M. Y. Sun

M. Y. Sun
Are you M. Y. Sun?

Claim your profile, edit publications, add additional information:

Contact Details

M. Y. Sun

Pubs By Year

Pub Categories

Computer Science - Computer Vision and Pattern Recognition (13)
Computer Science - Computation and Language (13)
High Energy Astrophysical Phenomena (7)
Astrophysics of Galaxies (5)
Computer Science - Learning (4)
Computer Science - Artificial Intelligence (4)
Computer Science - Multimedia (3)
Cosmology and Nongalactic Astrophysics (3)
Computer Science - Cryptography and Security (2)
Statistics - Machine Learning (2)
Computer Science - Software Engineering (1)
Mathematics - Algebraic Geometry (1)
Physics - Soft Condensed Matter (1)
Mathematics - Commutative Algebra (1)
Instrumentation and Methods for Astrophysics (1)
Physics - Mesoscopic Systems and Quantum Hall Effect (1)
Computer Science - Databases (1)
Physics - Optics (1)
Computer Science - Logic in Computer Science (1)
Physics - Materials Science (1)
Computer Science - Robotics (1)
Mathematics - Combinatorics (1)
Computer Science - Graphics (1)
Solar and Stellar Astrophysics (1)
Computer Science - Distributed; Parallel; and Cluster Computing (1)
Mathematics - Information Theory (1)
Computer Science - Information Theory (1)
Physics - Physics and Society (1)
Nuclear Experiment (1)

Publications Authored By M. Y. Sun

For survival, a living agent must have the ability to assess risk (1) by temporally anticipating accidents before they occur, and (2) by spatially localizing risky regions in the environment to move away from threats. In this paper, we take an agent-centric approach to study the accident anticipation and risky region localization tasks. We propose a novel soft-attention Recurrent Neural Network (RNN) which explicitly models both spatial and appearance-wise non-linear interaction between the agent triggering the event and another agent or static-region involved. Read More

Knowledge graphs (KGs) can provide significant relational information and have been widely utilized in various tasks. However, there may exist amounts of noises and conflicts in KGs, especially in those constructed automatically with less human supervision. To address this problem, we propose a novel confidence-aware knowledge representation learning framework (CKRL), which detects possible noises in KGs while learning knowledge representations with confidence simultaneously. Read More

We propose a max-pooling based loss function for training Long Short-Term Memory (LSTM) networks for small-footprint keyword spotting (KWS), with low CPU, memory, and latency requirements. The max-pooling loss training can be further guided by initializing with a cross-entropy loss trained network. A posterior smoothing based evaluation approach is employed to measure keyword spotting performance. Read More

Watching a 360{\deg} sports video requires a viewer to continuously select a viewing angle, either through a sequence of mouse clicks or head movements. To relieve the viewer from this "360 piloting" task, we propose "deep 360 pilot" -- a deep learning-based agent for piloting through 360{\deg} sports videos automatically. At each frame, the agent observes a panoramic image and has the knowledge of previously selected viewing angles. Read More

We present a $Chandra$ study of the hot intragroup medium (hIGM) of the galaxy group NCG2563. The $Chandra$ mosaic observations, with a total exposure time of ~430 ks, allow the gas density to be detected beyond $R_{200}$ and the gas temperature out to 0.75 $R_{200}$. Read More

Despite the recent success of deep-learning based semantic segmentation, deploying a pre-trained road scene segmenter to a city whose images are not presented in the training set would not achieve satisfactory performance due to dataset biases. Instead of collecting a large number of annotated images of each city of interest to train or refine the segmenter, we propose an unsupervised learning approach to adapt road scene segmenters across different cities. By utilizing Google Street View and its time-machine feature, we can collect unannotated images for each road scene at different times, so that the associated static-object priors can be extracted accordingly. Read More

We present $Suzaku$ off-center observations of two poor galaxy groups, NGC 3402 and NGC 5129, with temperatures below 1 keV. Through spectral decomposition, we measure their surface brightnesses and temperatures out to 330 and 680 times the critical density of the universe for NGC 3402 and NGC 5129, respectively. These quantities are consistent with extrapolations from existing inner measurements of the two groups. Read More

While part-of-speech (POS) tagging and dependency parsing are observed to be closely related, existing work on joint modeling with manually crafted feature templates suffers from the feature sparsity and incompleteness problems. In this paper, we propose an approach to joint POS tagging and dependency parsing using transition-based neural networks. Three neural network based classifiers are designed to resolve shift/reduce, tagging, and labeling conflicts. Read More

We investigate the dependence of black-hole accretion rate (BHAR) on host-galaxy star formation rate (SFR) and stellar mass ($M_*$) in the CANDELS/GOODS-South field in the redshift range of $0.5\leq z < 2.0$. Read More

The concept of metric dimension has applications in a variety of fields, such as chemistry, robotic navigation, and combinatorial optimization. We show bounds for graphs with $n$ vertices and metric dimension $\beta$. For Hamiltonian outerplanar graphs, we have $\beta \leq \left\lceil\frac{n}2\right\rceil$; for outerplanar graphs in general, we have $\beta \leq \left\lfloor\frac{2n}{3}\right\rfloor$; for maximal planar graphs, we have $\beta \leq \left\lfloor\frac{3n}{4}\right\rfloor$. Read More

We have discovered large amounts of molecular gas, as traced by CO emission, in the ram pressure stripped gas tail of the Coma cluster galaxy D100 (GMP 2910), out to large distances of about 50 kpc. D100 has a 60 kpc long, strikingly narrow tail which is bright in X-rays and H{\alpha}. Our observations with the IRAM 30m telescope reveal in total ~ 10^9 M_sun of H_2 (assuming the standard CO-to-H_2 conversion) in several regions along the tail, thus indicating that molecular gas may dominate its mass. Read More

Robot vision is a fundamental device for human-robot interaction and robot complex tasks. In this paper, we use Kinect and propose a feature graph fusion (FGF) for robot recognition. Our feature fusion utilizes RGB and depth information to construct fused feature from Kinect. Read More

We introduce two tactics to attack agents trained by deep reinforcement learning algorithms using adversarial examples, namely the strategically-timed attack and the enchanting attack. In the strategically-timed attack, the adversary aims at minimizing the agent's reward by only attacking the agent at a small subset of time steps in an episode. Limiting the attack activity to this subset helps prevent detection of the attack by the agent. Read More

Motivated by the discovery of a handful of pulsating, extremely low mass white dwarfs (ELM WDs, mass $M \lesssim 0.17\, M_\odot$) which likely have WD companions, this paper discusses binary formation models for these systems. Formation of an ELM WD by unstable mass transfer (MT) or a common envelope (CE) event is unlikely. Read More

The Bose polaron is a quasi-particle of a mobile impurity dressed by surrounding bosons. Since it is known from the few-body physics that an impurity can form a sequence of Efimov bound states with two bosons on the vicinity of a Feshbach resonance, one would expect that this Efimov correlation can manifest itself in the Bose polaron problem. Nevertheless, no signature of Efimov physics has been reported in the spectroscopy measurements of Bose polarons up to date. Read More

Acoustic unit discovery (AUD) is a process of automatically identifying a categorical acoustic unit inventory from speech and producing corresponding acoustic unit tokenizations. AUD provides an important avenue for unsupervised acoustic model training in a zero resource setting where expert-provided linguistic knowledge and transcribed speech are unavailable. Therefore, to further facilitate zero-resource AUD process, in this paper, we demonstrate acoustic feature representations can be significantly improved by (i) performing linear discriminant analysis (LDA) in an unsupervised self-trained fashion, and (ii) leveraging resources of other languages through building a multilingual bottleneck (BN) feature extractor to give effective cross-lingual generalization. Read More

Photo composition is an important factor affecting the aesthetics in photography. However, it is a highly challenging task to model the aesthetic properties of good compositions due to the lack of globally applicable rules to the wide variety of photographic styles. Inspired by the thinking process of photo taking, we treat the photo composition problem as a view finding process which successively examines pairs of views and determines the aesthetic preference. Read More

The dynamical behavior of liquids is frequently characterized by the fragility, which can be defined from the temperature dependence of the shear viscosity, {\eta}. For a strong liquid, the activation energy for {\eta} changes little with cooling towards the glass transition temperature, Tg. The change is much greater in fragile liquids, with the activation energy becoming very large near Tg. Read More

As a complement to cloud computing, fog computing can offer many benefits in terms of avoiding the long wide-area network (WAN) propagation delay and relieving the network bandwidth burden by providing local services to nearby end users, resulting in a reduced revenue loss associated with the WAN propagation delay and network bandwidth cost for a cloud provider. However, serving the requests of end-users would lead to additional energy costs for fog devices, thus the could provider must compensate fog devices for their losses. In this paper, we investigate the problem of minimizing the total cost of a cloud provider without sacrificing the interests of fog devices. Read More

Emoji is an essential component in dialogues which has been broadly utilized on almost all social platforms. It could express more delicate feelings beyond plain texts and thus smooth the communications between users, making dialogue systems more anthropomorphic and vivid. In this paper, we focus on automatically recommending appropriate emojis given the contextual information in multi-turn dialogue systems, where the challenges locate in understanding the whole conversations. Read More

Kes 79 (G33.6+0.1) is an aspherical thermal composite supernova remnant (SNR) observed across the electromagnetic spectrum and showing an unusual highly-structured morphology, in addition to harboring a central compact object (CCO). Read More

Android, the most popular mobile OS, has around 78% of the mobile market share. Due to its popularity, it attracts many malware attacks. In fact, people have discovered around one million new malware samples per quarter, and it was reported that over 98% of these new malware samples are in fact "derivatives" (or variants) from existing malware families. Read More

Network representation learning (NRL) aims to build low-dimensional vectors for vertices in a network. Most existing NRL methods focus on learning representations from local context of vertices (such as their neighbors). Nevertheless, vertices in many complex networks also exhibit significant global patterns widely known as communities. Read More

While recent neural machine translation approaches have delivered state-of-the-art performance for resource-rich language pairs, they suffer from the data scarcity problem for resource-scarce language pairs. Although this problem can be alleviated by exploiting a pivot language to bridge the source and target languages, the source-to-pivot and pivot-to-target translation models are usually independently trained. In this work, we introduce a joint training algorithm for pivot-based neural machine translation. Read More

In an Internet of Things network, multiple sensors send information to a fusion center for it to infer a public hypothesis of interest. However, the same sensor information may be used by the fusion center to make inferences of a private nature that the sensors wish to protect. To model this, we adopt a decentralized hypothesis testing framework with binary public and private hypotheses. Read More

Joint representation learning of text and knowledge within a unified semantic space enables us to perform knowledge graph completion more accurately. In this work, we propose a novel framework to embed words, entities and relations into the same continuous vector space. In this model, both entity and relation embeddings are learned by taking knowledge graph and plain text into consideration. Read More

We propose a scalable approach to learn video-based question answering (QA): answer a "free-form natural language question" about a video content. Our approach automatically harvests a large number of videos and descriptions freely available online. Then, a large number of candidate QA pairs are automatically generated from descriptions rather than manually annotated. Read More

We present X-ray source catalogs for the $\approx7$ Ms exposure of the Chandra Deep Field-South (CDF-S), which covers a total area of 484.2 arcmin$^2$. Utilizing WAVDETECT for initial source detection and ACIS Extract for photometric extraction and significance assessment, we create a main source catalog containing 1008 sources that are detected in up to three X-ray bands: 0. Read More

Masses of $^{52g,52m}$Co were measured for the first time with an accuracy of $\sim 10$ keV, an unprecedented precision reached for short-lived nuclei in the isochronous mass spectrometry. Combining our results with the previous $\beta$-$\gamma$ measurements of $^{52}$Ni, the $T=2$, $J^{\pi}=0^+$ isobaric analog state (IAS) in $^{52}$Co was newly assigned, questioning the conventional identification of IASs from the $\beta$-delayed proton emissions. Using our energy of the IAS in $^{52}$Co, the masses of the $T=2$ multiplet fit well into the Isobaric Multiplet Mass Equation. Read More

We consider exciton-photon coupling in semiconductor microcavities in which separate periodic potentials have been embedded for excitons and photons. We show theoretically that this system supports degenerate ground-states appearing at non-zero in-plane momenta, corresponding to multiple valleys in reciprocal space, which are further separated in polarization corresponding to a polarization-valley coupling in the system. Aside forming a basis for valleytronics, the multivalley dispersion is predicted to allow for spontaneous momentum symmetry breaking and two-mode squeezing under non-resonant and resonant excitation, respectively. Read More

The hyperluminous X-ray source (HLX-1, the peak X-ray luminosity $\sim 10^{42}\rm erg\ s^{-1}$) near the spiral galaxy ESO 243-49 is possibly the best candidate for intermediate mass black hole (IMBH), which underwent recurrent outbursts with a period of $\sim 400$ days. The physical reason for this quasi-periodic variability is still unclear. We explore the possibility of radiation-pressure instability in accretion disk by modeling the light curve of HLX-1, and find that it can roughly reproduce the duration, period and amplitude of the recurrent outbursts HLX-1 with an IMBH of ~10^5Msun. Read More

Convolutional Neural Networks (CNN) have demon- strated its successful applications in computer vision, speech recognition, and natural language processing. For object recog- nition, CNNs might be limited by its strict label requirement and an implicit assumption that images are supposed to be target- object-dominated for optimal solutions. However, the labeling procedure, necessitating laying out the locations of target ob- jects, is very tedious, making high-quality large-scale dataset prohibitively expensive. Read More

Motivated by order bounds for algorithms for algebraic differential equations (for example, effective differential elimination and Nullstellensatz), representing the radical of a given polynomial ideal or, equivalently, the corresponding affine variety, using triangular sets is of primary interest. An algorithm to construct such a representation was proposed by A. Szanto. Read More

Entity images could provide significant visual information for knowledge representation learning. Most conventional methods learn knowledge representations merely from structured triples, ignoring rich visual information extracted from entity images. In this paper, we propose a novel Image-embodied Knowledge Representation Learning model (IKRL), where knowledge representations are learned with both triple facts and images. Read More

Single feature is inefficient to describe content of an image, which is a shortcoming in traditional image retrieval task. We know that one image can be described by different features. Multi-feature fusion ranking can be utilized to improve the ranking list of query. Read More

Distantly supervised relation extraction has been widely used to find novel relational facts from plain text. To predict the relation between a pair of two target entities, existing methods solely rely on those direct sentences containing both entities. In fact, there are also many sentences containing only one of the target entities, which provide rich and useful information for relation extraction. Read More

Textual information is considered as significant supplement to knowledge representation learning (KRL). There are two main challenges for constructing knowledge representations from plain texts: (1) How to take full advantages of sequential contexts of entities in plain texts for KRL. (2) How to dynamically select those informative sentences of the corresponding entities for KRL. Read More

Manipulating the self-assembly nanostructures with combined different control measures is emerging as a promising route for numerous applications to generate templates and scaffolds for nanostructured materials. Here, the two different control measures are a cylindrical confinement and an oscillatory shear flow. We study the phase behavior of diblock copolymer confined in nanopore under oscillatory shear by considering different $D/L_0$ ($D$ is the diameter of the cylindrical nanopore, $L_0$ is the domain spacing) and different shears via Cell Dynamics Simulation. Read More

Hybrid systems exhibit both continuous and discrete behavior. Analyzing hybrid systems is known to be hard. Inspired by the idea of concolic testing (of programs), we investigate whether we can combine random sampling and symbolic execution in order to effectively verify hybrid systems. Read More

We perform long-term ($\approx 15$ yr, observed-frame) X-ray variability analyses of the 68 brightest radio-quiet active galactic nuclei (AGNs) in the 6 Ms $Chandra$ Deep Field-South (CDF-S) survey; the majority are in the redshift range of $0.6-3.1$, providing access to penetrating rest-frame X-rays up to $\approx 10-30$ keV. Read More

A great video title describes the most salient event compactly and captures the viewer's attention. In contrast, video captioning tends to generate sentences that describe the video as a whole. Although generating a video title automatically is a very useful task, it is much less addressed than video captioning. Read More

Neural models have recently been used in text summarization including headline generation. The model can be trained using a set of document-headline pairs. However, the model does not explicitly consider topical similarities and differences of documents. Read More

Graph mining to extract interesting components has been studied in various guises, e.g., communities, dense subgraphs, cliques. Read More

A residual-networks family with hundreds or even thousands of layers dominates major image recognition tasks, but building a network by simply stacking residual blocks inevitably limits its optimization ability. This paper proposes a novel residual-network architecture, Residual networks of Residual networks (RoR), to dig the optimization ability of residual networks. RoR substitutes optimizing residual mapping of residual mapping for optimizing original residual mapping. Read More

HLX-1, currently the best intermediate-mass black hole candidate, has undergone seven violent outbursts, each with a peak X-ray luminosity of $L_{\mathrm{peak},\mathrm{X}}\sim 10^{42}\ \rm{erg\ s^{-1}}$. Interestingly, the properties of the HLX-1 outbursts evolve with time. In this work, we aim to constrain the physical parameters of the central engine of the HLX-1 outbursts in the framework of the black hole accretion. Read More

As an alternative to conventional multi-pixel cameras, single-pixel cameras enable images to be recorded using a single detector that measures the correlations between the scene and a set of patterns. However, to fully sample a scene in this way requires at least the same number of correlation measurements as there are pixels in the reconstructed image. Therefore single-pixel imaging systems typically exhibit low frame-rates. Read More

The accelerated growth of mobile trajectories in location-based services brings valuable data resources to understand users' moving behaviors. Apart from recording the trajectory data, another major characteristic of these location-based services is that they also allow the users to connect whomever they like. A combination of social networking and location-based services is called as location-based social networks (LBSN). Read More

While end-to-end neural machine translation (NMT) has made remarkable progress recently, NMT systems only rely on parallel corpora for parameter estimation. Since parallel corpora are usually limited in quantity, quality, and coverage, especially for low-resource languages, it is appealing to exploit monolingual corpora to improve NMT. We propose a semi-supervised approach for training NMT models on the concatenation of labeled (parallel corpora) and unlabeled (monolingual corpora) data. Read More

We introduce an agreement-based approach to learning parallel lexicons and phrases from non-parallel corpora. The basic idea is to encourage two asymmetric latent-variable translation models (i.e. Read More