arXiv subjects

Yiming Xu

Publications and source records attributed to Yiming Xu.

At least 109 records · Page 6Linked to original sources

arXiv2021

Randomized weakly admissible meshes

A weakly admissible mesh (WAM) on a continuum real-valued domain is a sequence of discrete grids such that the discrete maximum norm of polynomials on the grid is comparable to the supremum norm of polynomials on the domain. The asymptotic rate of growth of the grid sizes and of the comparability constants must grow in a controlled manner. In this paper, we generalize the notion of a WAM to a hierarchical subspaces of not necessarily polynomial functions, and we analyze particular strategies for random sampling as a technique for generating WAM's. Our main results show that WAM's and their stronger variant, admissible meshes (AM's), can be generated by random sampling, and our analysis provides concrete estimates for growth of both the meshes and the discrete-continuum comparability constants.

math.NA↗

arXiv2021

Estimating Wildfire Evacuation Decision and Departure Timing Using Large-Scale GPS Data

With increased frequency and intensity due to climate change, wildfires have become a growing global concern. This creates severe challenges for fire and emergency services as well as communities in the wildland-urban interface (WUI). To reduce wildfire risk and enhance the safety of WUI communities, improving our understanding of wildfire evacuation is a pressing need. To this end, this study proposes a new methodology to analyze human behavior during wildfires by leveraging a large-scale GPS dataset. This methodology includes a home-location inference algorithm and an evacuation-behavior inference algorithm, to systematically identify different groups of wildfire evacuees (i.e., self-evacuee, shadow evacuee, evacuee under warning, and ordered evacuee). We applied the methodology to the 2019 Kincade Fire in Sonoma County, CA. We found that among all groups of evacuees, self-evacuees and shadow evacuees accounted for more than half of the evacuees during the Kincade Fire. The results also show that inside of the evacuation warning/order zones, the total evacuation compliance rate was around 46% among all the categorized people. The findings of this study can be used by emergency managers and planners to better target public outreach campaigns, training protocols, and emergency communication strategies to prepare WUI households for future wildfire events.

cs.CY↗

arXiv2021

Do e-scooters fill mobility gaps and promote equity before and during COVID-19? A spatiotemporal analysis using open big data

The growing popularity of e-scooters and their rapid expansion across urban streets has attracted widespread attention. A major policy question is whether e-scooters substitute existing mobility options or fill the service gaps left by them. This study addresses this question by analyzing the spatiotemporal patterns of e-scooter service availability and use in Washington DC, focusing on their spatial relationships with public transit and bikesharing. Results from an analysis of three open big datasets suggest that e-scooters have both competing and complementary effects on transit and bikesharing services. The supply of e-scooters significantly overlaps with the service areas of transit and bikesharing, and we classify a majority of e-scooter trips as substitutes to transit and bikesharing uses. A travel-time-based analysis further reveals that when choosing e-scooters over transit, travelers pay a price premium and save some travel time. The price premium is greater during the COVID-19 pandemic but the associated travel-time savings are smaller. This implies that public health considerations rather than time-cost tradeoffs are the main driver for many to choose e-scooters over transit during COVID. In addition, we find that e-scooters complement bikesharing and transit by providing services to underserved neighborhoods. A sizeable proportion (about 10 percent) of e-scooter trips are taken to connect with the rail services. Future research may combine the big-data-based analysis presented here with traditional methods to further shed light on the interactions between e-scooter services, bikesharing, and public transit.

cs.CY↗

arXiv2021

Analysis of The Ratio of $\ell_1$ and $\ell_2$ Norms in Compressed Sensing

We first propose a novel criterion that guarantees that an $s$-sparse signal is the local minimizer of the $\ell_1/\ell_2$ objective; our criterion is interpretable and useful in practice. We also give the first uniform recovery condition using a geometric characterization of the null space of the measurement matrix, and show that this condition is easily satisfied for a class of random matrices. We also present analysis on the robustness of the procedure when noise pollutes data. Numerical experiments are provided that compare $\ell_1/\ell_2$ with some other popular non-convex methods in compressed sensing. Finally, we propose a novel initialization approach to accelerate the numerical optimization procedure. We call this initialization approach \emph{support selection}, and we demonstrate that it empirically improves the performance of existing $\ell_1/\ell_2$ algorithms.

math.NA↗

arXiv2021

Infections Forecasting and Intervention Effect Evaluation for COVID-19 via a Data-Driven Markov Process and Heterogeneous Simulation

The Coronavirus Disease 2019 (COVID-19) pandemic has caused tremendous amount of deaths and a devastating impact on the economic development all over the world. Thus, it is paramount to control its further transmission, for which purpose it is necessary to find the mechanism of its transmission process and evaluate the effect of different control strategies. To deal with these issues, we describe the transmission of COVID-19 as an explosive Markov process with four parameters. The state transitions of the proposed Markov process can clearly disclose the terrible explosion and complex heterogeneity of COVID-19. Based on this, we further propose a simulation approach with heterogeneous infections. Experimentations show that our approach can closely track the real transmission process of COVID-19, disclose its transmission mechanism, and forecast the transmission under different non-drug intervention strategies. More importantly, our approach can helpfully develop effective strategies for controlling COVID-19 and appropriately compare their control effect in different countries/cities.

stat.AP↗

arXiv2020

Open Set Domain Adaptation by Extreme Value Theory

Common domain adaptation techniques assume that the source domain and the target domain share an identical label space, which is problematic since when target samples are unlabeled we have no knowledge on whether the two domains share the same label space. When this is not the case, the existing methods fail to perform well because the additional unknown classes are also matched with the source domain during adaptation. In this paper, we tackle the open set domain adaptation problem under the assumption that the source and the target label spaces only partially overlap, and the task becomes when the unknown classes exist, how to detect the target unknown classes and avoid aligning them with the source domain. We propose to utilize an instance-level reweighting strategy for domain adaptation where the weights indicate the likelihood of a sample belonging to known classes and to model the tail of the entropy distribution with Extreme Value Theory for unknown class detection. Experiments on conventional domain adaptation datasets show that the proposed method outperforms the state-of-the-art models.

stat.ML↗

arXiv2020

Concept Drift and Covariate Shift Detection Ensemble with Lagged Labels

In model serving, having one fixed model during the entire often life-long inference process is usually detrimental to model performance, as data distribution evolves over time, resulting in lack of reliability of the model trained on historical data. It is important to detect changes and retrain the model in time. The existing methods generally have three weaknesses: 1) using only classification error rate as signal, 2) assuming ground truth labels are immediately available after features from samples are received and 3) unable to decide what data to use to retrain the model when change occurs. We address the first problem by utilizing six different signals to capture a wide range of characteristics of data, and we address the second problem by allowing lag of labels, where labels of corresponding features are received after a lag in time. For the third problem, our proposed method automatically decides what data to use to retrain based on the signals. Extensive experiments on structured and unstructured data for different type of data changes establish that our method consistently outperforms the state-of-the-art methods by a large margin.

cs.AI↗

arXiv2020

Micromobility Trip Origin and Destination Inference Using General Bikeshare Feed Specification (GBFS) Data

Emerging micromobility services (e.g., e-scooters) have a great potential to enhance urban mobility but more knowledge on their usage patterns is needed. The General Bikeshare Feed Specification (GBFS) data are a possible source for examining micromobility trip patterns, but efforts are needed to infer trips from the GBFS data. Existing trip inference methods are usually based upon the assumption that the vehicle ID of a micromobility option (e-scooter or e-bike) does not change, and so they cannot deal with data with vehicle IDs that change over time. In this study, we propose a comprehensive package of algorithms to infer trip origins and destinations from GBFS data with different types of vehicle ID. We implement the algorithms in Washington DC by analyzing one-week (last week of February 2020) of GBFS data published by six vendors, and we evaluate the inference accuracy of the proposed algorithms by R-squared, mean absolute error, and sum absolute error. We find that the R-squared measure is larger than 0.9 and the MAE measure is smaller than 2 when the algorithms are evaluated with a 400m*400m grid, and the absolute errors are relatively larger in the downtown area. The accuracy of the trip-inference algorithms is sufficiently high for most practical applications.

cs.CY↗

arXiv2020

Consistency of archetypal analysis

Archetypal analysis is an unsupervised learning method that uses a convex polytope to summarize multivariate data. For fixed $k$, the method finds a convex polytope with $k$ vertices, called archetype points, such that the polytope is contained in the convex hull of the data and the mean squared distance between the data and the polytope is minimal. In this paper, we prove a consistency result that shows if the data is independently sampled from a probability measure with bounded support, then the archetype points converge to a solution of the continuum version of the problem, of which we identify and establish several properties. We also obtain the convergence rate of the optimal objective values under appropriate assumptions on the distribution. If the data is independently sampled from a distribution with unbounded support, we also prove a consistency result for a modified method that penalizes the dispersion of the archetype points. Our analysis is supported by detailed computational experiments of the archetype points for data sampled from the uniform distribution in a disk, the normal distribution, an annular distribution, and a Gaussian mixture model.

math.ST↗

arXiv2019

k-Nearest Neighbors by Means of Sequence to Sequence Deep Neural Networks and Memory Networks

k-Nearest Neighbors is one of the most fundamental but effective classification models. In this paper, we propose two families of models built on a sequence to sequence model and a memory network model to mimic the k-Nearest Neighbors model, which generate a sequence of labels, a sequence of out-of-sample feature vectors and a final label for classification, and thus they could also function as oversamplers. We also propose 'out-of-core' versions of our models which assume that only a small portion of data can be loaded into memory. Computational experiments show that our models on structured datasets outperform k-Nearest Neighbors, a feed-forward neural network, XGBoost, lightGBM, random forest and a memory network, due to the fact that our models must produce additional output and not just the label. On image and text datasets, the performance of our model is close to many state-of-the-art deep models. As an oversampler on imbalanced datasets, the sequence to sequence kNN model often outperforms Synthetic Minority Over-sampling Technique and Adaptive Synthetic Sampling.

cs.LG↗

arXiv2019

Open-Ended Visual Question Answering by Multi-Modal Domain Adaptation

We study the problem of visual question answering (VQA) in images by exploiting supervised domain adaptation, where there is a large amount of labeled data in the source domain but only limited labeled data in the target domain with the goal to train a good target model. A straightforward solution is to fine-tune a pre-trained source model by using those limited labeled target data, but it usually cannot work well due to the considerable difference between the data distributions of the source and target domains. Moreover, the availability of multiple modalities (i.e., images, questions and answers) in VQA poses further challenges to model the transferability between those different modalities. In this paper, we tackle the above issues by proposing a novel supervised multi-modal domain adaptation method for VQA to learn joint feature embeddings across different domains and modalities. Specifically, we align the data distributions of the source and target domains by considering all modalities together as well as separately for each individual modality. Based on the extensive experiments on the benchmark VQA 2.0 and VizWiz datasets for the realistic open-ended VQA task, we demonstrate that our proposed method outperforms the existing state-of-the-art approaches in this challenging domain adaptation setting for VQA.

cs.CV↗

arXiv2019

Automatic Ontology Learning from Domain-Specific Short Unstructured Text Data

Ontology learning is a critical task in industry, dealing with identifying and extracting concepts captured in text data such that these concepts can be used in different tasks, e.g. information retrieval. Ontology learning is non-trivial due to several reasons with limited amount of prior research work that automatically learns a domain specific ontology from data. In our work, we propose a two-stage classification system to automatically learn an ontology from unstructured text data. We first collect candidate concepts, which are classified into concepts and irrelevant collocates by our first classifier. The concepts from the first classifier are further classified by the second classifier into different concept types. The proposed system is deployed as a prototype at a company and its performance is validated by using complaint and repair verbatim data collected in automotive industry from different data sources.

cs.IR↗

arXiv2019

A setup for extreme-ultraviolet ultrafast angle-resolved photoelectron spectroscopy at 50-kHz repetition rate

Time- and angle-resolved photoelectron spectroscopy (trARPES) is a powerful method to track the ultrafast dynamics of quasiparticles and electronic bands in energy and momentum space. We present a setup for trARPES with 22.3 eV extreme-ultraviolet (XUV) femtosecond pulses at 50-kHz repetition rate, which enables fast data acquisition and access to dynamics across momentum space with high sensitivity. The design and operation of the XUV beamline, pump-probe setup, and UHV endstation are described in detail. By characterizing the effect of space-charge broadening, we determine an ultimate source-limited energy resolution of 60 meV, with typically 80-100 meV obtained at 1-2e10 photons/s probe flux on the sample. The instrument capabilities are demonstrated via both equilibrium and time-resolved ARPES studies of transition-metal dichalcogenides. The 50-kHz repetition rate enables sensitive measurements of quasiparticles at low excitation fluences in semiconducting MoSe$_2$, with an instrumental time resolution of 65 fs. Moreover, photo-induced phase transitions can be driven with the available pump fluence, as shown by charge density wave melting in 1T-TiSe$_2$. The high repetition-rate setup thus provides a versatile platform for sensitive XUV trARPES, from quenching of electronic phases down to the perturbative limit.

physics.ins-det↗

arXiv2018

Low Rank Variation Dictionary and Inverse Projection Group Sparse Representation Model for Breast Tumor Classification

Sparse representation classification achieves good results by addressing recognition problem with sufficient training samples per subject. However, SRC performs not very well for small sample data. In this paper, an inverse-projection group sparse representation model is presented for breast tumor classification, which is based on constructing low-rank variation dictionary. The proposed low-rank variation dictionary tackles tumor recognition problem from the viewpoint of detecting and using variations in gene expression profiles of normal and patients, rather than directly using these samples. The inverse projection group sparsity representation model is constructed based on taking full using of exist samples and group effect of microarray gene data. Extensive experiments on public breast tumor microarray gene expression datasets demonstrate the proposed technique is competitive with state-of-the-art methods. The results of Breast-1, Breast-2 and Breast-3 databases are 80.81%, 89.10% and 100% respectively, which are better than the latest literature.

cs.CV↗

arXiv2017

Bright high-repetition-rate source of narrowband extreme-ultraviolet harmonics beyond 22 eV

Novel table-top sources of extreme-ultraviolet light based on high-harmonic generation yield unique insight into the fundamental properties of molecules, nanomaterials, or correlated solids, and enable advanced applications in imaging or metrology. Extending high-harmonic generation to high repetition rates portends great experimental benefits, yet efficient extreme-ultraviolet conversion of correspondingly weak driving pulses is challenging. Here, we demonstrate a highly efficient source of femtosecond extreme-ultraviolet pulses at 50-kHz repetition rate, utilizing the ultraviolet second-harmonic focused tightly into Kr gas. In this cascaded scheme, a photon flux beyond ~3e13 per second is generated at 22.3 eV, with 5e-5 conversion efficiency that surpasses similar harmonics directly driven by the fundamental by two orders of magnitude. The enhancement arises from both wavelength scaling of the atomic dipole and improved spatio-temporal phase-matching, confirmed by simulations. Spectral isolation of a single 72-meV wide harmonic renders this bright, 50-kHz extreme-ultraviolet source a powerful tool for ultrafast photoemission, nanoscale imaging and other applications.

physics.optics↗

arXiv2017

Cooperative Transmission for Physical Layer Security by Exploring Social Awareness

Social awareness and social ties are becoming increasingly fashionable with emerging mobile and handheld devices. Social trust degree describing the strength of the social ties has drawn lots of research interests in many fields including secure cooperative communications. Such trust degree reflects the users' willingness for cooperation, which impacts the selection of the cooperative users in the practical networks. In this paper, we propose a cooperative relay and jamming selection scheme to secure communication based on the social trust degree under a stochastic geometry framework. We aim to analyze the involved secrecy outage probability (SOP) of the system's performance. To achieve this target, we propose a double Gamma ratio (DGR) approach through Gamma approximation. Based on this, the SOP is tractably obtained in closed form. The simulation results verify our theoretical findings, and validate that the social trust degree has dramatic influences on the network's secrecy performance.

cs.IT↗

arXiv2014

A solution of 2D QCD at Finite $N$ using a conformal basis

We study 2D QCD with a fundamental fermion at small-$N$ using the recently proposed conformal basis approach. We find that effective conformal dominance still holds, namely that the spectrum converges efficiently, with high scaling-dimension operators decoupling exponentially quickly from the stable single-particle states. Consequently, for these stable bound states, accurate, analytic expressions for wavefunctions and parton distribution functions can be given, even for $N=3$.

hep-th↗

arXiv2014

Seeking Lorentz Violation from the Higgs

The recently discovered Higgs particle with a mass near $126$ GeV presents new opportunities to explore Lorentz violation. Ultra-high-energy cosmic rays are one of the most sensitive testing grounds for Lorentz symmetry, and can be used to seek for and limit departures from Lorentz invariance in the Higgs sector. If the Higgs were to have a super- or sub-luminal maximal speed both Higgs and weak interaction physics would be modified. Consideration of such modifications allow us to constrain the Higgs maximal velocity to agree with that of other Standard Model particles to parts in $10^{14}$.

hep-ph↗