Searcharxiv⌕ Search

arXiv subjects

Tao Ma

Publications and source records attributed to Tao Ma.

At least 55 records · Page 3Linked to original sources
arXiv2021

Perception Entropy: A Metric for Multiple Sensors Configuration Evaluation and Design

Sensor configuration, including the sensor selections and their installation locations, serves a crucial role in autonomous driving. A well-designed sensor configuration significantly improves the performance upper bound of the perception system. However, as leveraging multiple sensors is becoming the mainstream setting, existing methods mainly focusing on single-sensor configuration problems are hardly utilized in practice. To tackle these issues, we propose a novel method based on conditional entropy in Bayesian theory to evaluate the sensor configurations containing both cameras and LiDARs. Correspondingly, an evaluation metric, perception entropy, is introduced to measure the difference between two configurations, which considers both the perception algorithm performance and the selections of the sensors. To the best of our knowledge, this is the first method to tackle the multi-sensor configuration problem for autonomous vehicles. The simulation results, extensive comparisons, and analysis all demonstrate the superior performance of our proposed approach.

cs.RO↗
arXiv2021

CRLF: Automatic Calibration and Refinement based on Line Feature for LiDAR and Camera in Road Scenes

For autonomous vehicles, an accurate calibration for LiDAR and camera is a prerequisite for multi-sensor perception systems. However, existing calibration techniques require either a complicated setting with various calibration targets, or an initial calibration provided beforehand, which greatly impedes their applicability in large-scale autonomous vehicle deployment. To tackle these issues, we propose a novel method to calibrate the extrinsic parameter for LiDAR and camera in road scenes. Our method introduces line features from static straight-line-shaped objects such as road lanes and poles in both image and point cloud and formulates the initial calibration of extrinsic parameters as a perspective-3-lines (P3L) problem. Subsequently, a cost function defined under the semantic constraints of the line features is designed to perform refinement on the solved coarse calibration. The whole procedure is fully automatic and user-friendly without the need to adjust environment settings or provide an initial calibration. We conduct extensive experiments on KITTI and our in-house dataset, quantitative and qualitative results demonstrate the robustness and accuracy of our method.

cs.CV↗
arXiv2020

Contamination of TEM Holders Quantified and Mitigated with Open-Hardware, High-Vacuum Bakeout System

Hydrocarbon contamination plagues high-resolution and analytical electron microscopy by depositing carbonaceous layers onto surfaces during electron irradiation, which can render carefully prepared specimens useless. Increased specimen thickness degrades resolution with beam broadening alongside loss of contrast. The large inelastic cross-section of carbon hampers accurate atomic species detection. Oxygen and water molecules pose problems of lattice damage by chemically etching the specimen during imaging. These constraints on high-resolution and spectroscopic imaging demand clean, high-vacuum microscopes with dry pumps. Here, we present an open-hardware design of a high-vacuum manifold for transmission electron microscopy (TEM) holders to mitigate hydrocarbon and residual species exposure. We quantitatively show that TEM holders are inherently dirty and introduce a range of unwanted chemical species. Overnight storage in our manifold reduces contaminants by 1-2 orders of magnitude and promotes 2-4 times faster vacuum recovery. A built-in bakeout system further reduces contaminants partial pressure to below $10^{-10}$ Torr (~4 orders of magnitude down from ambient storage) and alleviates monolayer adsorption during a typical TEM experiment. We determine that bakeout of TEM holder with specimen held therein is the optimal cleaning method. Our high-vacuum manifold design is published with open-source blueprints, parts list, and cost.

physics.ins-det↗
arXiv2020

Speech Fusion to Face: Bridging the Gap Between Human's Vocal Characteristics and Facial Imaging

While deep learning technologies are now capable of generating realistic images confusing humans, the research efforts are turning to the synthesis of images for more concrete and application-specific purposes. Facial image generation based on vocal characteristics from speech is one of such important yet challenging tasks. It is the key enabler to influential use cases of image generation, especially for business in public security and entertainment. Existing solutions to the problem of speech2face renders limited image quality and fails to preserve facial similarity due to the lack of quality dataset for training and appropriate integration of vocal features. In this paper, we investigate these key technical challenges and propose Speech Fusion to Face, or SF2F in short, attempting to address the issue of facial image quality and the poor connection between vocal feature domain and modern image generation models. By adopting new strategies on data model and training, we demonstrate dramatic performance boost over state-of-the-art solution, by doubling the recall of individual identity, and lifting the quality score from 15 to 19 based on the mutual information score with VGGFace classifier.

cs.CV↗
arXiv2020

ASAPP-ASR: Multistream CNN and Self-Attentive SRU for SOTA Speech Recognition

In this paper we present state-of-the-art (SOTA) performance on the LibriSpeech corpus with two novel neural network architectures, a multistream CNN for acoustic modeling and a self-attentive simple recurrent unit (SRU) for language modeling. In the hybrid ASR framework, the multistream CNN acoustic model processes an input of speech frames in multiple parallel pipelines where each stream has a unique dilation rate for diversity. Trained with the SpecAugment data augmentation method, it achieves relative word error rate (WER) improvements of 4% on test-clean and 14% on test-other. We further improve the performance via N-best rescoring using a 24-layer self-attentive SRU language model, achieving WERs of 1.75% on test-clean and 4.46% on test-other.

eess.AS↗
arXiv2020

Calibration and performance of the neutron detector onboard of the DAMPE mission

The DArk Matter Particle Explorer (DAMPE), one of the four space-based scientific missions within the framework of the Strategic Pioneer Program on Space Science of the Chinese Academy of Sciences, has been successfully launched on Dec. 17th 2015 from Jiuquan launch center. One of the most important scientific goals of DAMPE is to search for the evidence of dark matter indirectly by measuring the spectrum of high energy cosmic-ray electrons. The neutron detector, one of the four sub-payloads of DAMPE, is designed to distinguish high energy electrons from hadron background by measuring the secondary neutrons produced in the shower. In this paper, a comprehensive introduction of the neutron detector is presented, including the design, the calibration and the performance. The analysis with simulated data and flight data indicates a powerful proton rejection capability of the neutron detector, which plays an essential role for TeV electron identification of DAMPE.

physics.ins-det↗
arXiv2020

Environmental tests of the HXI spectrometer for the ASO-S mission

The advanced space-based solar observatory (ASO-S) is the first Chinese solar mission which is scheduled to be launched in the end of 2021. It is designed to study the magnetic field, the solar coronal mass ejections the solar flares as well as their relationships. ASO-S includes three scientific payloads, the Full-disk vector MagnetoGraph (FMG), the Lyman-alpha Solar Telescope (LST) and the Hard X-ray Imager (HXI). As a key part on ASO-S, HXI will improve our understanding on the solar flares during the 25th solar maximum. The HXI is composed of three instruments, the collimator (HXI-C), the spectrometer (HXI-S) and the electrical control box (HXI-E). This paper describes the procedure and results of the environmental tests including the mechanical and thermal which were done on the qualification model of the HXI-S. The functional tests of the spectrometer are further carried out, and the results show that the detector could operate normally. 10

physics.ins-det↗
arXiv2019

PasteGAN: A Semi-Parametric Method to Generate Image from Scene Graph

Despite some exciting progress on high-quality image generation from structured(scene graphs) or free-form(sentences) descriptions, most of them only guarantee the image-level semantical consistency, i.e. the generated image matching the semantic meaning of the description. They still lack the investigations on synthesizing the images in a more controllable way, like finely manipulating the visual appearance of every object. Therefore, to generate the images with preferred objects and rich interactions, we propose a semi-parametric method, PasteGAN, for generating the image from the scene graph and the image crops, where spatial arrangements of the objects and their pair-wise relationships are defined by the scene graph and the object appearances are determined by the given object crops. To enhance the interactions of the objects in the output, we design a Crop Refining Network and an Object-Image Fuser to embed the objects as well as their relationships into one map. Multiple losses work collaboratively to guarantee the generated images highly respecting the crops and complying with the scene graphs while maintaining excellent image quality. A crop selector is also proposed to pick the most-compatible crops from our external object tank by encoding the interactions around the objects in the scene graph if the crops are not provided. Evaluated on Visual Genome and COCO-Stuff dataset, our proposed method significantly outperforms the SOTA methods on Inception Score, Diversity Score and Fréchet Inception Distance. Extensive experiments also demonstrate our method's ability to generate complex and diverse images with given objects.

cs.CV↗
arXiv2019

Uncompensated Polarization in Incommensurate Modulations of Perovskite Antiferroelectrics

Complex polar structures of incommensurate modulations (ICMs) are revealed in chemically modified PbZrO$_3$ perovskite antiferroelectrics using advanced transmission electron microscopy techniques. The Pb-cation displacements, previously assumed to arrange in a fully-compensated antiparallel fashion, are found to be either antiparallel but with different magnitudes, or in a nearly orthogonal arrangement in adjacent stripes in the ICMs. Ab initio calculations corroborate the low-energy state of these arrangements. Our discovery corrects the atomic understanding of ICMs in PbZrO$_3$-based perovskite antiferroelectrics.

cond-mat.mtrl-sci↗
arXiv2019

State-of-the-Art Speech Recognition Using Multi-Stream Self-Attention With Dilated 1D Convolutions

Self-attention has been a huge success for many downstream tasks in NLP, which led to exploration of applying self-attention to speech problems as well. The efficacy of self-attention in speech applications, however, seems not fully blown yet since it is challenging to handle highly correlated speech frames in the context of self-attention. In this paper we propose a new neural network model architecture, namely multi-stream self-attention, to address the issue thus make the self-attention mechanism more effective for speech recognition. The proposed model architecture consists of parallel streams of self-attention encoders, and each stream has layers of 1D convolutions with dilated kernels whose dilation rates are unique given stream, followed by a self-attention layer. The self-attention mechanism in each stream pays attention to only one resolution of input speech frames and the attentive computation can be more efficient. In a later stage, outputs from all the streams are concatenated then linearly projected to the final embedding. By stacking the proposed multi-stream self-attention encoder blocks and rescoring the resultant lattices with neural network language models, we achieve the word error rate of 2.2% on the test-clean dataset of the LibriSpeech corpus, the best number reported thus far on the dataset.

cs.CL↗
arXiv2019

Fatigue-resistant high-performance elastocaloric materials via additive manufacturing

Elastocaloric cooling, which exploits the latent heat released and absorbed as stress-induced phase transformations are reversibly cycled in shape memory alloys, has recently emerged as a frontrunner in non-vapor-compression cooling technologies. The intrinsically high thermodynamic efficiency of elastocaloric materials is limited only by work hysteresis. Here, we report on creating high-performance low-hysteresis elastocaloric cooling materials via additive manufacturing of Titanium-Nickel (Ti-Ni) alloys. Contrary to established knowledge of the physical metallurgy of Ti-Ni alloys, intermetallic phases are found to be beneficial to elastocaloric performances when they are combined with the binary Ti-Ni compound in nanocomposite configurations. The resulting microstructure gives rise to quasi-linear stress-strain behaviors with extremely small hysteresis, leading to enhancement in the materials efficiency by a factor of five. Furthermore, despite being composed of more than 50% intermetallic phases, the reversible, repeatable elastocaloric performance of this material is shown to be stable over one million cycles. This result opens the door for direct implementation of additive manufacturing to elastocaloric cooling systems where versatile design strategy enables both topology optimization of heat exchangers as well as unique microstructural control of metallic refrigerants.

cond-mat.mtrl-sci↗
arXiv2019

Boundedness of differential transforms for one-sided fractional Poisson-type operator sequence

In this paper, we analyze the convergence speed of a series related with $\mathcal{P}_τ^αf$ by discussing the behavior of the family of operators \begin{equation*} T_N^αf(t) = \sum_{j=N_1}^{N_2} v_j(\mathcal{P}_{a_{j+1}}^αf(t)-\mathcal{P}_{a_j}^αf(t)),\quad ~N=(N_1,N_2)\in \mathbb{Z}^2\quad \hbox{with} \quad N_1<N_2, \end{equation*} where $\{v_j\}_{j\in \mathbb Z}$ is a bounded number sequence, and $\{a_j\}_{j\in \mathbb{Z}}$ is a $ρ$-lacunary sequence of positive numbers, that is, $1<ρ\leq a_{j+1}/a_j, \text{for all}\ j\in \mathbb{Z}.$ We shall show the boundedness of the maximal operator \begin{equation*}T^*f(t)=\sup_N |T_N^αf(t)|, \quad t\in\mathbb{R}, \end{equation*} in the one-sided weighted Lebesgue spaces $L^p(\mathbb{R},ω)(ω\in A_p^-$), $1< p < \infty$. As a consequence we infer the existence of the limit, in norm and almost everywhere, of the family $T_N^αf$ for functions in $L^p(\mathbb{R},ω)$. Results for $L^1(\mathbb{R},ω)(ω\in A_1^-)$, $L^\infty(\mathbb{R})$ and $BMO(\mathbb{R})$ are also obtained. It is also shown that the local size of $T^*f$, for functions $f$ having local support, is the same with the order of a singular integral. Moreover, if $\{v_j\}_{j\in \mathbb Z}\in \ell^p(\mathbb Z)$, we get an intermediate size between the local size of singular integrals and Hardy-Littlewood maximal operator.

math.CA↗
arXiv2019

Vector-valued $q$-variational inequalities for averaging operators and Hilbert transform

Recently, in \cite{GXHTM}, the authors established $L^p$-boundedness of vector-valued $q$-variational inequalities for averaging operators which take values in the Banach space satisfying martingale cotype $q$ property. In this paper, we prove that martingale cotype $q$ property is also necessary for the vector-valued $q$-variational inequalities, which is a question left open. Moreover, we characterize UMD property and martingale cotype $q$ property in terms of vector valued $q$-variational inequalities for Hilbert transform.

math.CA↗
arXiv2019

Crystal growth, microstructure and physical properties of SrMnSb$_2$

We report on the crystal and magnetic structures, magnetic, and transport properties of SrMnSb$_2$ single crystals grown by the self-flux method. Magnetic susceptibility measurements reveal an antiferromagnetic (AFM) transition at $T_{\rm N} = 295(3)$ K. Above $T_{\rm N}$, the susceptibility slightly increases and forms a broad peak at $T \sim 420$ K, which is a typical feature of two-dimensional magnetic systems. Neutron diffraction measurements on single crystals confirm the previously reported C-type AFM structure below $T_{\rm N}$. Both de Haas-van Alphen (dHvA) and Shubnikov-de Haas (SdH) effects are observed in SrMnSb$_2$ single crystals. Analysis of the oscillatory component by a Fourier transform shows that the prominent frequencies obtained by the two different techniques are practically the same within error regardless of sample size or saturated magnetic moment. Transmission electron microscopy (TEM) reveals the existence of stacking faults in the crystals, which result from a horizontal shift of Sb atomic layers suggesting possible ordering of Sb vacancies in the crystals. Increase of temperature in susceptibility measurements leads to the formation of a strong peak at $T \sim {570}$ K that upon cooling under magnetic field the susceptibility shows a ferromagnetic transition at $T_{\rm C} \sim 580$ K. Neutron powder diffraction on crushed single-crystals does not support an FM phase above $T_{\rm N}$. Furthermore, X-ray magnetic circular dichroism (XMCD) measurements of a single crystal at the $L_{2,3}$ edge of Mn shows a signal due to induced canting of AFM moments by the applied magnetic field. All evidence strongly suggests that a chemical transformation at the surface of single crystals occurs above 500 K concurrently producing a minute amount of ferromagnetic impurity phase.

cond-mat.mtrl-sci↗
arXiv2019

Multi-task Learning for Financial Forecasting

Financial forecasting is challenging and attractive in machine learning. There are many classic solutions, as well as many deep learning based methods, proposed to deal with it yielding encouraging performance. Stock time series forecasting is the most representative problem in financial forecasting. Due to the strong connections among stocks, the information valuable for forecasting is not only included in individual stocks, but also included in the stocks related to them. However, most previous works focus on one single stock, which easily ignore the valuable information in others. To leverage more information, in this paper, we propose a jointly forecasting approach to process multiple time series of related stocks simultaneously, using multi-task learning framework. Compared to the previous works, we use multiple networks to forecast multiple related stocks, using the shared and private information of them simultaneously through multi-task learning. Moreover, we propose an attention method learning an optimized weighted combination of shared and private information based on the idea of Capital Asset Pricing Model (CAPM) to help forecast. Experimental results on various data show improved forecasting performance over baseline methods.

cs.LG↗
arXiv2014

Vector valued $q$-variation for differential operators and semigroups I

In this paper, we establish $\mathcal B$-valued variational inequalities for differential operators, ergodic averages and symmetric diffusion semigroups under the condition that Banach space $\mathcal B$ has martingale cotype property. These results generalize, on the one hand Pisier and Xu's result on the variational inequalities for $\mathcal B$-valued martingales, on the other hand many classical variational inequalities in harmonic analysis and ergodic theory. Moreover, we show that Rademacher cotype $q$ is necessary for the $\mathcal B$-valued $q$-variational inequalities. As applications of the variational inequalities, we deduce the jump estimates and obtain quantitative information on the rate of convergence. It turns out the rate of convergence depends on the geometric property of the Banach space under consideration, which considerably improve Cowling and Leinert's result where it is shown that the convergence always holds for all Banach spaces.

math.FA↗
arXiv2014

Vector valued $q$-variation for differential operators and semigroups II

In this paper, we establish UMD lattice-valued variational inequalities for differential operators, ergodic averages and analytic semigroups. These results generalize, on the one hand some scalar-valued variational inequalities in ergodic theory, on the other hand Xu's very recent result on UMD lattice-valued maximal inequality. As a consequence, we deduce the jump estimates and obtain quantitative information on the rate of the pointwise convergence.

math.FA↗