Fig. 1. Masked images and corresponding inpainted results using our partialconvolution based network.Abstract. Existing deep learning based image inpainting methods use a standard convolutional network over the corrupted image, using convolutional filter responses conditioned on both valid pixels as well as the substitute values in the masked holes (typically the mean value). This often leads to artifacts such as color discrepancy and blurriness. Postprocessing is usually used to reduce such artifacts, but are expensive and may fail. We propose the use of partial convolutions, where the convolution is masked and renormalized to be conditioned on only valid pixels. We further include a mechanism to automatically generate an updated mask for the next layer as part of the forward pass. Our model outperforms other methods for irregular masks. We show qualitative and quantitative comparisons with other methods to validate our approach.

The prevalence and energetics of quasar feedback is a major unresolved problem in galaxy formation theory. In this paper, we present Gemini Integral Field Unit observations of ionized gas around eleven luminous, obscured, radio-quiet quasars at z ∼ 0.5 out to ∼ 15 kpc from the quasar; specifically, we measure the kinematics and morphology of [O iii]λ5007Å emission. The round morphologies of the nebulae and the large line-of-sight velocity widths (with velocities containing 80% of the emission as high as 10 3 km s −1 ) combined with relatively small velocity difference across them (from 90 to 520 km s −1 ) point toward wide-angle quasi-spherical outflows. We use the observed velocity widths to estimate a median outflow velocity of 760 km s −1 , similar to or above the escape velocities from the host galaxies. The line-of-sight velocity dispersion declines slightly toward outer parts of the nebulae (by 3% per kpc on average). The majority of nebulae show blueshifted excesses in their line profiles across most of their extents, signifying gas outflows. For the median outflow velocity, we finḋ E kin between 4 × 10 44 and 3 × 10 45 erg s −1 andṀ between 2 × 10 3 and 2 × 10 4 M ⊙ yr −1 . These values are large enough for the observed quasar winds to have a significant impact on their host galaxies. The median rate of converting bolometric luminosity to kinetic energy of ionized gas clouds is ∼2%. We report four new candidates for "super-bubbles" -outflows that may have broken out of the denser regions of the host galaxy.

Black hole feedback -the strong interaction between the energy output of supermassive black holes and their surrounding environments -is routinely invoked to explain the absence of overly luminous galaxies, the black hole vs. bulge correlations and the similarity of black hole accretion and star formation histories. Yet direct probes of this process in action are scarce and limited to small samples of active nuclei. In this paper we present Gemini Integral Field Unit observations of the distribution of ionized gas around luminous, obscured, radio-quiet quasars at z ∼ 0.5. We detect extended ionized gas nebulae via [O iii]λ5007Å emission in every case, with a mean diameter of 28 kpc. These nebulae are nearly perfectly

Well-defined ultrathin nanoribbons have been fabricated from an amphiphilic electron donor-acceptor (D-A) supramolecule comprising perylene tetracarboxylic diimide as the backbone scaffold to enforce the one-dimensional intermolecular assembly via strong pi-stacking. These nanoribbons demonstrated high photoconductivity upon illumination with white light. The high photoconductivity thus obtained is likely due to the optimal molecular design that enables a good kinetic balance between the two competitive processes, the intramolecular charge recombination (between D and A) and the intermolecular charge transport along the nanoribbon. The photoconduction response has also proven to be prompt and reproducible with the light turning on and off. The photogenerated electrons within the nanoribbon can be efficiently trapped by the adsorbed oxygen molecules or other oxidizing species, leading to depletion of the charge carriers (and thus the electrical conductivity) of the nanoribbon, as typically observed for n-type semiconductor materials as applied in chemiresistors. Combination of this sensitive modulation of conductivity with the unique features intrinsic to the nanoribbon morphology (large surface area and continuous nanoporosity when deposited on a substrate to form a fibril film) enables efficient vapor sensing of nitro-based explosives.