithm to preserve annotations, avoid artifacts, and stabilize training: (i) a 'self-regularization' term, (ii) a local adversarial loss, and (iii) updating the discriminator using a history of refined images. We show that this enables generation of highly realistic images, which we demonstrate both qualitatively and with a user study. We quantitatively evaluate the generated images by training models for gaze estimation and hand pose estimation. We show a significant improvement over using synthetic images, and achieve state-of-the-art results on the MPIIGaze dataset without any labeled real data.

Diatom is an important group of marine algae and contributes to around 20% of the global photosynthetic carbon fixation. Photosystem I (PSI) of diatoms is associated with a large number of fucoxanthin-chlorophyll a/c proteins (FCPIs). We report the structure of PSI-FCPI from a diatom Chaetoceros gracilis at 2.38 Å resolution by single-particle cryo-electron microscopy. PSI-FCPI is a monomeric supercomplex consisting of 12 core and 24 antenna subunits (FCPIs), and 326 chlorophylls a, 34 chlorophylls c, 102 fucoxanthins, 35 diadinoxanthins, 18 β-carotenes and some electron transfer cofactors. Two subunits designated PsaR and PsaS were found in the core, whereas several subunits were lost. The large number of pigments constitute a unique and huge network ensuring efficient energy harvesting, transfer and dissipation. These results provide a firm structural basis for unraveling the mechanisms of light-energy harvesting, transfer and quenching in the diatom PSI-FCPI, and also important clues to evolutionary changes of PSI-LHCI.

No abstract

A self-healing hydrogel ionic conductor has been developed by combining dynamic covalent chemistry with nanofiller reinforcement and micelle crosslinking, and used for sensing of diverse human activities.