arXiv subjects

Yichen Zhang

Publications and source records attributed to Yichen Zhang.

At least 181 records · Page 10Linked to original sources

arXiv2021

Unsupervised data augmentation for object detection

Data augmentation has always been an effective way to overcome overfitting issue when the dataset is small. There are already lots of augmentation operations such as horizontal flip, random crop or even Mixup. However, unlike image classification task, we cannot simply perform these operations for object detection task because of the lack of labeled bounding boxes information for corresponding generated images. To address this challenge, we propose a framework making use of Generative Adversarial Networks(GAN) to perform unsupervised data augmentation. To be specific, based on the recently supreme performance of YOLOv4, we propose a two-step pipeline that enables us to generate an image where the object lies in a certain position. In this way, we can accomplish the goal that generating an image with bounding box label.

cs.CV↗

arXiv2021

The Perseus ALMA Chemistry Survey (PEACHES). I. The Complex Organic Molecules in Perseus Embedded Protostars

To date, about two dozen low-mass embedded protostars exhibit rich spectra with lines of complex organic molecule (COM). These protostars seem to possess different enrichment in COMs. However, the statistics of COM abundance in low-mass protostars are limited by the scarcity of observations. This study introduces the Perseus ALMA Chemistry Survey (PEACHES), which aims at unbiasedly characterizing the chemistry of COMs toward the embedded (Class 0/I) protostars in the Perseus molecular cloud. Of 50 embedded protostars surveyed, 58% of them have emission from COMs. A 56%, 32%, and 40% of the protostars have CH$_3$OH, CH$_3$OCHO, and N-bearing COMs, respectively. The detectability of COMs depends neither on the averaged continuum brightness temperature, a proxy of the H$_2$ column density, nor on the bolometric luminosity and the bolometric temperature. For the protostars with detected COMs, CH$_3$OH has a tight correlation with CH$_3$CN, spanning more than two orders of magnitude in column densities normalized by the continuum brightness temperature, suggesting a chemical relation between CH$_3$OH and CH$_3$CN and a large chemical diversity in the PEACHES samples at the same time. A similar trend with more scatter is also found between all identified COMs, hinting at a common chemistry for the sources with COMs. The correlation between COMs is insensitive to the protostellar properties, such as the bolometric luminosity and the bolometric temperature. The abundance of larger COMs (CH$_3$OCHO and CH$_3$OCH$_3$) relative to that of smaller COMs (CH$_3$OH and CH$_3$CN) increases with the inferred gas column density, hinting at an efficient production of complex species in denser envelopes.

astro-ph.GA↗

arXiv2021

Provably Correct Controller Synthesis of Switched Stochastic Systems with Metric Temporal Logic Specifications: A Case Study on Power Systems

In this paper, we present a provably correct controller synthesis approach for switched stochastic control systems with metric temporal logic (MTL) specifications with provable probabilistic guarantees. We first present the stochastic control bisimulation function for switched stochastic control systems, which bounds the trajectory divergence between the switched stochastic control system and its nominal deterministic control system in a probabilistic fashion. We then develop a method to compute optimal control inputs by solving an optimization problem for the nominal trajectory of the deterministic control system with robustness against initial state variations and stochastic uncertainties. We implement our robust stochastic controller synthesis approach on both a four-bus power system and a nine-bus power system under generation loss disturbances, with MTL specifications expressing requirements for the grid frequency deviations, wind turbine generator rotor speed variations and the power flow constraints at different power lines.

eess.SY↗

arXiv2021

The Hot and Dynamic Birth of Massive Stars from the ngVLA Perspective

The Next Generation Very Large Array (ngVLA) has excellent capabilities to unveil various dynamical and chemical processes in massive star formation at the unexplored innermost regions. Based on the recent observations of ALMA/VLA as well as theoretical predictions, we propose several intriguing topics in massive star formation from the perspective of the ngVLA. In the disk scale of $\lesssim$ 100 au around massive protostars, dust grains are expected to be destructed/sublimated because the physical conditions of temperature, shocks, and radiation are much more intense than those in the envelopes, which are typically observed as hot cores. The high sensitivity and resolution of the ngVLA will enable us to detect the gaseous refractories released by dust destruction, e.g., SiO, NaCl, and AlO, which trace disk kinematics and give new insights into the metallic elements in star-forming regions, i.e., astromineralogy. The multi-epoch survey by the ngVLA will provide demographics of forming massive multiples with separations of $\lesssim$ 10 au with their proper motion. Combining with observations of refractory molecular lines and hydrogen recombination lines, we can reproduce the three-dimensional orbital motions of massive proto-binaries. Moreover, the 1-mas resolution of the ngVLA could possibly take the first-ever picture of the photospheric surface of an accreting protostar, if it is bloated to the au scale by the high accretion rates of mass and thermal energy.

astro-ph.GA↗

arXiv2021

Robust Trajectory-Constrained Frequency Control for Microgrids Considering Model Linearization Error

The capability to switch between grid-connected and islanded modes has promoted adoption of microgrid technology for powering remote locations. Stabilizing frequency during the islanding event, however, is a challenging control task, particularly under high penetration of converter-interfaced sources. In this paper, a numerical optimal control (NOC)-based control synthesis methodology is proposed for preparedness of microgrid islanding that ensure guaranteed performance. The key feature of the proposed paradigm is near real-time centralized scheduling for real-time decentralized executing. For tractable computation, linearized models are used in the problem formulation. To accommodate the linearization errors, interval analysis is employed to compute linearization-induced uncertainty as numerical intervals so that the NOC problem can be formulated into a robust mixed-integer linear program. The proposed control is verified on the full nonlinear model in Simulink. The simulation results shown effectiveness of the proposed control paradigm and the necessity of considering linearization-induced uncertainty.

eess.SY↗

arXiv2021

First-order Newton-type Estimator for Distributed Estimation and Inference

This paper studies distributed estimation and inference for a general statistical problem with a convex loss that could be non-differentiable. For the purpose of efficient computation, we restrict ourselves to stochastic first-order optimization, which enjoys low per-iteration complexity. To motivate the proposed method, we first investigate the theoretical properties of a straightforward Divide-and-Conquer Stochastic Gradient Descent (DC-SGD) approach. Our theory shows that there is a restriction on the number of machines and this restriction becomes more stringent when the dimension $p$ is large. To overcome this limitation, this paper proposes a new multi-round distributed estimation procedure that approximates the Newton step only using stochastic subgradient. The key component in our method is the proposal of a computationally efficient estimator of $Σ^{-1} w$, where $Σ$ is the population Hessian matrix and $w$ is any given vector. Instead of estimating $Σ$ (or $Σ^{-1}$) that usually requires the second-order differentiability of the loss, the proposed First-Order Newton-type Estimator (FONE) directly estimates the vector of interest $Σ^{-1} w$ as a whole and is applicable to non-differentiable losses. Our estimator also facilitates the inference for the empirical risk minimizer. It turns out that the key term in the limiting covariance has the form of $Σ^{-1} w$, which can be estimated by FONE.

stat.ML↗

arXiv2021

The ALMA Early Science view of FUor/EXor objects. I. Through the looking-glass of V2775 Ori

As part of an ALMA survey to study the origin of episodic accretion in young eruptive variables, we have observed the circumstellar environment of the star V2775 Ori. This object is a very young, pre-main sequence object which displays a large amplitude outburst characteristic of the FUor class. We present Cycle-2 band 6 observations of V2775 Ori with a continuum and CO (2-1) isotopologue resolution of 0.25\as (103 au). We report the detection of a marginally resolved circumstellar disc in the ALMA continuum with an integrated flux of $106 \pm 2$ mJy, characteristic radius of $\sim$ 30 au, inclination of $14.0^{+7.8}_{-14.5}$ deg, and is oriented nearly face-on with respect to the plane of the sky. The \co~emission is separated into distinct blue and red-shifted regions that appear to be rings or shells of expanding material from quasi-episodic outbursts. The system is oriented in such a way that the disc is seen through the outflow remnant of V2775 Ori, which has an axis along our line-of-sight. The $^{13}$CO emission displays similar structure to that of the \co, while the C$^{18}$O line emission is very weak. We calculated the expansion velocities of the low- and medium-density material with respect to the disc to be of -2.85 km s$^{-1}$ (blue), 4.4 km s$^{-1}$ (red) and -1.35 and 1.15 km s$^{-1}$ (for blue and red) and we derived the mass, momentum and kinetic energy of the expanding gas. The outflow has an hourglass shape where the cavities are not seen. We interpret the shapes that the gas traces as cavities excavated by an ancient outflow. We report a detection of line emission from the circumstellar disc and derive a lower limit of the gas mass of 3 \MJup.

astro-ph.SR↗

arXiv2021

FAUST II. Discovery of a Secondary Outflow in IRAS 15398-3359: Variability in Outflow Direction during the Earliest Stage of Star Formation?

We have observed the very low-mass Class 0 protostar IRAS 15398-3359 at scales ranging from 50 au to 1800 au, as part of the ALMA Large Program FAUST. We uncover a linear feature, visible in H2CO, SO, and C18O line emission, which extends from the source along a direction almost perpendicular to the known active outflow. Molecular line emission from H2CO, SO, SiO, and CH3OH further reveals an arc-like structure connected to the outer end of the linear feature and separated from the protostar, IRAS 15398-3359, by 1200 au. The arc-like structure is blue-shifted with respect to the systemic velocity. A velocity gradient of 1.2 km/s over 1200 au along the linear feature seen in the H2CO emission connects the protostar and the arc-like structure kinematically. SO, SiO, and CH3OH are known to trace shocks, and we interpret the arc-like structure as a relic shock region produced by an outflow previously launched by IRAS 15398-3359. The velocity gradient along the linear structure can be explained as relic outflow motion. The origins of the newly observed arc-like structure and extended linear feature are discussed in relation to turbulent motions within the protostellar core and episodic accretion events during the earliest stage of protostellar evolution.

astro-ph.SR↗

arXiv2020

FUEL: Fast UAV Exploration using Incremental Frontier Structure and Hierarchical Planning

Autonomous exploration is a fundamental problem for various applications of unmanned aerial vehicles. Existing methods, however, were demonstrated to insufficient exploration rate, due to the lack of efficient global coverage, conservative motion plans and low decision frequencies. In this paper, we propose FUEL, a hierarchical framework that can support Fast UAV Exploration in complex unknown environments. We maintain crucial information in the entire space required by exploration planning by a frontier information structure (FIS), which can be updated incrementally when the space is explored. Supported by the FIS, a hierarchical planner plans exploration motions in three steps, which find efficient global coverage paths, refine a local set of viewpoints and generate minimum-time trajectories in sequence. We present extensive benchmark and real-world tests, in which our method completes the exploration tasks with unprecedented efficiency (3-8 times faster) compared to state-of-the-art approaches. Our method will be made open source to benefit the community.

cs.RO↗

arXiv2020

Deep Active Learning for Solvability Prediction in Power Systems

Traditional methods for solvability region analysis can only have inner approximations with inconclusive conservatism. Machine learning methods have been proposed to approach the real region. In this letter, we propose a deep active learning framework for power system solvability prediction. Compared with the passive learning methods where the training is performed after all instances are labeled, the active learning selects most informative instances to be label and therefore significantly reduce the size of labeled dataset for training. In the active learning framework, the acquisition functions, which correspond to different sampling strategies, are defined in terms of the on-the-fly posterior probability from the classifier. The IEEE 39-bus system is employed to validate the proposed framework, where a two-dimensional case is illustrated to visualize the effectiveness of the sampling method followed by the full-dimensional numerical experiments.

cs.LG↗

arXiv2020

Ring formation by coagulation of dust aggregates in early phase of disk evolution around a protostar

Ring structures are observed by (sub-)millimeter dust continuum emission in various circumstellar disks from early stages of Class 0 and I to late stage of Class II young stellar objects (YSOs). In this paper, we study one of the possible scenarios of such ring formation in early stage, which is coagulation of dust aggregates. The dust grains grow in an inside-out manner because the growth timescale is roughly proportional to the orbital period. The boundary of the dust evolution can be regarded as the growth front, where the growth time is comparable to the disk age. With radiative transfer calculations based on the dust coagulation model, we find that the growth front can be observed as a ring structure because dust surface density is sharply changed at this position. Furthermore, we confirm that the observed ring positions in the YSOs with an age of $\lesssim1$ Myr are consistent with the growth front. The growth front could be important to create the ring structure in particular for early stage of the disk evolution such as Class 0 and I sources.

astro-ph.EP↗

arXiv2020

Approximating Trajectory Constraints with Machine Learning -- Microgrid Islanding with Frequency Constraints

In this paper, we introduce a deep learning aided constraint encoding method to tackle the frequency-constraint microgrid scheduling problem. The nonlinear function between system operating condition and frequency nadir is approximated by using a neural network, which admits an exact mixed-integer formulation (MIP). This formulation is then integrated with the scheduling problem to encode the frequency constraint. With the stronger representation power of the neural network, the resulting commands can ensure adequate frequency response in a realistic setting in addition to islanding success. The proposed method is validated on a modified 33-node system. Successful islanding with a secure response is simulated under the scheduled commands using a detailed three-phase model in Simulink. The advantages of our model are particularly remarkable when the inertia emulation functions from wind turbine generators are considered.

eess.SY↗

arXiv2020

Online Detection of Low-Quality Synchrophasor Data Considering Frequency Similarity

This letter proposes a new approach for online detection of low-quality synchrophasor data under both normal and event conditions. The proposed approach utilizes the features of synchrophasor data in time and frequency domains to distinguish multiple regional PMU signals and detect low-quality synchrophasor data. The proposed approach does not require any offline study and it is more effective to detect low-quality data with apparently indistinguishable profiles. Case studies from recorded synchrophasor measurements verify the effectiveness of the proposed approach.

eess.SY↗

arXiv2020

Pathological Visual Question Answering

Is it possible to develop an "AI Pathologist" to pass the board-certified examination of the American Board of Pathology (ABP)? To build such a system, three challenges need to be addressed. First, we need to create a visual question answering (VQA) dataset where the AI agent is presented with a pathology image together with a question and is asked to give the correct answer. Due to privacy concerns, pathology images are usually not publicly available. Besides, only well-trained pathologists can understand pathology images, but they barely have time to help create datasets for AI research. The second challenge is: since it is difficult to hire highly experienced pathologists to create pathology visual questions and answers, the resulting pathology VQA dataset may contain errors. Training pathology VQA models using these noisy or even erroneous data will lead to problematic models that cannot generalize well on unseen images. The third challenge is: the medical concepts and knowledge covered in pathology question-answer (QA) pairs are very diverse while the number of QA pairs available for modeling training is limited. How to learn effective representations of diverse medical concepts based on limited data is technically demanding. In this paper, we aim to address these three challenges. To our best knowledge, our work represents the first one addressing the pathology VQA problem. To deal with the issue that a publicly available pathology VQA dataset is lacking, we create PathVQA dataset. To address the second challenge, we propose a learning-by-ignoring approach. To address the third challenge, we propose to use cross-modal self-supervised learning. We perform experiments on our created PathVQA dataset and the results demonstrate the effectiveness of our proposed learning-by-ignoring method and cross-modal self-supervised learning methods.

cs.CV↗

arXiv2020

The SOFIA Massive (SOMA) Star Formation Survey. III. From Intermediate- to High-Mass Protostars

We present $\sim10-40\,μ$m SOFIA-FORCAST images of 14 intermediate-mass protostar candidates as part of the SOFIA Massive (SOMA) Star Formation Survey. We build spectral energy distributions (SEDs), also utilizing archival Spitzer, Herschel and IRAS data. We then fit the SEDs with radiative transfer (RT) models of Zhang & Tan (2018), based on Turbulent Core Accretion theory, to estimate key protostellar properties. With the addition of these intermediate-mass sources, SOMA protostars span luminosities from $\sim10^{2}-10^{6}\:L_{\odot}$, current protostellar masses from $\sim0.5-30\:M_{\odot}$ and ambient clump mass surface densities, $Σ_{\rm cl}$ from $0.1-3\:{\rm{g\:cm}^{-2}}$. A wide range of evolutionary states of the individual protostars and of the protocluster environments are also probed. We have also considered about 50 protostars identified in Infrared Dark Clouds and expected to be at the earliest stages of their evolution. With this global sample, most of the evolutionary stages of high- and intermediate-mass protostars are probed. From the best fitting models, there is no evidence of a threshold value of protocluster clump mass surface density being needed to form protostars up to $\sim25\:M_\odot$. However, to form more massive protostars, there is tentative evidence that $Σ_{\rm{cl}}$ needs to be $\gtrsim1\:{\rm{g\,cm}}^{-2}$. We discuss how this is consistent with expectations from core accretion models that include internal feedback from the forming massive star.

astro-ph.GA↗

arXiv2020

Salt, Hot Water, and Silicon Compounds Tracing Massive Twin Disks

We report results of 0.05"-resolution observations toward the O-type proto-binary system IRAS 16547-4247 with the Atacama Large Millimeter/submillimeter Array (ALMA). We present dynamical and chemical structures of the circumbinary disk, circumstellar disks, outflows and jets, illustrated by multi-wavelength continuum and various molecular lines. In particular, we detect sodium chloride, silicon compounds, and vibrationally-excited water lines as probes of the individual protostellar disks at a scale of 100 au. These are complementary to typical hot-core molecules tracing the circumbinary structures on a 1000-au scale. The H2O line tracing inner-disks has an upper-state energy of Eu/k>3000K, indicating a high temperature of the disks. On the other hand, despite the detected transitions of NaCl, SiO, and SiS not necessarily having high upper-state energies, they are enhanced only in the vicinity of the protostars. We interpret that these molecules are the products of dust destruction, which only happens in the inner disks. This is the second detection of alkali metal halide in protostellar systems after the case of the disk of Orion Source I, and also one of few massive protostellar disks associated with high-energy transition water and silicon compounds. These new results suggest these "hot-disk" lines may be common in innermost disks around massive protostars, and have great potential for future research of massive star formation. We also tentatively find that the twin disks are counter-rotating, which might give a hint of the origin of the massive proto-binary system IRAS 16547-4247.

astro-ph.SR↗

arXiv2020

High-speed Gaussian modulated continuous-variable quantum key distribution with a local local oscillator based on pilot-tone-assisted phase compensation

A high-speed Gaussian modulated continuous-variable quantum key distribution (CVQKD) with a local local oscillator (LLO) is experimentally demonstrated based on pilot-tone-assisted phase compensation. In the proposed scheme, the frequency-multiplexing and polarization-multiplexing techniques are used for the separate transmission and heterodyne detection between quantum signal and pilot tone, guaranteeing no crosstalk from strong pilot tone to weak quantum signal and different detection requirements of low-noise for quantum signal and high-saturation limitation for pilot tone. Moreover, compared with the conventional CVQKD based on homodyne detection, the proposed LLO-CVQKD scheme can measure X and P quadrature simultaneously using heterodyne detection without need of extra random basis selection. Besides, the phase noise, which contains the fast-drift phase noise due to the relative phase of two independent lasers and the slow-drift phase noise introduced by quantum channel disturbance, has been compensated experimentally in real time, so that a low level of excess noise with a 25km optical fiber channel is obtained for the achievable secure key rate of 7.04 Mbps in the asymptotic regime and 1.85 Mbps under the finite-size block of 10^7.

quant-ph↗

arXiv2020

COVID-CT-Dataset: A CT Scan Dataset about COVID-19

During the outbreak time of COVID-19, computed tomography (CT) is a useful manner for diagnosing COVID-19 patients. Due to privacy issues, publicly available COVID-19 CT datasets are highly difficult to obtain, which hinders the research and development of AI-powered diagnosis methods of COVID-19 based on CTs. To address this issue, we build an open-sourced dataset -- COVID-CT, which contains 349 COVID-19 CT images from 216 patients and 463 non-COVID-19 CTs. The utility of this dataset is confirmed by a senior radiologist who has been diagnosing and treating COVID-19 patients since the outbreak of this pandemic. We also perform experimental studies which further demonstrate that this dataset is useful for developing AI-based diagnosis models of COVID-19. Using this dataset, we develop diagnosis methods based on multi-task learning and self-supervised learning, that achieve an F1 of 0.90, an AUC of 0.98, and an accuracy of 0.89. According to the senior radiologist, models with such performance are good enough for clinical usage. The data and code are available at https://github.com/UCSD-AI4H/COVID-CT

cs.LG↗