A significant fraction of high redshift star-forming disc galaxies are known to host giant clumps, whose formation, nature and role in galaxy evolution are yet to be understood. In this work we first present a new automated method based on deep neural networks to detect clumps in galaxy images and show that it is more sensitive and faster than previous proposed methods. We then use this method to systematically detect clumps in the rest-frame optical and UV images of a complete sample of $\sim1500$ star forming galaxies at $1<z<3$ in the CANDELS survey as well as in images from the VELA forward modeled zoom-in cosmological simulations. We show that observational effects have a dramatic impact on the derived clump properties leading to an overestimation of the clump mass up to a factor of 10, which highlights the importance of fair and direct comparisons between observations and simulations. After correcting for these effects with a simple mixture density network, we estimate that the clump stellar mass function follows a power-law down to the completeness limit ($10^{7}$ solar masses) with the majority of the clumps being less massive than $10^9$ solar masses. We find that between $20-40\%$ of $\sim10^{10}$ solar mass star-forming galaxies contain at least one massive ($M_c>10^{7}M_\odot$) off-centered clump but only $\sim2-5\%$ of the total galaxy stellar mass is in those clumps. We also show indications that the contribution of clumps to the stellar mass is more important in extended and low mass galaxies. The simulations explored in this work overall reproduce the shape of the observed clump stellar mass function when confronted under the same conditions although tend to lie in the lower limit of the confidence intervals of the observations. This agreement suggests that most of the observed clumps are formed in-situ through violent disk instabilities.

Using analytic modeling and simulations, we address the origin of an abundance of star-forming, clumpy, extended gas rings about massive central bodies in massive galaxies at $z \!<\! 4$. Rings form by high-angular-momentum streams and survive in galaxies of $M_{\rm star} \!>\! 10^{9.5-10} M_\odot$ where merger-driven spin flips and supernova feedback are ineffective. The rings survive after events of compaction to central nuggets. Ring longevity was unexpected based on inward mass transport driven by torques from violent disc instability. However, evaluating the torques from a tightly wound spiral structure, we find that the timescale for transport per orbital time is long and $\propto\! \delta_{\rm d}^{-3}$, with $\delta_{\rm d}$ the cold-to-total mass ratio interior to the ring. A long-lived ring forms when the ring transport is slower than its replenishment by accretion and the interior depletion by SFR, both valid for $\delta_{\rm d} \!<\! 0.3$. The central mass that lowers $\delta_{\rm d}$ is a compaction-driven bulge and/or dark matter, aided by the lower gas fraction at $z \!<\! 4$, provided that it is not too low. The ring is Toomre unstable for clump and star formation. The high-$z$ dynamic rings are not likely to arise form secular resonances or collisions. AGN feedback is not expected to affect the rings. Mock images of simulated rings through dust indicate qualitative consistency with observed rings about bulges in massive $z\!\sim\!0.5\!-\!3$ galaxies, in $H_{\alpha}$ and deep HST imaging. ALMA mock images indicate that $z\!\sim\!0.5\!-\!1$ rings should be detectable. We quote expected observable properties of rings and their central nuggets.

Using the NIHAO suite of cosmological zoom-in simulations, we explore a parametric function for the dark-matter halo density profile subject to baryonic effects. This profile has variable inner slope and concentration parameter, and it allows analytic expressions for the gravitational potential, velocity dispersion, and lensing properties. We show that this profile provides better fits to simulated profiles than the Einasto profile and the generalized NFW profile with variable inner slope, in particular towards the halo centers, and that its parameters are correlated with the stellar-to-halo mass ratio $M_{\rm star}/M_{\rm vir}$. This defines a mass-dependent density profile describing the dark matter profiles in all galaxies, which can be directly applied to observed rotation curves of galaxies, gravitational lenses, and semi-analytic models of galaxy formation and satellite-galaxy evolution. The effect of baryons manifests itself in the profile parameters by a sharp drop of the inner density slope and a 20% decrease of the concentration parameter for $\log(M_{\rm star}/M_{\rm vir})$ between -3.5 and -2, which corresponds to $M_{\rm star}$ between $10^7$ and $10^{10}~\rm M_\odot$. The accuracy by which this profile fits simulated galaxies is similar to certain multi-parameter, mass-dependent profiles, but its fewer parameters and analytic nature make it most desirable for many purposes.

We predict, analytically and by simulations, that gas discs tend to survive only in haloes above a threshold mass $\sim 2 \times 10^{11} M_\odot$ (stellar mass $\sim 10^9 M_\odot$), with only a weak redshift dependence. At lower masses, the disc spins typically flip in less than an orbital time due to mergers associated with a change in the pattern of the feeding cosmic-web streams. This threshold arises from the halo merger rate when accounting for the mass dependence of the ratio of galactic baryons and halo mass. Above the threshold, wet compactions lead to massive central nuggets that allow the longevity of extended clumpy gas rings. Supernova feedback has a major role in disrupting discs below the critical mass, by driving the stellar-to-halo mass ratio that affects the merger rate, by stirring up turbulence and suppressing high-angular-momentum gas supply, and by confining major compactions to the critical mass. Our predictions seem consistent with current observed fractions of gas discs, to be explored by future observations that will resolve galaxies below $10^9 M_\odot$ at high redshifts, e.g. by JWST.