arXiv subjects

Xinyi Zhang

Publications and source records attributed to Xinyi Zhang.

At least 127 records · Page 7Linked to original sources

arXiv2021

LA-HCN: Label-based Attention for Hierarchical Multi-label TextClassification Neural Network

Hierarchical multi-label text classification (HMTC) has been gaining popularity in recent years thanks to its applicability to a plethora of real-world applications. The existing HMTC algorithms largely focus on the design of classifiers, such as the local, global, or a combination of them. However, very few studies have focused on hierarchical feature extraction and explore the association between the hierarchical labels and the text. In this paper, we propose a Label-based Attention for Hierarchical Mutlti-label Text Classification Neural Network (LA-HCN), where the novel label-based attention module is designed to hierarchically extract important information from the text based on the labels from different hierarchy levels. Besides, hierarchical information is shared across levels while preserving the hierarchical label-based information. Separate local and global document embeddings are obtained and used to facilitate the respective local and global classifications. In our experiments, LA-HCN outperforms other state-of-the-art neural network-based HMTC algorithms on four public HMTC datasets. The ablation study also demonstrates the effectiveness of the proposed label-based attention module as well as the novel local and global embeddings and classifications. By visualizing the learned attention (words), we find that LA-HCN is able to extract meaningful information corresponding to the different labels which provides explainability that may be helpful for the human analyst.

cs.CL↗

arXiv2021

mSHINE: A Multiple-meta-paths Simultaneous Learning Framework for Heterogeneous Information Network Embedding

Heterogeneous information networks(HINs) become popular in recent years for its strong capability of modelling objects with abundant information using explicit network structure. Network embedding has been proved as an effective method to convert information networks into lower-dimensional space, whereas the core information can be well preserved. However, traditional network embedding algorithms are sub-optimal in capturing rich while potentially incompatible semantics provided by HINs. To address this issue, a novel meta-path-based HIN representation learning framework named mSHINE is designed to simultaneously learn multiple node representations for different meta-paths. More specifically, one representation learning module inspired by the RNN structure is developed and multiple node representations can be learned simultaneously, where each representation is associated with one respective meta-path. By measuring the relevance between nodes with the designed objective function, the learned module can be applied in downstream link prediction tasks. A set of criteria for selecting initial meta-paths is proposed as the other module in mSHINE which is important to reduce the optimal meta-path selection cost when no prior knowledge of suitable meta-paths is available. To corroborate the effectiveness of mSHINE, extensive experimental studies including node classification and link prediction are conducted on five real-world datasets. The results demonstrate that mSHINE outperforms other state-of-the-art HIN embedding methods.

cs.LG↗

arXiv2021

Natural inflation with a nonminimal coupling to gravity

Although natural inflation is a theoretically well-motivated model for cosmic inflation, it is in tension with recent Planck cosmic microwave background anisotropy measurements. We present a way to alleviate this tension by considering a very weak nonminimal coupling of the inflaton field to gravity in both contexts of metric and Palatini formulations of general relativity. We start our discussions with a generic form of the inflaton coupling to the Ricci scalar, then focus on a simple form to do phenomenological study. Our results show that such an extension can bring natural inflation's predictions to a good agreement with the Planck data. Depending on values of the coupling constant $ξ$ and the symmetry breaking scale $f$, we find that with $|ξ|\sim 10^{-3}$ and $f\gtrsim 2.0 M_{\mathrm{pl}}$ predictions of the model stay inside $68\%$ CL allowed region until $f$ increases up to $7.7 M_{\mathrm{pl}}$, then only inside $95\%$ CL region after $f$ exceeds the latter value. The predictions from the metric and the Palatini theories are very similar due to the simple form of the coupling function we use and the small magnitude of the coupling $ξ$. Successful reheating can also be realized in this model.

astro-ph.CO↗

arXiv2021

Baconian: A Unified Open-source Framework for Model-Based Reinforcement Learning

Model-Based Reinforcement Learning (MBRL) is one category of Reinforcement Learning (RL) algorithms which can improve sampling efficiency by modeling and approximating system dynamics. It has been widely adopted in the research of robotics, autonomous driving, etc. Despite its popularity, there still lacks some sophisticated and reusable open-source frameworks to facilitate MBRL research and experiments. To fill this gap, we develop a flexible and modularized framework, Baconian, which allows researchers to easily implement a MBRL testbed by customizing or building upon our provided modules and algorithms. Our framework can free users from re-implementing popular MBRL algorithms from scratch thus greatly save users' efforts on MBRL experiments.

cs.LG↗

arXiv2020

HEU Emotion: A Large-scale Database for Multi-modal Emotion Recognition in the Wild

The study of affective computing in the wild setting is underpinned by databases. Existing multimodal emotion databases in the real-world conditions are few and small, with a limited number of subjects and expressed in a single language. To meet this requirement, we collected, annotated, and prepared to release a new natural state video database (called HEU Emotion). HEU Emotion contains a total of 19,004 video clips, which is divided into two parts according to the data source. The first part contains videos downloaded from Tumblr, Google, and Giphy, including 10 emotions and two modalities (facial expression and body posture). The second part includes corpus taken manually from movies, TV series, and variety shows, consisting of 10 emotions and three modalities (facial expression, body posture, and emotional speech). HEU Emotion is by far the most extensive multi-modal emotional database with 9,951 subjects. In order to provide a benchmark for emotion recognition, we used many conventional machine learning and deep learning methods to evaluate HEU Emotion. We proposed a Multi-modal Attention module to fuse multi-modal features adaptively. After multi-modal fusion, the recognition accuracies for the two parts increased by 2.19% and 4.01% respectively over those of single-modal facial expression recognition.

cs.CV↗

arXiv2020

TESS Asteroseismic Analysis of the Known Exoplanet Host Star HD 222076

The Transiting Exoplanet Survey Satellite (TESS) is an all-sky survey mission aiming to search for exoplanets that transit bright stars. The high-quality photometric data of TESS are excellent for the asteroseismic study of solar-like stars. In this work, we present an asteroseismic analysis of the red-giant star HD~222076 hosting a long-period (2.4 yr) giant planet discovered through radial velocities. Solar-like oscillations of HD~222076 are detected around $203 \, μ$Hz by TESS for the first time. Asteroseismic modeling, using global asteroseismic parameters as input, yields a determination of the stellar mass ($M_\star = 1.12 \pm 0.12\, M_\odot$), radius ($R_\star = 4.34 \pm 0.21\,R_\odot$), and age ($7.4 \pm 2.7\,$Gyr), with precisions greatly improved from previous studies. The period spacing of the dipolar mixed modes extracted from the observed power spectrum reveals that the star is on the red-giant branch burning hydrogen in a shell surrounding the core. We find that the planet will not escape the tidal pull of the star and be engulfed into it within about $800\,$Myr, before the tip of the red-giant branch is reached.

astro-ph.SR↗

arXiv2020

Multi-Scale Boosted Dehazing Network with Dense Feature Fusion

In this paper, we propose a Multi-Scale Boosted Dehazing Network with Dense Feature Fusion based on the U-Net architecture. The proposed method is designed based on two principles, boosting and error feedback, and we show that they are suitable for the dehazing problem. By incorporating the Strengthen-Operate-Subtract boosting strategy in the decoder of the proposed model, we develop a simple yet effective boosted decoder to progressively restore the haze-free image. To address the issue of preserving spatial information in the U-Net architecture, we design a dense feature fusion module using the back-projection feedback scheme. We show that the dense feature fusion module can simultaneously remedy the missing spatial information from high-resolution features and exploit the non-adjacent features. Extensive evaluations demonstrate that the proposed model performs favorably against the state-of-the-art approaches on the benchmark datasets as well as real-world hazy images.

cs.CV↗

arXiv2020

Gated Fusion Network for Degraded Image Super Resolution

Single image super resolution aims to enhance image quality with respect to spatial content, which is a fundamental task in computer vision. In this work, we address the task of single frame super resolution with the presence of image degradation, e.g., blur, haze, or rain streaks. Due to the limitations of frame capturing and formation processes, image degradation is inevitable, and the artifacts would be exacerbated by super resolution methods. To address this problem, we propose a dual-branch convolutional neural network to extract base features and recovered features separately. The base features contain local and global information of the input image. On the other hand, the recovered features focus on the degraded regions and are used to remove the degradation. Those features are then fused through a recursive gate module to obtain sharp features for super resolution. By decomposing the feature extraction step into two task-independent streams, the dual-branch model can facilitate the training process by avoiding learning the mixed degradation all-in-one and thus enhance the final high-resolution prediction results. We evaluate the proposed method in three degradation scenarios. Experiments on these scenarios demonstrate that the proposed method performs more efficiently and favorably against the state-of-the-art approaches on benchmark datasets.

cs.CV↗

arXiv2020

Achieving Super-Linear Speedup across Multi-FPGA for Real-Time DNN Inference

Real-time Deep Neural Network (DNN) inference with low-latency requirement has become increasingly important for numerous applications in both cloud computing (e.g., Apple's Siri) and edge computing (e.g., Google/Waymo's driverless car). FPGA-based DNN accelerators have demonstrated both superior flexibility and performance; in addition, for real-time inference with low batch size, FPGA is expected to achieve further performance improvement. However, the performance gain from the single-FPGA design is obstructed by the limited on-chip resource. In this paper, we employ multiple FPGAs to cooperatively run DNNs with the objective of achieving super-linear speed-up against single-FPGA design. In implementing such systems, we found two barriers that hinder us from achieving the design goal: (1) the lack of a clear partition scheme for each DNN layer to fully exploit parallelism, and (2) the insufficient bandwidth between the off-chip memory and the accelerator due to the growing size of DNNs. To tackle these issues, we propose a general framework, "Super-LIP", which can support different kinds of DNNs. In this paper, we take Convolutional Neural Network (CNN) as a vehicle to illustrate Super-LIP. We first formulate an accurate system-level model to support the exploration of best partition schemes. Then, we develop a novel design methodology to effectively alleviate the heavy loads on memory bandwidth by moving traffic from memory bus to inter-FPGA links. We implement Super-LIP based on ZCU102 FPGA boards. Results demonstrate that Super-LIP with 2 FPGAs can achieve 3.48x speedup, compared to the state-of-the-art single-FPGA design. What is more, as the number of FPGAs scales up, the system latency can be further reduced while maintaining high energy efficiency.

cs.DC↗

arXiv2020

Low Overhead Online Data Flow Tracking for Intermittently Powered Non-volatile FPGAs

Energy harvesting is an attractive way to power future IoT devices since it can eliminate the need for battery or power cables. However, harvested energy is intrinsically unstable. While FPGAs have been widely adopted in various embedded systems, it is hard to survive unstable power since all the memory components in FPGA are based on volatile SRAMs. The emerging non-volatile memory based FPGAs provide promising potentials to keep configuration data on the chip during power outages. Few works have considered implementing efficient runtime intermediate data checkpoint on non-volatile FPGAs. To realize accumulative computation under intermittent power on FPGA, this paper proposes a low-cost design framework, Data-Flow-Tracking FPGA (DFT-FPGA), which utilizes binary counters to track intermediate data flow. Instead of keeping all on-chip intermediate data, DFT-FPGA only targets on necessary data that is labeled by off-line analysis and identified by an online tracking system. The evaluation shows that compared with state-of-the-art techniques, DFT-FPGA can realize accumulative computing with less off-line workload and significantly reduce online roll-back time and resource utilization.

cs.AR↗

arXiv2019

RVH: Range-Vector Hash for Fast Online Packet Classification

Packet classification according to multi-field ruleset is a key component for many network applications. Emerging software defined networking and cloud computing need to update the rulesets frequently for flexible policy configuration. Their success depends on the availability of the new generation of classifiers that can support both fast ruleset updating and high-speed packet classification. However, existing packet classification approaches focus either on high-speed packet classification or fast rule update, but no known scheme meets both requirements. In this paper, we propose Range-vector Hash (RVH) to effectively accelerate the packet classification with a hash-based algorithm while ensuring the fast rule update. RVH is built on our key observation that the number of distinct combinations of each field prefix lengths is not evenly distributed. To reduce the number of hash tables for fast classification, we introduce a novel concept range-vector with each specified the length range of each field prefix of the projected rules. RVH can overcome the major obstacle that hinders hash-based packet classification by balancing the number of hash tables and the probability of hash collision. Experimental results demonstrate that RVH can achieve the classification speed up to 15.7 times and the update speed up to 2.3 times that of the state-of-the-art algorithms on average, while only consuming 44% less memory.

cs.NI↗

arXiv2019

Testing Lorentz invariance and CPT symmetry using gamma-ray burst neutrinos

A recent work [Y. Huang and B.-Q. Ma, Commun. Phys. {\bf 1}, 62 (2018)] associated all four PeV neutrinos observed by IceCube to gamma-ray bursts (GRBs), and revealed a regularity which indicates a Lorentz violation scale $E_{\rm LV}=(6.5\pm0.4)\times10^{17}$ GeV with opposite sign factors $s=\pm 1$ between neutrinos and antineutrinos. The association of "time delay" and "time advance" events with neutrinos and antineutrinos (or vice versa) is only a hypothesis since the IceCube detector cannot tell the chirality of the neutrinos, and further experimental tests are needed to verify this hypothesis. We derive the values of the CPT-odd Lorentz violating parameters in the standard-model extension (SME) framework, and perform a threshold analysis on the electron-positron pair emission of the superluminal neutrinos (or antineutrinos). We find that different neutrino/antineutrino propagation properties, suggested by Y. Huang and B.-Q. Ma, can be described in the SME framework with both Lorentz invariance and CPT symmetry violation, but with a threshold energy constraint. A viable way on testing the CPT symmetry violation between neutrinos and antineutrinos is suggested.

hep-ph↗

arXiv2019

Accuracy vs. Efficiency: Achieving Both through FPGA-Implementation Aware Neural Architecture Search

A fundamental question lies in almost every application of deep neural networks: what is the optimal neural architecture given a specific dataset? Recently, several Neural Architecture Search (NAS) frameworks have been developed that use reinforcement learning and evolutionary algorithm to search for the solution. However, most of them take a long time to find the optimal architecture due to the huge search space and the lengthy training process needed to evaluate each candidate. In addition, most of them aim at accuracy only and do not take into consideration the hardware that will be used to implement the architecture. This will potentially lead to excessive latencies beyond specifications, rendering the resulting architectures useless. To address both issues, in this paper we use Field Programmable Gate Arrays (FPGAs) as a vehicle to present a novel hardware-aware NAS framework, namely FNAS, which will provide an optimal neural architecture with latency guaranteed to meet the specification. In addition, with a performance abstraction model to analyze the latency of neural architectures without training, our framework can quickly prune architectures that do not satisfy the specification, leading to higher efficiency. Experimental results on common data set such as ImageNet show that in the cases where the state-of-the-art generates architectures with latencies 7.81x longer than the specification, those from FNAS can meet the specs with less than 1% accuracy loss. Moreover, FNAS also achieves up to 11.13x speedup for the search process. To the best of the authors' knowledge, this is the very first hardware aware NAS.

cs.DC↗

arXiv2018

On the Radical of Multiperfect Numbers and Applications

It is conjectured that for a perfect number $m,$ $\rm{rad}(m)\ll m^{\frac{1}{2}}.$ We prove bounds on the radical of multiperfect number $m$ depending on its abundancy index. Assuming the ABC conjecture, we apply this result to study gaps between multiperfect numbers, multiperfect numbers represented by polynomials. Finally, we prove that there are only finitely many multiperfect multirepdigit numbers in any base $g$ where the number of digits in the repdigit is a power of $2.$ This generalizes previous works of several authors including O. Klurman, F. Luca, P. Polack, C. Pomerance and others.

math.NT↗

arXiv2018

DAC-SDC Low Power Object Detection Challenge for UAV Applications

The 55th Design Automation Conference (DAC) held its first System Design Contest (SDC) in 2018. SDC'18 features a lower power object detection challenge (LPODC) on designing and implementing novel algorithms based object detection in images taken from unmanned aerial vehicles (UAV). The dataset includes 95 categories and 150k images, and the hardware platforms include Nvidia's TX2 and Xilinx's PYNQ Z1. DAC-SDC'18 attracted more than 110 entries from 12 countries. This paper presents in detail the dataset and evaluation procedure. It further discusses the methods developed by some of the entries as well as representative results. The paper concludes with directions for future improvements.

cs.CV↗

arXiv2018

Gated Fusion Network for Joint Image Deblurring and Super-Resolution

Single-image super-resolution is a fundamental task for vision applications to enhance the image quality with respect to spatial resolution. If the input image contains degraded pixels, the artifacts caused by the degradation could be amplified by super-resolution methods. Image blur is a common degradation source. Images captured by moving or still cameras are inevitably affected by motion blur due to relative movements between sensors and objects. In this work, we focus on the super-resolution task with the presence of motion blur. We propose a deep gated fusion convolution neural network to generate a clear high-resolution frame from a single natural image with severe blur. By decomposing the feature extraction step into two task-independent streams, the dual-branch design can facilitate the training process by avoiding learning the mixed degradation all-in-one and thus enhance the final high-resolution prediction results. Extensive experiments demonstrate that our method generates sharper super-resolved images from low-resolution inputs with high computational efficiency.

cs.CV↗

arXiv2018

Frequency identification and asteroseismic analysis of the red giant KIC 9145955: fundamental parameters and helium core size

We have analyzed 18 quarters of long-cadence data of KIC 9145955 provided by \emph{Kepler}, and extracted 61 oscillation frequencies from these high precision photometric data. The oscillation frequencies include 7 $l = 0$ modes, 44 $l = 1$ modes, 7 $l = 2$ modes, and 3 $l = 3$ modes. We identify $l = 0$ modes as p modes and $l = 2$ modes as p-dominated modes. For $l = 1$ modes, all of them are identified as mixed modes. These mixed modes can be used to determine the size of the helium core. We conduct a series of asteroseismic models and the size of the helium core is determined to be $M_{\rm He}$ = 0.210 $\pm$ 0.002 $M_{\odot}$ and $R_{\rm He}$ = 0.0307 $\pm$ 0.0002 $R_{\odot}$. Furthermore, we find that only the acoustic radius $τ_{0}$ can be precisely determined with the asteroseismic method independently. The value of $τ_{0}$ is determined to be 0.494 $\pm$ 0.001 days. By combining asteroseismic results and spectroscopic observations, we obtain the best-fitting model. The physical parameters of this model are $M$ = 1.24 $M_{\odot}$, $Z$ = 0.009, $α$ = 2.0, $T_{\rm eff}$ = 5069 K, $\log g$ = 3.029, $R$ = 5.636 $R_{\odot}$, and $L$ = 18.759 $L_{\odot}$. In addition, we think that the observed frequency F39 (96.397 $μ$Hz) is more appropriate to be identified as a mixed mode of the most p-dominated.

astro-ph.SR↗

arXiv2018

An $S_4$ model inspired from self-complementary neutrino mixing

We build an $S_4$ model for neutrino masses and mixings based on the self-complementary (SC) neutrino mixing pattern. The SC mixing is constructed from the self-complementarity relation plus $δ_{\rm CP}=-\fracπ{2}$. We elaborately construct the model at a percent level of accuracy to reproduce the structure given by the SC mixing. After performing a numerical study on the model's parameter space, we find that in the case of normal ordering, the model can give predictions on the observables that are compatible with their $3σ$ ranges, and give predictions for the not-yet observed quantities like the lightest neutrino mass $m_1\in [0.003,0.010]$ eV and the Dirac CP violating phase $δ_{\rm CP}\in[256.72^\circ,283.33^\circ]$.

hep-ph↗