From large-scale environment to CGM angular momentum to star forming activities -- I: star-forming galaxies

The connection between halo gas acquisition through the circumgalactic medium (CGM) and galaxy star formation has long been studied. In this series of two papers, we put this interplay within the context of the galaxy environment on large scales (several hundreds of kpc), which, to a certain degree, maps out various paths for galaxy interactions. We use the IllustrisTNG-100 simulation to demonstrate that the large-scale environment modulates the circumgalactic gas angular momentum, resulting in either enhanced (Paper I) or suppressed (Paper II) star formation inside a galaxy. In this paper (Paper I), we show that the large-scale environment around a star-forming galaxy is often responsible for triggering new episodes of star formation. Such an episodic star formation pattern is well synced with a pulsating motion of the circumgalactic gas, which, on the one hand receives angular momentum modulations from the large-scale environment, yielding in-spiralling gas to fuel the star-forming reservoir, while, on the other hand, is affected by the feedback activities from the galaxy centre. As a result, a present-day star-forming galaxy may have gone through several cycles of star-forming and quiescent phases during its evolutionary history, with the circumgalactic gas carrying out a synchronized cadence of breathing in and out motions out to $sim 100$ kpc.

From large-scale environment to CGM angular momentum to star forming activities -- II: quenched galaxies

The gas needed to sustain star formation in galaxies is supplied by the circumgalactic medium (CGM), which in turn is affected by accretion from large scales. In a series of two papers, we examine the interplay between a galaxys ambient CGM and central star formation within the context of the large-scale environment. We use the IllustrisTNG-100 simulation to show that the influence exerted by the large-scale galaxy environment on the CGM gas angular momentum results in either enhanced (Paper I) or suppressed (Paper II, this paper) star formation inside a galaxy. We find that for present-day quenched galaxies, both the large-scale environments and the ambient CGM have always had higher angular momenta throughout their evolutionary history since at least $z=2$, in comparison to those around present-day star-forming disk galaxies, resulting in less efficient gas inflow into the central star-forming gas reservoirs. A sufficiently high CGM angular momentum, as inherited from the larger-scale environment, is thus an important factor in keeping a galaxy quenched, once it is quenched. The process above naturally renders two key observational signatures: (1) a coherent rotation pattern existing across multiple distances from the large-scale galaxy environment, to the circumgalactic gas, to the central stellar disk; and (2) an anti-correlation between galaxy star-formation rates and orbital angular momenta of interacting galaxy pairs or groups.

Randomized Substitution and Vote for Textual Adversarial Example Detection

A line of work has shown that natural text processing models are vulnerable to adversarial examples. Correspondingly, various defense methods are proposed to mitigate the threat of textual adversarial examples, e.g. adversarial training, certified defense, input pre-processing, detection, etc. In this work, we treat the optimization process for synonym substitution based textual adversarial attacks as a specific sequence of word replacement, in which each word mutually influences other words. We identify that we could destroy such mutual interaction and eliminate the adversarial perturbation by randomly substituting a word with its synonyms. Based on this observation, we propose a novel textual adversarial example detection method, termed Randomized Substitution and Vote (RS&V), which votes the prediction label by accumulating the logits of k samples generated by randomly substituting the words in the input text with synonyms. The proposed RS&V is generally applicable to any existing neural networks without modification on the architecture or extra training, and it is orthogonal to prior work on making the classification network itself more robust. Empirical evaluations on three benchmark datasets demonstrate that RS&V could detect the textual adversarial examples more successfully than the existing detection methods while maintaining the high classification accuracy on benign samples.

Achieving Deterministic Service in Mobile Edge Computing (MEC) Networks

Mobile edge computing (MEC) is proposed to boost high-efficient and time-sensitive 5G applications. However, the microburst may occur even in lightly-loaded scenarios, which leads to the indeterministic service latency (i.e., unpredictable delay or delay variation), hence hindering the deployment of MEC. Deterministic IP networking (DIP) has been proposed that can provide bounds on latency, and high reliability in the large-scale networks. Nevertheless, the direct migration of DIP into the MEC network is non-trivial owing to its original design for the Ethernet with homogeneous devices. Meanwhile, DIP also faces the challenges on the network throughput and scheduling flexibility. In this paper, we delve into the adoption of DIP for the MEC networks and some of the relevant aspects. A deterministic MEC (D-MEC) network is proposed to deliver the deterministic service (i.e., providing the MEC service with bounded service latency). In the D-MEC network, two mechanisms, including the cycle mapping and cycle shifting, are designed to enable: (i) seamless and deterministic transmission with heterogeneous underlaid resources; and (ii) traffic shaping on the edges to improve the resource utilization. We also formulate a joint configuration to maximize the network throughput with deterministic QoS guarantees. Extensive simulations verify that the proposed D-MEC network can achieve a deterministic MEC service, even in the highly-loaded scenarios.

Localize, Group, and Select: Boosting Text-VQA by Scene Text Modeling

As an important task in multimodal context understanding, Text-VQA (Visual Question Answering) aims at question answering through reading text information in images. It differentiates from the original VQA task as Text-VQA requires large amounts of scene-text relationship understanding, in addition to the cross-modal grounding capability. In this paper, we propose Localize, Group, and Select (LOGOS), a novel model which attempts to tackle this problem from multiple aspects. LOGOS leverages two grounding tasks to better localize the key information of the image, utilizes scene text clustering to group individual OCR tokens, and learns to select the best answer from different sources of OCR (Optical Character Recognition) texts. Experiments show that LOGOS outperforms previous state-of-the-art methods on two Text-VQA benchmarks without using additional OCR annotation data. Ablation studies and analysis demonstrate the capability of LOGOS to bridge different modalities and better understand scene text.

Human Pose and Shape Estimation from Single Polarization Images

This paper focuses on a new problem of estimating human pose and shape from single polarization images. Polarization camera is known to be able to capture the polarization of reflected lights that preserves rich geometric cues of an object surface. Inspired by the recent applications in surface normal reconstruction from polarization images, in this paper, we attempt to estimate human pose and shape from single polarization images by leveraging the polarization-induced geometric cues. A dedicated two-stage pipeline is proposed: given a single polarization image, stage one (Polar2Normal) focuses on the fine detailed human body surface normal estimation; stage two (Polar2Shape) then reconstructs clothed human shape from the polarization image and the estimated surface normal. To empirically validate our approach, a dedicated dataset (PHSPD) is constructed, consisting of over 500K frames with accurate pose and shape annotations. Empirical evaluations on this real-world dataset as well as a synthetic dataset, SURREAL, demonstrate the effectiveness of our approach. It suggests polarization camera as a promising alternative to the more conventional RGB camera for human pose and shape estimation.

Object Wake-up: 3-D Object Reconstruction, Animation, and in-situ Rendering from a Single Image

Given a picture of a chair, could we extract the 3-D shape of the chair, animate its plausible articulations and motions, and render in-situ in its original image space? The above question prompts us to devise an automated approach to extract and manipulate articulated objects in single images. Comparing with previous efforts on object manipulation, our work goes beyond 2-D manipulation and focuses on articulable objects, thus introduces greater flexibility for possible object deformations. The pipeline of our approach starts by reconstructing and refining a 3-D mesh representation of the object of interest from an input image; its control joints are predicted by exploiting the semantic part segmentation information; the obtained object 3-D mesh is then rigged & animated by non-rigid deformation, and rendered to perform in-situ motions in its original image space. Quantitative evaluations are carried out on 3-D reconstruction from single images, an established task that is related to our pipeline, where our results surpass those of the SOTAs by a noticeable margin. Extensive visual results also demonstrate the applicability of our approach.

Unsupervised 3D Human Mesh Recovery from Noisy Point Clouds

This paper presents a novel unsupervised approach to reconstruct human shape and pose from noisy point cloud. Traditional approaches search for correspondences and conduct model fitting iteratively where a good initialization is critical. Relying on large amount of dataset with ground-truth annotations, recent learning-based approaches predict correspondences for every vertice on the point cloud; Chamfer distance is usually used to minimize the distance between a deformed template model and the input point cloud. However, Chamfer distance is quite sensitive to noise and outliers, thus could be unreliable to assign correspondences. To address these issues, we model the probability distribution of the input point cloud as generated from a parametric human model under a Gaussian Mixture Model. Instead of explicitly aligning correspondences, we treat the process of correspondence search as an implicit probabilistic association by updating the posterior probability of the template model given the input. A novel unsupervised loss is further derived that penalizes the discrepancy between the deformed template and the input point cloud conditioned on the posterior probability. Our approach is very flexible, which works with both complete point cloud and incomplete ones including even a single depth image as input. Our network is trained from scratch with no need to warm-up the network with supervised data. Compared to previous unsupervised methods, our method shows the capability to deal with substantial noise and outliers. Extensive experiments conducted on various public synthetic datasets as well as a very noisy real dataset (i.e. CMU Panoptic) demonstrate the superior performance of our approach over the state-of-the-art methods. Code can be found url{https://github.com/wangsen1312/unsupervised3dhuman.git}

Quenched, bulge-dominated, but dynamically cold galaxies in IllustrisTNG and their real-world counterparts

Galaxy morphologies, kinematics, and stellar populations are thought to link to each other. However, both simulations and observations have pointed out mismatches therein. In this work, we study the nature and origin of the present-day quenched, bulge-dominated, but dynamically cold galaxies within a stellar mass range of $10.3<log,M_{ast}/mathrm{M_{odot}}<11.2$ in the IllustrisTNG-100 Simulation, as a companion paper of Lu et al.(2021), which aimed at the star-forming but dynamically hot disc galaxies within a lower stellar mass range of $9.7<log,M_{ast}/mathrm{M_{odot}}<10.3$. We compare cold quenched population with a population of normal star-forming dynamically cold disc galaxies and a population of normal quenched dynamically hot elliptical galaxies within the same mass range. The populations of the present-day quenched and bulge-dominated galaxies (both being dynamically cold and hot) used to have significantly higher star-formation rates and thinner morphologies at redshifts of z~2. They have experienced more frequent larger mass-ratio mergers below z~0.7 in comparison to their star-forming disc counterparts, which is responsible for the formation of their bulge-dominated morphologies. The dynamically cold populations (both being star-forming and quenched) have experienced more frequent prograde and tangential mergers especially below z~1, in contrast to the dynamically hot ellipticals, which have had more retrograde and radial mergers. Such different merging histories can well explain the differences on the cold and hot dynamical status among these galaxies. We point out that the real-world counterparts of these dynamically cold and hot bulge-dominated quenched populations are the fast- and slow-rotating early-type galaxies, respectively, as seen in observations and hence reveal the different evolution paths of these two distinct populations of early-type galaxies.

A Multi-task Mean Teacher for Semi-supervised Facial Affective Behavior Analysis

Affective Behavior Analysis is an important part in human-computer interaction. Existing multi-task affective behavior recognition methods suffer from the problem of incomplete labeled datasets. To tackle this problem, this paper presents a semi-supervised model with a mean teacher framework to leverage additional unlabeled data. To be specific, a multi-task model is proposed to learn three different kinds of facial affective representations simultaneously. After that, the proposed model is assigned to be student and teacher networks. When training with unlabeled data, the teacher network is employed to predict pseudo labels for student network training, which allows it to learn from unlabeled data. Experimental results showed that our proposed method achieved much better performance than baseline model and ranked 4th in both competition track 1 and track 2, and 6th in track 3, which verifies that the proposed network can effectively learn from incomplete datasets.