arXiv subjects

Tao Yang

Publications and source records attributed to Tao Yang.

At least 289 records · Page 16Linked to original sources

arXiv2021

Rethinking Content and Style: Exploring Bias for Unsupervised Disentanglement

Content and style (C-S) disentanglement intends to decompose the underlying explanatory factors of objects into two independent subspaces. From the unsupervised disentanglement perspective, we rethink content and style and propose a formulation for unsupervised C-S disentanglement based on our assumption that different factors are of different importance and popularity for image reconstruction, which serves as a data bias. The corresponding model inductive bias is introduced by our proposed C-S disentanglement Module (C-S DisMo), which assigns different and independent roles to content and style when approximating the real data distributions. Specifically, each content embedding from the dataset, which encodes the most dominant factors for image reconstruction, is assumed to be sampled from a shared distribution across the dataset. The style embedding for a particular image, encoding the remaining factors, is used to customize the shared distribution through an affine transformation. The experiments on several popular datasets demonstrate that our method achieves the state-of-the-art unsupervised C-S disentanglement, which is comparable or even better than supervised methods. We verify the effectiveness of our method by downstream tasks: domain translation and single-view 3D reconstruction. Project page at https://github.com/xrenaa/CS-DisMo.

cs.CV↗

arXiv2021

Safety-aware Motion Prediction with Unseen Vehicles for Autonomous Driving

Motion prediction of vehicles is critical but challenging due to the uncertainties in complex environments and the limited visibility caused by occlusions and limited sensor ranges. In this paper, we study a new task, safety-aware motion prediction with unseen vehicles for autonomous driving. Unlike the existing trajectory prediction task for seen vehicles, we aim at predicting an occupancy map that indicates the earliest time when each location can be occupied by either seen and unseen vehicles. The ability to predict unseen vehicles is critical for safety in autonomous driving. To tackle this challenging task, we propose a safety-aware deep learning model with three new loss functions to predict the earliest occupancy map. Experiments on the large-scale autonomous driving nuScenes dataset show that our proposed model significantly outperforms the state-of-the-art baselines on the safety-aware motion prediction task. To the best of our knowledge, our approach is the first one that can predict the existence of unseen vehicles in most cases. Project page at {\url{https://github.com/xrenaa/Safety-Aware-Motion-Prediction}}.

cs.CV↗

arXiv2021

Linear Convergence of First- and Zeroth-Order Primal-Dual Algorithms for Distributed Nonconvex Optimization

This paper considers the distributed nonconvex optimization problem of minimizing a global cost function formed by a sum of local cost functions by using local information exchange. We first consider a distributed first-order primal-dual algorithm. We show that it converges sublinearly to a stationary point if each local cost function is smooth and linearly to a global optimum under an additional condition that the global cost function satisfies the Polyak-Łojasiewicz condition. This condition is weaker than strong convexity, which is a standard condition for proving linear convergence of distributed optimization algorithms, and the global minimizer is not necessarily unique. Motivated by the situations where the gradients are unavailable, we then propose a distributed zeroth-order algorithm, derived from the considered first-order algorithm by using a deterministic gradient estimator, and show that it has the same convergence properties as the considered first-order algorithm under the same conditions. The theoretical results are illustrated by numerical simulations.

math.OC↗

arXiv2021

ULTRA: An Unbiased Learning To Rank Algorithm Toolbox

Learning to rank systems has become an important aspect of our daily life. However, the implicit user feedback that is used to train many learning to rank models is usually noisy and suffered from user bias (i.e., position bias). Thus, obtaining an unbiased model using biased feedback has become an important research field for IR. Existing studies on unbiased learning to rank (ULTR) can be generalized into two families-algorithms that attain unbiasedness with logged data, offline learning, and algorithms that achieve unbiasedness by estimating unbiased parameters with real-time user interactions, namely online learning. While there exist many algorithms from both families, there lacks a unified way to compare and benchmark them. As a result, it can be challenging for researchers to choose the right technique for their problems or for people who are new to the field to learn and understand existing algorithms. To solve this problem, we introduced ULTRA, which is a flexible, extensible, and easily configure ULTR toolbox. Its key features include support for multiple ULTR algorithms with configurable hyperparameters, a variety of built-in click models that can be used separately to simulate clicks, different ranking model architecture and evaluation metrics, and simple learning to rank pipeline creation. In this paper, we discuss the general framework of ULTR, briefly describe the algorithms in ULTRA, detailed the structure, and pipeline of the toolbox. We experimented on all the algorithms supported by ultra and showed that the toolbox performance is reasonable. Our toolbox is an important resource for researchers to conduct experiments on ULTR algorithms with different configurations as well as testing their own algorithms with the supported features.

cs.IR↗

arXiv2021

Acoustically Manipulating Internal Structure of Disk-in-Sphere Endoskeletal Droplets

Manipulation of micro/nano particles has been well studied and demonstrated by optical, electromagnetic, and acoustic approaches, or their combinations. Manipulation of internal structure of droplet/particle is rarely explored and remains challenging due to its complicated nature. Here we demonstrated the manipulation of internal structure of disk-in-sphere endoskeletal droplets using acoustic wave for the first time. We developed a model to investigate the physical mechanisms behind this novel phenomenon. Theoretical analysis of the acoustic interactions indicated that these assembly dynamics arise from a balance of the primary and secondary radiation forces. Additionally, the disk orientation was found to change with acoustic driving frequency, which allowed on-demand, reversible adjusting disk orientations with respect to the substrate. This novel dynamic behavior leads to unique reversible arrangements of the endoskeletal droplets and their internal architecture, which may provide a new avenue for directed assembly of novel hierarchical colloidal architectures and intracellular organelles or intra-organoid structures.

physics.app-ph↗

arXiv2021

Efficient Online Transfer Learning for 3D Object Classification in Autonomous Driving

Autonomous driving has achieved rapid development over the last few decades, including the machine perception as an important issue of it. Although object detection based on conventional cameras has achieved remarkable results in 2D/3D, non-visual sensors such as 3D LiDAR still have incomparable advantages in the accuracy of object position detection. However, the challenge also exists with the difficulty in properly interpreting point cloud generated by LiDAR. This paper presents a multi-modal-based online learning system for 3D LiDAR-based object classification in urban environments, including cars, cyclists and pedestrians. The proposed system aims to effectively transfer the mature detection capabilities based on visual sensors to the new model learning based on non-visual sensors through a multi-target tracker (i.e. using one sensor to train another). In particular, it integrates the Online Random Forests (ORF) [1] method, which inherently has the abilities of fast and multi-class learning. Through experiments, we show that our system is capable of learning a high-performance model for LiDAR-based 3D object classification on-the-fly, which is especially suitable for robotics in-situ deployment while responding to the widespread challenge of insufficient detector generalization capabilities.

cs.RO↗

arXiv2021

The performance of IHEP-NDL LGAD sensors after neutron irradiation

The performances of Low Gain Avalanche diode (LGAD) sensors from a neutron irradiation campaign with fluences of 0.8 x 10^15, 15 x 10^15 and 2.5 x 10^15 neq/cm2 are reported in this article. These LGAD sensors are developed by the Institute of High Energy Physics, Chinese Academy of Sciences and the Novel Device Laboratory for the High Granularity Timing Detector of the High Luminosity Large Hadron Collider. The timing resolution and collected charge of the LGAD sensors were measured with electrons from a beta source. After irradiation with a fluence of 2.5 x 10^15 neq/cm2, the collected charge decreases from 40 fC to 7 fC, the signal-to-noise ratio deteriorates from 48 to 12, and the timing resolution increases from 29 ps to 39 ps.

physics.ins-det↗

arXiv2021

Psycholinguistic Tripartite Graph Network for Personality Detection

Most of the recent work on personality detection from online posts adopts multifarious deep neural networks to represent the posts and builds predictive models in a data-driven manner, without the exploitation of psycholinguistic knowledge that may unveil the connections between one's language usage and his psychological traits. In this paper, we propose a psycholinguistic knowledge-based tripartite graph network, TrigNet, which consists of a tripartite graph network and a BERT-based graph initializer. The graph network injects structural psycholinguistic knowledge from LIWC, a computerized instrument for psycholinguistic analysis, by constructing a heterogeneous tripartite graph. The graph initializer is employed to provide initial embeddings for the graph nodes. To reduce the computational cost in graph learning, we further propose a novel flow graph attention network (GAT) that only transmits messages between neighboring parties in the tripartite graph. Benefiting from the tripartite graph, TrigNet can aggregate post information from a psychological perspective, which is a novel way of exploiting domain knowledge. Extensive experiments on two datasets show that TrigNet outperforms the existing state-of-art model by 3.47 and 2.10 points in average F1. Moreover, the flow GAT reduces the FLOPS and Memory measures by 38% and 32%, respectively, in comparison to the original GAT in our setting.

cs.CL↗

arXiv2021

Regret and Cumulative Constraint Violation Analysis for Online Convex Optimization with Long Term Constraints

This paper considers online convex optimization with long term constraints, where constraints can be violated in intermediate rounds, but need to be satisfied in the long run. The cumulative constraint violation is used as the metric to measure constraint violations, which excludes the situation that strictly feasible constraints can compensate the effects of violated constraints. A novel algorithm is first proposed and it achieves an $\mathcal{O}(T^{\max\{c,1-c\}})$ bound for static regret and an $\mathcal{O}(T^{(1-c)/2})$ bound for cumulative constraint violation, where $c\in(0,1)$ is a user-defined trade-off parameter, and thus has improved performance compared with existing results. Both static regret and cumulative constraint violation bounds are reduced to $\mathcal{O}(\log(T))$ when the loss functions are strongly convex, which also improves existing results. %In order to bound the regret with respect to any comparator sequence, In order to achieve the optimal regret with respect to any comparator sequence, another algorithm is then proposed and it achieves the optimal $\mathcal{O}(\sqrt{T(1+P_T)})$ regret and an $\mathcal{O}(\sqrt{T})$ cumulative constraint violation, where $P_T$ is the path-length of the comparator sequence. Finally, numerical simulations are provided to illustrate the effectiveness of the theoretical results.

cs.LG↗

arXiv2021

On Problems with Cosmography in Cosmic Dark Ages

Quasars show considerable promise as standard candles in a high-redshift window beyond Type Ia supernovae. Recently, Risaliti, Lusso \& collaborators \cite{Risaliti:2018reu, Lusso:2019akb, Lusso:2020pdb} have succeeded in producing a high redshift Hubble diagram ($ z \lesssim 7$) that supports "a trend whereby the Hubble diagram of quasars is well reproduced by the standard flat $Λ$CDM model up to $z \sim 1.5-2$, but strong deviations emerge at higher redshifts". This conclusion hinges upon a log polynomial expansion for the luminosity distance. In this note, we demonstrate that this expansion (or "improvements" thereof) typically can only be trusted up to $z \sim 1.5-2$. As a result, a breakdown in the validity of the expansion may be misinterpreted as a (phantom) deviation from flat $Λ$CDM. We further illustrate the problem through mock data examples.

astro-ph.CO↗

arXiv2021

Gravitational-Wave Detector Networks: Standard Sirens on Cosmology and Modified Gravity Theory

We construct the catalogues of standard sirens (StS) based on the future gravitational wave (GW) detector networks, i.e., the second-generation ground-based advanced LIGO+advanced Virgo+KAGRA+LIGO-India (HLVKI), the third-generation ground-based Einstein Telescope+two Cosmic Explorer (ET+2CE), and the space-based LISA+Taiji. From the corresponding electromagnetic (EM) counterpart detectors for each networks, we sample the joint GW+EM detections from the probability to construct the Hubble diagram of standard sirens for 10 years detections of HLVKI, 5 years detections of ET+2CE, and 5 years of detections of LISA+Taiji, which we estimate would be available and released in the 2030s. Thus we construct a combined Hubble diagram from these ground and spaced-based detector networks to explore the expansion history of our Universe from redshift 0 to 7. We give a conservative and realistic estimation of the catalogue and Hubble diagram of GW standard sirens and their potential on studying cosmology and modified gravity theory in the 2030s. We adopt two strategies for the forecasts. One is the traditional model-fitting Markov-Chain Monte-Carlo method (MCMC). The results show that the combined StS alone can constrain the Hubble constant at the precision level of $0.34\%$, 1.76 times more tightly than the current most precise measurement from \textit{Planck}+BAO+Pantheon. The joint StS with current EM experiments will improve the constraints of cosmological parameters significantly. The modified gravity theory can be constrained with $0.46\%$ error from the GW propagation. In the second strategy, we use the machine-learning nonparametric reconstruction techniques, i.e., the Gaussian process (GP) with the Artificial Neural Networks (ANN) as a comparison. GP reconstructions can give comparable results with MCMC. We anticipate more works and research on these topics.

astro-ph.CO↗

arXiv2021

GAN Prior Embedded Network for Blind Face Restoration in the Wild

Blind face restoration (BFR) from severely degraded face images in the wild is a very challenging problem. Due to the high illness of the problem and the complex unknown degradation, directly training a deep neural network (DNN) usually cannot lead to acceptable results. Existing generative adversarial network (GAN) based methods can produce better results but tend to generate over-smoothed restorations. In this work, we propose a new method by first learning a GAN for high-quality face image generation and embedding it into a U-shaped DNN as a prior decoder, then fine-tuning the GAN prior embedded DNN with a set of synthesized low-quality face images. The GAN blocks are designed to ensure that the latent code and noise input to the GAN can be respectively generated from the deep and shallow features of the DNN, controlling the global face structure, local face details and background of the reconstructed image. The proposed GAN prior embedded network (GPEN) is easy-to-implement, and it can generate visually photo-realistic results. Our experiments demonstrated that the proposed GPEN achieves significantly superior results to state-of-the-art BFR methods both quantitatively and qualitatively, especially for the restoration of severely degraded face images in the wild. The source code and models can be found at https://github.com/yangxy/GPEN.

cs.CV↗

arXiv2021

Running Hubble Tension and a H0 Diagnostic

Hubble tension is routinely presented as a mismatch between the Hubble constant $H_0$ determined locally and a value inferred from the flat $Λ$CDM cosmology. In essence, the tension boils down to a disagreement between two numbers. Here, assuming the tension is cosmological in origin, we predict that within flat $Λ$CDM there should be other inferred values of $H_0$, and that a "running of $H_0$ with redshift" can be expected. These additional determinations of $H_0$ may be traced to a difference between the effective equation of state (EoS) of the Universe within the Friedmann-Lemaître-Robertson-Walker (FLRW) cosmology framework and the current standard model. We introduce a diagnostic that flags such a running of $H_0$.

astro-ph.CO↗

arXiv2021

Hubble Sinks In The Low-Redshift Swampland

Local determinations of the Hubble constant $H_0$ favour a higher value than Planck based on CMB and $Λ$CDM. Through a model-independent expansion, we show that low redshift ($z \lesssim 0.7$) data comprising baryon acoustic oscillations (BAO), cosmic chronometers and Type Ia supernovae has a preference for Quintessence models that lower $H_0$ relative to $Λ$CDM. In addition, we confirm that an exponential coupling to dark matter cannot alter this conclusion in the same redshift range. Our results leave open the possibility that a coupling in the matter-dominated epoch, potentially even in the dark ages, may yet save $H_0$ from sinking in the string theory Swampland.

astro-ph.CO↗

arXiv2021

On Liquid Viscosity Effects on Droplet Splash and Receding Breakup on a Smooth Solid Surface at Atmospheric Pressure

Experimental and modelling study is presented for the effect of a wide range of liquid viscosities on the droplet impact on a smooth solid surface at atmospheric pressure. A non-monotonic variation of threshold between droplet deposition and splash was observed experimentally. Specifically, the critical Weber number separating deposition from splashing decreases and then increases with increasing the Ohnesorge number. The observations of splash in low viscosity region and receding breakup in high viscosity region were analyzed qualitatively from the perspectives of Kelvin-Helmholtz instability and Rayleigh-Taylor instability, respectively. Based on instability analysis for the viscosityinduced nonmonotonicity, a new semi-empirical correlation of droplet splashing thresholds is proposed by fitting experimental results from previous and present data and shows better performance than previous correlation formulas.

physics.flu-dyn↗

arXiv2021

Normalized solutions to a class of Kirchhoff equations with Sobolev critical exponent

In this paper, we consider the existence and asymptotic properties of solutions to the following Kirchhoff equation \begin{equation}\label{1}\nonumber - \Bigl(a+b\int_{\R^3} {{{\left| {\nabla u} \right|}^2}}\Bigl) Δu =λu+ {| u |^{p - 2}}u+μ{| u |^{q - 2}}u \text { in } \mathbb{R}^{3} \end{equation} under the normalized constraint $\int_{{\mathbb{R}^3}} {{u}^2}=c^2$, where $a\!>\!0$, $b\!>\!0$, $c\!>\!0$, $2\!<\!q\!<\!\frac{14}{3}\!<\! p\!\leq\!6$ or $\frac{14}{3}\!<\!q\!<\! p\!\leq\! 6$, $μ\!>\!0$ and $λ\!\in\!\R$ appears as a Lagrange multiplier. In both cases for the range of $p$ and $q$, the Sobolev critical exponent $p\!=\!6$ is involved and the corresponding energy functional is unbounded from below on $S_c=\Big\{ u \in H^{1}({\mathbb{R}^3}): \int_{{\mathbb{R}^3}} {{u}^2}=c^2 \Big\}$. If $2\!<\!q\!<\!\frac{10}{3}$ and $\frac{14}{3}\!<\! p\!<\!6$, we obtain a multiplicity result to the equation. If $2\!<\!q\!<\!\frac{10}{3}\!<\! p\!=\!6$ or $\frac{14}{3}\!<\!q\!<\! p\!\leq\! 6$, we get a ground state solution to the equation. Furthermore, we derive several asymptotic results on the obtained normalized solutions. Our results extend the results of N. Soave (J. Differential Equations 2020 $\&$ J. Funct. Anal. 2020), which studied the nonlinear Schrödinger equations with combined nonlinearities, to the Kirchhoff equations. To deal with the special difficulties created by the nonlocal term $({\int_{\R^3} {\left| {\nabla u} \right|} ^2}) Δu$ appearing in Kirchhoff type equations, we develop a perturbed Pohozaev constraint approach and we find a way to get a clear picture of the profile of the fiber map via careful analysis. In the meantime, we need some subtle energy estimates under the $L^2$-constraint to recover compactness in the Sobolev critical case.

math.AP↗

arXiv2021

SME: ReRAM-based Sparse-Multiplication-Engine to Squeeze-Out Bit Sparsity of Neural Network

Resistive Random-Access-Memory (ReRAM) crossbar is a promising technique for deep neural network (DNN) accelerators, thanks to its in-memory and in-situ analog computing abilities for Vector-Matrix Multiplication-and-Accumulations (VMMs). However, it is challenging for crossbar architecture to exploit the sparsity in the DNN. It inevitably causes complex and costly control to exploit fine-grained sparsity due to the limitation of tightly-coupled crossbar structure. As the countermeasure, we developed a novel ReRAM-based DNN accelerator, named Sparse-Multiplication-Engine (SME), based on a hardware and software co-design framework. First, we orchestrate the bit-sparse pattern to increase the density of bit-sparsity based on existing quantization methods. Second, we propose a novel weigh mapping mechanism to slice the bits of a weight across the crossbars and splice the activation results in peripheral circuits. This mechanism can decouple the tightly-coupled crossbar structure and cumulate the sparsity in the crossbar. Finally, a superior squeeze-out scheme empties the crossbars mapped with highly-sparse non-zeros from the previous two steps. We design the SME architecture and discuss its use for other quantization methods and different ReRAM cell technologies. Compared with prior state-of-the-art designs, the SME shrinks the use of crossbars up to 8.7x and 2.1x using Resent-50 and MobileNet-v2, respectively, with less than 0.3% accuracy drop on ImageNet.

cs.AR↗

arXiv2021

A Driving Behavior Recognition Model with Bi-LSTM and Multi-Scale CNN

In autonomous driving, perceiving the driving behaviors of surrounding agents is important for the ego-vehicle to make a reasonable decision. In this paper, we propose a neural network model based on trajectories information for driving behavior recognition. Unlike existing trajectory-based methods that recognize the driving behavior using the hand-crafted features or directly encoding the trajectory, our model involves a Multi-Scale Convolutional Neural Network (MSCNN) module to automatically extract the high-level features which are supposed to encode the rich spatial and temporal information. Given a trajectory sequence of an agent as the input, firstly, the Bi-directional Long Short Term Memory (Bi-LSTM) module and the MSCNN module respectively process the input, generating two features, and then the two features are fused to classify the behavior of the agent. We evaluate the proposed model on the public BLVD dataset, achieving a satisfying performance.

cs.CV↗