This work presents Keep it Simple (KiS), a new approach to unsupervised text simplification which learns to balance a reward across three properties: fluency, salience and simplicity. We train the model with a novel algorithm to optimize the reward (
k-SCST), in which the model proposes several candidate simplifications, computes each candidates reward, and encourages candidates that outperform the mean reward. Finally, we propose a realistic text comprehension task as an evaluation method for text simplification. When tested on the English news domain, the KiS model outperforms strong supervised baselines by more than 4 SARI points, and can help people complete a comprehension task an average of 18% faster while retaining accuracy, when compared to the original text. Code available: https://github.com/tingofurro/keep_it_simple
We have followed up two ultra-diffuse galaxies (UDGs), detected adjacent to stellar streams, with Hubble Space Telescope (HST) imaging and HI mapping with the Jansky Very Large Array (VLA) in order to investigate the possibility that they might have
a tidal origin. With the HST F814W and F555W images we measure the globular cluster (GC) counts for NGC 2708-Dw1 and NGC 5631-Dw1 as $2^{+1}_{-1}$ and $5^{+1}_{-2}$, respectively. NGC 2708-Dw1 is undetected in HI down to a 3$sigma$ limit of $log (M_mathrm{HI}/mathrm{M_odot}) = 7.3$, and there is no apparent HI associated with the nearby stellar stream. There is a 2$sigma$ HI feature coincident with NGC 5631-Dw1. However, this emission is blended with a large gaseous tail emanating from NGC 5631 and is not necessarily associated with the UDG. The presence of any GCs and the lack of clear HI connections between the UDGs and their parent galaxies strongly disfavor a tidal dwarf galaxy origin, but cannot entirely rule it out. The GC counts are consistent with those of normal dwarf galaxies, and the most probable formation mechanism is one where these UDGs were born as normal dwarfs and were later tidally stripped and heated. We also identify an over-luminous ($M_mathrm{V} = -11.1$) GC candidate in NGC 2708-Dw1, which may be a nuclear star cluster transitioning to an ultra-compact dwarf as the surrounding dwarf galaxy gets stripped of stars.
We present atomic hydrogen (HI) observations using the Robert C. Byrd Green Bank Telescope along the lines-of-sight to 27 low surface brightness (LSB) dwarf galaxy candidates discovered in optical searches around M101. We detect HI reservoirs in 5 ta
rgets and place stringent upper limits on the remaining 22, implying that they are gas poor. The distances to our HI detections range from 7 Mpc --150 Mpc, demonstrating the utility of wide-bandpass HI observations as a follow-up tool. The systemic velocities of 3 detections are consistent with that of the NGC~5485 group behind M101, and we suggest that our 15 non-detections with lower distance limits from the optical are associated with and have been stripped by that group. We find that the gas richnesses of confirmed M101 satellites are broadly consistent with those of the Milky Way satellites, as well as with those of satellites around other hosts of comparable mass, when survey completeness is taken into account. This suggests that satellite quenching and gas stripping proceeds similarly around halos of similar mass, in line with theoretical expectations.
We report the discovery of two ultra-diffuse galaxies (UDGs) which show clear evidence for association with tidal material and interaction with a larger galaxy halo, found during a search of the Wide portion of the Canada-France-Hawaii Telescope Lega
cy Survey (CFHTLS). The two new UDGs, NGC2708-Dw1 and NGC5631-Dw1, are faint ($M_g$=$-$13.7 and $-$11.8 mag), extended ($r_h$=2.60 and 2.15 kpc) and have low central surface brightness ($mu(g,0)$=24.9 and 27.3 mag arcsec$^{-2}$), while the stellar stream associated with each has a surface brightness $mu(g)$$gtrsim$28.2 mag arcsec$^{-2}$. These observations provide evidence that the origin of some UDGs may connect to galaxy interactions, either by transforming normal dwarf galaxies by expanding them, or because UDGs can collapse out of tidal material (i.e. they are tidal dwarf galaxies). Further work is needed to understand the fraction of the UDG population `formed through galaxy interactions, and wide field searches for diffuse dwarf galaxies will provide further clues to the origin of these enigmatic stellar systems.
