arXiv subjects

Miao Li

Publications and source records attributed to Miao Li.

At least 127 records · Page 7Linked to original sources

arXiv2021

Data Generation for Learning to Grasp in a Bin-picking Scenario

The rise of deep learning has greatly transformed the pipeline of robotic grasping from model-based approach to data-driven stream. Along this line, a large scale of grasping data either collected from simulation or from real world examples become extremely important. In this paper, we present our recent work on data generation in simulation for a bin-picking scene. 77 objects from the YCB object data sets are used to generate the dataset with PyBullet, where different environment conditions are taken into account including lighting, camera pose, sensor noise and so on. In all, 100K data samples are collected in terms of ground truth segmentation, RGB, 6D pose and point cloud. All the data examples including the source code are made available online.

cs.RO↗

arXiv2021

Unknown Object Segmentation through Domain Adaptation

The ability to segment unknown objects in cluttered scenes has a profound impact on robot grasping. The rise of deep learning has greatly transformed the pipeline of robotic grasping from model-based approach to data-driven stream, which generally requires a large scale of grasping data either collected in simulation or from real-world examples. In this paper, we proposed a sim-to-real framework to transfer the object segmentation model learned in simulation to the real-world. First, data samples are collected in simulation, including RGB, 6D pose, and point cloud. Second, we also present a GAN-based unknown object segmentation method through domain adaptation, which consists of an image translation module and an image segmentation module. The image translation module is used to shorten the reality gap and the segmentation module is responsible for the segmentation mask generation. We used the above method to perform segmentation experiments on unknown objects in a bin-picking scenario. Finally, the experimental result shows that the segmentation model learned in simulation can be used for real-world data segmentation.

cs.RO↗

arXiv2021

Domain-Specific Suppression for Adaptive Object Detection

Domain adaptation methods face performance degradation in object detection, as the complexity of tasks require more about the transferability of the model. We propose a new perspective on how CNN models gain the transferability, viewing the weights of a model as a series of motion patterns. The directions of weights, and the gradients, can be divided into domain-specific and domain-invariant parts, and the goal of domain adaptation is to concentrate on the domain-invariant direction while eliminating the disturbance from domain-specific one. Current UDA object detection methods view the two directions as a whole while optimizing, which will cause domain-invariant direction mismatch even if the output features are perfectly aligned. In this paper, we propose the domain-specific suppression, an exemplary and generalizable constraint to the original convolution gradients in backpropagation to detach the two parts of directions and suppress the domain-specific one. We further validate our theoretical analysis and methods on several domain adaptive object detection tasks, including weather, camera configuration, and synthetic to real-world adaptation. Our experiment results show significant advance over the state-of-the-art methods in the UDA object detection field, performing a promotion of $10.2\sim12.2\%$ mAP on all these domain adaptation scenarios.

cs.CV↗

arXiv2020

Space Telescope and Optical Reverberation Mapping Project. IX. Velocity-Delay Maps for Broad Emission Lines in NGC 5548

We report velocity-delay maps for prominent broad emission lines, Ly_alpha, CIV, HeII and H_beta, in the spectrum of NGC5548. The emission-line responses inhabit the interior of a virial envelope. The velocity-delay maps reveal stratified ionization structure. The HeII response inside 5-10 light-days has a broad single-peaked velocity profile. The Ly_alpha, CIV, and H_beta responses peak inside 10 light-days, extend outside 20 light-days, and exhibit a velocity profile with two peaks separated by 5000 km/s in the 10 to 20 light-day delay range. The velocity-delay maps show that the M-shaped lag vs velocity structure found in previous cross-correlation analysis is the signature of a Keplerian disk with a well-defined outer edge at R=20 light-days. The outer wings of the M arise from the virial envelope, and the U-shaped interior of the M is the lower half of an ellipse in the velocity-delay plane. The far-side response is weaker than that from the near side, so that we see clearly the lower half, but only faintly the upper half, of the velocity--delay ellipse. The delay tau=(R/c)(1-sin(i))=5 light-days at line center is from the near edge of the inclined ring, giving the inclination i=45 deg. A black hole mass of M=7x10^7 Msun is consistent with the velocity-delay structure. A barber-pole pattern with stripes moving from red to blue across the CIV and possibly Ly_alpha line profiles suggests the presence of azimuthal structure rotating around the far side of the broad-line region and may be the signature of precession or orbital motion of structures in the inner disk. Further HST observations of NGC 5548 over a multi-year timespan but with a cadence of perhaps 10 days rather than 1 day could help to clarify the nature of this new AGN phenomenon.

astro-ph.GA↗

arXiv2020

Space Telescope and Optical Reverberation Mapping Project. XII. Broad-Line Region Modeling of NGC 5548

We present geometric and dynamical modeling of the broad line region for the multi-wavelength reverberation mapping campaign focused on NGC 5548 in 2014. The dataset includes photometric and spectroscopic monitoring in the optical and ultraviolet, covering the H$β$, C IV, and Ly$α$ broad emission lines. We find an extended disk-like H$β$ BLR with a mixture of near-circular and outflowing gas trajectories, while the C IV and Ly$α$ BLRs are much less extended and resemble shell-like structures. There is clear radial structure in the BLR, with C IV and Ly$α$ emission arising at smaller radii than the H$β$ emission. Using the three lines, we make three independent black hole mass measurements, all of which are consistent. Combining these results gives a joint inference of $\log_{10}(M_{\rm BH}/M_\odot) = 7.64^{+0.21}_{-0.18}$. We examine the effect of using the $V$ band instead of the UV continuum light curve on the results and find a size difference that is consistent with the measured UV-optical time lag, but the other structural and kinematic parameters remain unchanged, suggesting that the $V$ band is a suitable proxy for the ionizing continuum when exploring the BLR structure and kinematics. Finally, we compare the H$β$ results to similar models of data obtained in 2008 when the AGN was at a lower luminosity state. We find that the size of the emitting region increased during this time period, but the geometry and black hole mass remain unchanged, which confirms that the BLR kinematics suitably gauge the gravitational field of the central black hole.

astro-ph.GA↗

arXiv2020

Blind Inverse Gamma Correction with Maximized Differential Entropy

Unwanted nonlinear gamma distortion frequently occurs in a great diversity of images during the procedures of image acquisition, processing, and/or display. And the gamma distortion often varies with capture setup change and luminance variation. Blind inverse gamma correction, which automatically determines a proper restoration gamma value from a given image, is of paramount importance to attenuate the distortion. For blind inverse gamma correction, an adaptive gamma transformation method (AGT-ME) is proposed directly from a maximized differential entropy model. And the corresponding optimization has a mathematical concise closed-form solution, resulting in efficient implementation and accurate gamma restoration of AGT-ME. Considering the human eye has a non-linear perception sensitivity, a modified version AGT-ME-VISUAL is also proposed to achieve better visual performance. Tested on variable datasets, AGT-ME could obtain an accurate estimation of a large range of gamma distortion (0.1 to 3.0), outperforming the state-of-the-art methods. Besides, the proposed AGT-ME and AGT-ME-VISUAL were applied to three typical applications, including automatic gamma adjustment, natural/medical image contrast enhancement, and fringe projection profilometry image restoration. Furthermore, the AGT-ME/ AGT-ME-VISUAL is general and can be seamlessly extended to the masked image, multi-channel (color or spectrum) image, or multi-frame video, and free of the arbitrary tuning parameter. Besides, the corresponding Python code (https://github.com/yongleex/AGT-ME) is also provided for interested users.

eess.IV↗

arXiv2020

The Impact of Type Ia Supernovae in Quiescent Galaxies: I. Formation of the Multiphase Interstellar medium

A cool phase of the interstellar medium has been observed in many giant elliptical galaxies, but its origin remains unclear. We propose that uneven heating from Type Ia supernovae (SNe Ia), together with radiative cooling, can lead to the formation of the cool phase. The basic idea is that since SNe Ia explode randomly, gas parcels which are not directly heated by SN shocks will cool, forming multiphase gas. We run a series of idealized high-resolution numerical simulations, and find that cool gas develops even when the overall SNe heating rate $H$ exceeds the cooling rate $C$ by a factor as large as 1.4. We also find that the time for multiphase gas development depends on the gas temperature. When the medium has a temperature $T = 3\times 10^6$ K, the cool phase forms within one cooling time \tc; however, the cool phase formation is delayed to a few times \tc\ for higher temperatures. The main reason for the delay is turbulent mixing. Cool gas formed this way would naturally have a metallicity lower than that of the hot medium. For constant $H/C$, there is more turbulent mixing for higher temperature gas. We note that this mechanism of producing cool gas cannot be captured in cosmological simulations, which usually fail to resolve individual SN remnants.

astro-ph.GA↗

arXiv2020

The Impact of Type Ia Supernovae in Quiescent Galaxies: II. Energetics and Turbulence

Type Ia supernovae (SNe Ia) provide unique and important feedback in quiescent galaxies, but their impact has been underappreciated. In this paper, we analyze a series of high-resolution simulations to examine the energetics and turbulence of the medium under SNe Ia. We find that when SN remnants are resolved, their effects differ distinctly from a volumetric heating term, as is commonly assumed in unresolved simulations. First, the net heating is significantly higher than expected, by 30$\pm$10\% per cooling time. This is because a large fraction of the medium is pushed into lower densities which cool inefficiently. Second, the medium is turbulent; the root-mean-squared (RMS) velocity of the gas to 20-50 km s$^{-1}$ on a driving scale of tens of parsec. The velocity field of the medium is dominated by compressional modes, which are larger than the solenoidal components by a factor of 3-7. Third, the hot gas has a very broad density distribution. The ratio between the density fluctuations and the RMS Mach number, parameterized as $b$, is 2-20. This is in contrast to previous simulations of turbulent media, which have found $b\lesssim$ 1. The reason for the difference is mainly caused by the \textit{localized} heating of SNe Ia, which creates a large density contrast. Last, the typical length scale of a density fluctuation grows with time, forming increasingly larger bubbles and filamentary ridges. These underlying density fluctuations need to be included when X-ray observations are interpreted.

astro-ph.GA↗

arXiv2020

First results from SMAUG: Uncovering the Origin of the Multiphase Circumgalactic Medium with a Comparative Analysis of Idealized and Cosmological Simulations

We examine the properties of the circumgalactic medium (CGM) at low redshift in a range of simulated Milky Way mass halos. The sample is comprised of seven idealized simulations, an adaptive mesh refinement cosmological zoom-in simulation, and two groups of 50 halos with star forming or quiescent galaxies taken from the IllustrisTNG100 simulation. The simulations have very different setups, resolution, and feedback models, but are analyzed in a uniform manner. By comparing median radial profiles and mass distributions of CGM properties, we isolate key similarities and differences. In doing so, we advance the efforts of the SMAUG (Simulating Multiscale Astrophysics to Understand Galaxies) project that aims to understand the inherently multiscale galaxy formation process. In the cosmological simulations, the CGM exhibits nearly flat temperature distributions, and broad pressure and radial velocity distributions. In the idealized simulations, similar distributions are found in the inner CGM ($\lesssim 0.5 \, r_{\rm 200c}$) when strong galactic feedback models are employed, but the outer CGM ($\gtrsim 0.5 \, r_{\rm 200c}$) has a much less prominent cold phase, and narrower pressure and velocity distributions even in models with strong feedback. This comparative analysis demonstrates the dominant role feedback plays in shaping the inner CGM and the increased importance of cosmological effects, such as nonspherical accretion and satellite galaxies, in the outer CGM. Furthermore, our findings highlight that while cosmological simulations are required to capture the multiphase structure of the CGM at large radii, idealized simulations provide a robust framework to study how galactic feedback interacts with the inner CGM and thereby provide a reliable avenue to constrain feedback prescriptions.

astro-ph.GA↗

arXiv2020

How Do Supernovae Impact the Circumgalactic Medium? I. Large-Scale Fountains Around a Milky Way-Like Galaxy

Feedback is indispensable in galaxy formation. However, lacking resolutions, cosmological simulations often use ad hoc feedback parameters. Conversely, small-box simulations, while better resolving the feedback, cannot capture gas evolution beyond the simulation domain. We aim to bridge the gap by implementing small-box results of supernovae-driven outflows into dark matter halo-scale simulations and studying their impact on large scales. Galactic outflows are multiphase, but small-box simulations show that the hot phase (T$\approx$ 10$^{6-7}$ K) carries the majority of energy and metals. We implement hot outflows in idealized simulations of the Milky Way halo, and examine how they impact the circumgalactic medium (CGM). In this paper, we discuss the case when the star formation surface density is low and therefore the emerging hot outflows are gravitationally bound by the halo. We find that outflows form a large-scale, metal-enriched atmosphere with fountain motions. As hot gas accumulates, the inner atmosphere becomes "saturated". Cool gas condenses, with a rate balancing the injection of the hot outflows. This balance leads to a universal density profile of the hot atmosphere, independent of mass outflow rate. The atmosphere has a radially-decreasing temperature, naturally producing the observed X-ray luminosity and column densities of O VI, O VII, O VIII. The self-regulated atmosphere has a baryon and a metal mass of $(0.5-1.2)\times 10^{10}M_\odot$ and $(0.6-1.4)\times 10^8 M_\odot$, respectively, small compared to the ``missing" baryons and metals from the halo. We conjecture that the missing materials reside at even larger radii, ejected by more powerful outflows in the past.

astro-ph.GA↗

arXiv2020

The SPPD System for Schema Guided Dialogue State Tracking Challenge

This paper introduces one of our group's work on the Dialog System Technology Challenges 8 (DSTC8), the SPPD system for Schema Guided dialogue state tracking challenge. This challenge, named as Track 4 in DSTC8, provides a brand new and challenging dataset for developing scalable multi-domain dialogue state tracking algorithms for real world dialogue systems. We propose a zero-shot dialogue state tracking system for this task. The key components of the system is a number of BERT based zero-shot NLU models that can effectively capture semantic relations between natural language descriptions of services' schemas and utterances from dialogue turns. We also propose some strategies to make the system better to exploit information from longer dialogue history and to overcome the slot carryover problem for multi-domain dialogues. The experimental results show that the proposed system achieves a significant improvement compared with the baseline system.

cs.CL↗

arXiv2020

Chiral Magnetic Effect in Isobar Collisions from Stochastic Hydrodynamics

We study chiral magnetic effect in collisions of AuAu, RuRu and ZrZr at s = 200GeV. The axial charge evolution is modeled with stochastic hydrodynamics and geometrical quantities are calculated with Monte Carlo Glauber model. By adjusting the relaxation time of magnetic field, we find our results in good agreement with background subtracted data for AuAu collisions at the same energy. We also make prediction for RuRu and ZrZr collisions. We find a weak centrality dependence of initial chiral imbalance, which implies the centrality dependence of chiral magnetic effect signal comes mainly from those of magnetic field and volume factor. Our results also show an unexpected dependence on system size: while the system of AuAu has larger chiral imbalance and magnetic field, it turns out to have smaller signal for chiral magnetic effect due to the larger volume suppression factor.

nucl-th↗

arXiv2020

Properties of the Simulated Circumgalactic Medium

The circumgalactic medium (CGM) is closely linked to galaxy formation and evolution, but difficult to characterize observationally and typically poorly resolved in cosmological simulations. We use spherically-symmetric, idealized, high-resolution simulations of the CGM in $10^{12}M_\odot$ and $10^{11}M_\odot$ dark matter halos to characterize the gas pressure, turbulent and radial velocities, and degree of thermal and effective dynamic pressure support in the overall CGM as well as in its high- and low-temperature phases. We find that the $10^{12}M_\odot$ halo contains a CGM mostly formed of a hot gas halo in hydrostatic equilibrium out of which cold gas condenses and falls onto the central galaxy, while the $10^{11}M_\odot$ halo's CGM is not in hydrostatic equilibrium, has a wider spread of properties at a given galactocentric radius, does not have a clear separation of hot and cold phases, and is dominated by bulk motions. We also find that the degree of pressure support in the $10^{11}M_\odot$ halo is strongly dependent on the parameters of the galactic winds of the central galaxy. These results promote the idea that there is no "average" CGM and care must be taken when setting the initial conditions for a small-box simulation of a patch of the CGM.

astro-ph.GA↗

arXiv2020

Simple Yet Powerful: Hot Galactic Outflows Driven by Supernovae

Supernovae (SNe) drive multiphase galactic outflows, impacting galaxy formation; however, cosmological simulations mostly use \textit{ad hoc} feedback models for outflows, making outflow-related predictions from first principles problematic. Recent small-box simulations resolve individual SNe remnants in the interstellar medium (ISM), naturally driving outflows and permitting a determination of the wind loading factors of energy $η_E$, mass $η_m$, and metals $η_Z$. In this Letter, we compile small-box results, and find consensus that the hot outflows are much more powerful than the cool outflows: (i) hot outflows generally dominate the energy flux, and (ii) their specific energy $e_{s,h}$ is 10-1000 times higher than cool outflows. Moreover, the properties of hot outflows are remarkably simple: $e_{s,h} \propto η_{E,h}/η_{m,h}$ is almost invariant over four orders of magnitude of star formation surface density. Also, we find tentatively that $η_{E,h}/η_{Z,h} \sim$ 0.5. If corroborated by more simulation data, these correlations reduce the three hot phase loading factors into one. Finally, this one parameter is closely related to whether the ISM has a "breakout" condition. The narrow range of \esh\ indicates that hot outflows cannot escape dark matter halos with log $M_{\rm{halo}}\ [M_\odot] \gtrsim 12$. This mass is also where the galaxy mass-metallicity relation reaches its plateau, implying a deep connection between \textit{hot} outflows and galaxy formation. We argue that hot outflows should be included explicitly in cosmological simulations and (semi-)analytic modeling of galaxy formation.

astro-ph.GA↗

arXiv2019

Kinematics and Dynamics of Multiphase Outflows in Simulations of the Star-Forming Galactic ISM

Galactic outflows produced by stellar feedback are known to be multiphase in nature. Both observations and simulations indicate that the material within several kpc of galactic disk mid-planes consists of warm clouds embedded within a hot wind. A theoretical understanding of the outflow phenomenon, including both winds and fountain flows, requires study of the interactions among thermal phases. We develop a method to quantify these interactions via measurements of mass, momentum, and energy flux exchanges using temporally and spatially averaged quantities and conservation laws. We apply this method to a star-forming ISM MHD simulation based on the TIGRESS framework, for Solar neighbourhood conditions. To evaluate the extent of interactions among the phases, we first examine the validity of the ``ballistic model,'' which predicts trajectories of the warm phase ($5050\,\rm{K} 5\times10^5\,\rm{K}$) to the warm phase. The large energy flux from the hot outflow that is transferred to the warm and intermediate phases is quickly radiated away. A simple interaction model implies an effective warm cloud size in the fountain flow of a few 100~pc, showing that warm-hot flux exchange mainly involves a few large clouds rather than many small ones.

astro-ph.GA↗

arXiv2019

Direct Detection of Black Hole-Driven Turbulence in the Centers of Galaxy Clusters

Supermassive black holes (SMBHs) are thought to provide energy that prevents catastrophic cooling in the centers of massive galaxies and galaxy clusters. However, it remains unclear how this "feedback" process operates. We use high-resolution optical data to study the kinematics of multi-phase filamentary structures by measuring the velocity structure function (VSF) of the filaments over a wide range of scales in the centers of three nearby galaxy clusters: Perseus, Abell 2597 and Virgo. We find that the motions of the filaments are turbulent in all three clusters studied. There is a clear correlation between features of the VSFs and the sizes of bubbles inflated by SMBH driven jets. Our study demonstrates that SMBHs are the main driver of turbulent gas motions in the centers of galaxy clusters and suggests that this turbulence is an important channel for coupling feedback to the environment. Our measured amplitude of turbulence is in good agreement with Hitomi Doppler line broadening measurement and X-ray surface brightness fluctuation analysis, suggesting that the motion of the cold filaments is well-coupled to that of the hot gas. The smallest scales we probe are comparable to the mean free path in the intracluster medium (ICM). Our direct detection of turbulence on these scales provides the clearest evidence to date that isotropic viscosity is suppressed in the weakly-collisional, magnetized intracluster plasma.

astro-ph.GA↗

arXiv2019

Human-Like Hybrid Caching in Software-defined Edge Cloud

With the development of Internet of Things (IoT) and communication technology, the number of next-generation IoT devices has increased explosively, and the delay requirement for content requests is becoming progressively higher. Fortunately, the edge-caching scheme can satisfy users' demands for low latency of content. However, the existing caching schemes are not smart enough. To address these challenges, we propose a human-like hybrid caching architecture based on the software defined edge cloud, which simultaneously considers the content popularity and the fine-grained user characteristics. Then, an optimization problem with a caching hit ratio as an optimization objective is formulated. To solve this problem, using reinforcement learning, we design a human-like hybrid caching algorithm. Extensive experiments show that compared with popular caching schemes, human-like hybrid caching schemes can improve the cache hit ratio by 20%.

cs.NI↗

arXiv2019

Cosmological constraints from the redshift dependence of the Alcock-Paczynski effect: Fourier space analysis

The tomographic Alcock-Paczynski (AP) method utilizes the redshift evolution of the AP distortion to place constraints on cosmological parameters. It has proved to be a robust method that can separate the AP signature from the redshift space distortion (RSD) effect, and deliver powerful cosmological constraints using the $\lesssim 40h^{-1}\ \rm Mpc$ clustering region. In previous works, the tomographic AP method was performed via the anisotropic 2-point correlation function statistic. In this work we consider the feasibility of conducting the analysis in the Fourier domain and examine the pros and cons of this approach. We use the integrated galaxy power spectrum (PS) as a function of direction, $\hat P_{Δk}(μ)$, to quantify the magnitude of anisotropy in the large-scale structure clustering, and use its redshift variation to do the AP test. The method is tested on the large, high resolution Big-MultiDark Planck (BigMD) simulation at redshifts $z=0-1$, using the underlying true cosmology $Ω_m=0.3071,\ w=-1$. Testing the redshift evolution of $\hat P_{Δk}(μ)$ in the true cosmology and cosmologies deviating from the truth with $δΩ_m=0.1,\ δw=0.3$, we find that the redshift evolution of the AP distortion overwhelms the effects created by the RSD by a factor of $\sim1.7-3.6$. We test the method in the range of $k\in(0.2,1.8)\ h\ \rm Mpc^{-1}$, and find that it works well throughout the entire regime. We tune the halo mass within the range $2\times 10^{13}$ to $10^{14}\ M_{\odot}$, and find that the change of halo bias results in $\lesssim 5 \%$ change in $\hat P_{Δk}(μ)$, which is less significant compared with the cosmological effect. Our work shows that it is feasible to conduct the tomographic AP analysis in the Fourier space.

astro-ph.CO↗