Yilun Zhang - Academia.edu (original) (raw)

Papers by Yilun Zhang

Annals of Operations Research

Faced with dynamic and increasingly diversified public transport requirements, bus operators are ... more Faced with dynamic and increasingly diversified public transport requirements, bus operators are urged to propose operational innovations to sustain their competitiveness. In particular, ordinary bus operations are heavily constrained by well-established route options, and it is challenging to accommodate dynamic passenger flows effectively and with a good level of resource utilization performance. Inspired by the philosophy of sharing economy, many of the available transport resources on the road, such as minibuses and private vehicles, can offer opportunities for improvement if they can be effectively incorporated and exploited. In this regard, this paper proposes a metric learning-based prediction algorithm which can effectively capture the demand pattern and designs a route planning optimizer to help bus operators effectively deploy fixed routing and cooperative buses with traffic dynamics. Through extensive numerical studies, the performance of our proposed metric learning-base...

IEEE Transactions on Intelligent Transportation Systems

arXiv (Cornell University), Mar 15, 2019

Depth estimation is an important capability for autonomous vehicles to understand and reconstruct... more Depth estimation is an important capability for autonomous vehicles to understand and reconstruct 3D environments as well as avoid obstacles during the execution. Accurate depth sensors such as LiDARs are often heavy, expensive and can only provide sparse depth while lighter depth sensors such as stereo cameras are noiser in comparison. We propose an endto-end learning algorithm that is capable of using sparse, noisy input depth for refinement and depth completion. Our model also produces the camera pose as a byproduct, making it a great solution for autonomous systems. We evaluate our approach on both indoor and outdoor datasets. Empirical results show that our method performs well on the KITTI [1] dataset when compared to other competing methods, while having superior performance in dealing with sparse, noisy input depth on the TUM [2] dataset.

2020 IEEE Winter Conference on Applications of Computer Vision (WACV), 2020

This paper presents a semi-supervised keypoint detector by leveraging epipolar geometry and track... more This paper presents a semi-supervised keypoint detector by leveraging epipolar geometry and tracking as supervisionary signals. Our approach precisely detects the keypoints of realworld non-human species such as mouse, monkey, and dogs, for which limited labeled data are available (< 4%) without a pre-trained model.

Sustainable Design and Manufacturing 2019, 2019

Design for X (DFX) issue in product design and development is generally considered in the late de... more Design for X (DFX) issue in product design and development is generally considered in the late design process, e.g., detailed design phase. DFX issue plays a crucial role in determining cost, time and quality of product design. Due to the current DFX, research pays little attention to incorporate the industrial design, and the industrial design is rarely taken into account in the early design phase; design iteration is unavoidable. This paper identifies the new X issues requires in development of product life cycle design methodology, and then proposes a Design Merged X (DMX) framework to combine industrial design and product design with existing DFX methods. A process model of how to apply the DMX framework is also presented. An illustrative case is presented to show how the proposed DMX framework can be used to fulfill the new design process with considering the industrial design in the early design phase. Results of this study contribute new knowledge to DFX field.

2018 52nd Annual Conference on Information Sciences and Systems (CISS), 2018

The retina provides an excellent system for understanding the trade-offs that influence distribut... more The retina provides an excellent system for understanding the trade-offs that influence distributed information processing across multiple neuron types. We focus here on the problem faced by the visual system of allocating a limited number neurons to encode different visual features at different spatial locations. The retina needs to solve three competing goals: 1) encode different visual features, 2) maximize spatial resolution for each feature, and 3) maximize accuracy with which each feature is encoded at each location. There is no current understanding of how these goals are optimized together. While information theory provides a platform for theoretically solving these problems, evaluating information provided by the responses of large neuronal arrays is in general challenging. Here we present a solution to this problem in the case where multi-dimensional stimuli can be decomposed into approximately independent components that are subsequently coupled by neural responses. Using...

Rhythm can often be broken down in a binary fashion-strong and weak beats. Theories of rhythmic a... more Rhythm can often be broken down in a binary fashion-strong and weak beats. Theories of rhythmic attention posit that attention is heightened on strong metric positions that are accented or otherwise different from weak positions. Abecasis et al., (2004) used the magnitude of the P300 Event Related Potential (ERP) to determine the processing of an oddball at different metrical positions and found a preference for binary metrical structures, akin to a march, as opposed to a ternary metrical structure, akin to a waltz. This previous study did not, however, assess quaternary meter, a derivative of binary meter. Moreover, a quaternary meter could explain if an increased expectancy of a stimulus due to a delay of presentation (i.e., a sequential effect) contributes to the amplitude of a P300 response. This study tested the preference for a binary meter by introducing a new paradigm and quantification of the sequential effect. The data suggests that binary meter may be preferred as P300 average amplitude was highest on the strong beat even in quaternary meter, supporting rhythmic attention theory. The quaternary condition did not seem to follow rhythmic attention while the binary condition did, and the predicted sequential effect in both meters was not apparent

ELECTROPHORESIS, 2020

Y-chromosome SNP haplogroups exhibit geographic structuring in many populations around the world.... more Y-chromosome SNP haplogroups exhibit geographic structuring in many populations around the world. Therefore, Y chromosome haplogroups have been widely used to infer paternal bio-geographical ancestry and high-resolution paternal lineage classification. In the present study, we designed a customized panel containing 183 Y-SNPs based on previous studies and evaluated the genotyping performance and repeatability, concordance, sensitivity and ability of analysing case-type samples using a MALDI-TOF MS platform. The average call rate for duplicate typing of any one SNP in the panel was 97.0%. In the concordance and accuracy study, the results of haplogroup designation obtained from MALDI-TOF MS platform were fully consistent with those obtained from the next-generation sequencing (NGS) platform. The optimal amount of template DNA in the PCR seemed to be 10 ng. However, if less DNA (≥156.25pg) was available, it was still possible to obtain

It is known thatcis-acting DNA motifs play an important role in regulating gene expression. The g... more It is known thatcis-acting DNA motifs play an important role in regulating gene expression. The genome in a cell thus contains the information that not only encodes for the synthesis of proteins but also is necessary for regulating expression of genes. Therefore, the mRNA level of a gene may be predictable from the DNA sequence. Indeed, three deep neural network models were developed recently to predict the mRNA level of a gene directly or indirectly from the DNA sequence around the transcription start side of the gene. In this work, we develop a deep residual network model, named ExpResNet, to predict gene expression directly from DNA sequence. Applying ExpResNet to the GTEx data, we demonstrate that ExpResNet outperforms the three existing models across four tissues tested. Our model may be useful in the investigation of gene regulation.

Journal of Computational Biology, 2020

Precision medicine has become a trend in the field of biomedical research. Analysis of mRNA and m... more Precision medicine has become a trend in the field of biomedical research. Analysis of mRNA and microRNA (miRNA) is a basis of it. Although the analyses of mRNA and miRNA in different cancers have been studied widely, only a few integrated studies concerning their interactions on large patient cohorts have been done. In this article, we study the integrated significance of mRNA, miRNA, and their interactions in the survival analysis of breast cancer (BRCA). We detect differential expression genes, then we analyze significant genes of BRCA by the Cox model. The model described in this article has better performance than the model without predictors of interactions of miRNA and mRNA. The average C index considering miRNA-target interactions is 0.7, which achieves 84% improvement compared with the model without interactions, which is 0.38.

Science of The Total Environment, 2020

This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

MATEC Web of Conferences, 2019

With the increasingly intense competition in the market, more attention has been paid to customer... more With the increasingly intense competition in the market, more attention has been paid to customer’s perceptual requirements. Combining human sensibility and technology, Kansei Engineering plays a crucial role for industrial designers in making decisions at an earlier design stage in product development. However, sometimes it is challenging to define the customer requirements during the market research phase, especially for the perceptual requirements. Given this problem, this paper develops a process for establishing perceptual indexes based on the axiomatic design approach to improve design efficiency. To achieve this, an initial perceptual index is first built based on data previously collected from designers or design agencies. Then, the theory of axiomatic design is used to classify the schemes. The final perceptual index is filtered and built according to the frequency of occurrence.

Forensic Sciences Research, 2019

Short tandem repeat (STR) profiling is one of the mostly used systems for forensic applications. ... more Short tandem repeat (STR) profiling is one of the mostly used systems for forensic applications. In certain circumstances, STR profiling is time-consuming and costly, which potentially leads to delays in criminal investigations. LGC (Laboratory of the Government Chemist, UK) Forensics has developed a robust STR profiling platform called the ParaDNA V R Intelligence Test System which can provide early tactical intelligence and aid investigators in making informed decisions on sample prioritization for detection. Here, we validated the ParaDNA intelligence test for its application in forensic cases using a range of mock evidence items following guidelines set by the Scientific Working Group on DNA Analysis Methods (SWGDAM). Specifically, we tested the sensitivity and accuracy of the ParaDNA intelligence test, as well as the success rates for detecting mock samples and for use in case scenarios. Our findings demonstrate that the ParaDNA intelligence test generates useful DNA profiles, especially for samples such as blood, saliva, and semen that contain ample DNA, indicating the benefits of including ParaDNA as a prior step in forensic STR profiling pipelines.

IEEE Transactions on Instrumentation and Measurement, 2019

Blackbody cavity temperature sensor, combining the advantages of contact and noncontact thermomet... more Blackbody cavity temperature sensor, combining the advantages of contact and noncontact thermometry, has been well used for continuously measuring the molten steel temperature in continuous casting. However, the conventional cavity is composed of the opaque wall, and thermal conduction lag from outer to inner wall leads to the relatively slow response speed of the sensor, which impedes its application in the production processes with rapid temperature variation, such as nonferrous and ferrous metal refining. In this article, a novel blackbody cavity temperature sensor is introduced by replacing the blackbody cavity's opaque wall with a transparent wall, thus eliminating the response lag of the temperature sensor. Experiments are conducted on the response speed among the conventional and the novel blackbody cavity temperature sensors as well as two typical thermocouples. The response speed is evaluated by comparing measurements of different temperature sensors after some amount of time when the measured liquid is heated up very fast. Results show that the novel blackbody cavity temperature sensor's response speed is faster than both the traditional one and the thermocouples. Anticipated results indicate the rapidity of the novel sensor and its potential application in an accurate and rapid field with harsh conditions.

Chemical Geology, 2019

Forensic Science International: Genetics Supplement Series, 2019

Chinese Optics Letters, 2019

Design of a conventional zoom lens is always challengeable because it requires not only sophistic... more Design of a conventional zoom lens is always challengeable because it requires not only sophisticated optical design strategy, but also complex and precise mechanical structures for lens adjustment. In this paper, we propose a continuous zoom lens consisting of two chiral geometric metasurfaces with dielectric nanobrick arrays sitting on a transparent substrate. The metalens can continuously vary the focal length by rotating either of the two metasurfaces along its optical axis without changing any other conditions. More importantly, because of the polarization dependence of the geometric metasurface, the positive and negative polarities are interchangeable in one identical metalens only by changing the handedness of the incident circularly polarized light, which can generate varyingfocal lengths ranging from − to + in principle. On account of its advantages of compactness, flexibility and easiness in design, the proposed zoom metalens can provide new perspectives for the development of continuous-zoom optical system and it can find applications in fields which require ultracompact and continuous-zoom imaging and reconfigurable beam wavefront steering.

Applied Clay Science, 2019

Kaolinite dissolution rates at ambient temperature and pH 7.1-8.5 were measured with the isotope ... more Kaolinite dissolution rates at ambient temperature and pH 7.1-8.5 were measured with the isotope tracer method. A rare Si isotope 29 Si was introduced to the experimental solutions, which reacted with Georgia kaolinite (KGa-1b) composed of mostly 28 Si. Reaction rates were tracked by 29 Si/ 28 Si ratios of reacted solutions. The 16 batch experiments were designed with a grid of solutions that ranged from near saturation to supersaturation with respect to kaolinite. An average dissolution rate (unidirectional) of 5.4 ± 1.6 × 10 −14 mol (kaol) s −1 m −2 consistently fitted the 29 Si/ 28 Si ratios for all 16 experiments, indicating the dissolution rates were independent of pH in near-neutral pH waters and independent from the levels of departure from equilibrium. In other words, it appears that the dissolution reaction mechanisms do not change across from the kaolinite-undersaturated to kaolinite-supersaturated solutions near-equilibrium. The near-equilibrium kaolinite dissolution rates in this study are a new type of rates-unidirectional rates (from the isotope tracer method), which differ from all near-equilibrium kaolinite dissolution rates in the literature that are based upon Si or Al concentrations and are net rates (dissolution minus precipitation rates). Kaolinite dissolution was non-stoichiometric in all experiments. The Si and Al concentrations were sometimes systematic but more often erratic, resulting from the precipitation of AleSi secondary phases. The experimental solutions were grossly supersaturated with respect to gibbsite, allophanes, and imogolites. This confirms our hypothesis that the scatter and conflicts of near-equilibrium data are caused by unaccounted-for secondary phase precipitation, but the isotope tracer method successfully circumvents this experimental pitfall. Most natural waters are supersaturated with clays and are near-neutral pH. Our experimental rates are more applicable to the studies of natural waters than the majority rates available in the literature, which have mostly been measured at far-from-equilibrium, acidic pH, and high temperatures.

Computers & Geosciences, 2019

Abstract Rate equations and kinetic parameters for about 100 minerals were programmed into a libr... more Abstract Rate equations and kinetic parameters for about 100 minerals were programmed into a library of callable Basic language scripts for the geochemical modeling program Phreeqc (version 3.5.0) to facilitate the application of kinetics in geochemical modeling. For most minerals, the following general equation is used: r n e t = S A ∑ j A j e − E a , j / R T ∏ i a i , j n i ( 1 − Ω p j ) q j where rnet stands for the net dissolution rate of a mineral phase (mol kgw−1 s−1); j the jth reaction mechanism; SA the surface area per unit water mass (m2 kgw−1); Aj the Arrhenius pre-exponential factor (mol m−2 s−1); Ea,j the apparent reaction activation energy (J mol−1); R the universal gas constant (8.31446 J mol−1 K−1); T the temperature (K); ai the activity of aqueous species i; Ω the mineral saturation quotient. pj and qj are empirical fitting parameters. j stands for the specific mechanisms of reaction. Other forms of rate equations and associated parameters programmed in the library include parallel mechanisms, Langmuir adsorption isotherm, and empirical rate equations that apply to a specific reaction mechanism or geochemical system. A separate file of PHASEs, which define the chemical stoichiometry of the phases, dissolution reactions, and equilibrium constants of the dissolution reactions, is also provided. PHREEQC requires that the names in PHASES and RATES blocks match with each other. The Basic language scripts can also be used as templates for writing other rate equations which users might wish to use. To illustrate the application of the script library, we simulated the reaction path of albite dissolution at 25 °C and 1 bar, using three rate equations and compared the results. The script and phase library and supporting materials can be downloaded from https://github.com/HydrogeoIU/PHREEQC-Kinetic-Library and doi.org/10.5967/41gq-yr13.

E3S Web of Conferences, 2019

Since the time of J. H. van’t Hoff [1], it has been known that chemical equilibrium is dynamic, m... more Since the time of J. H. van’t Hoff [1], it has been known that chemical equilibrium is dynamic, meaning that at equilibrium, chemical reactions do not cease, but instead the forward and backward reaction rates are equal. The constant concentrations at equilibrium preclude the use of concentrations to measure reaction rates at equilibrium. Therefore, with the exception of a few special cases, no reaction rates at equilibrium have been published in the literature of chemistry, physics, or chemical engineering. Here we report dissolution and precipitation rates at equilibrium for quartz and barite with the isotope-doping method. Experimental data show that dissolution and precipitation rates are equal at equilibrium, indicating the principle of detailed balance (PDB) appear to be applicable at these experimental conditions. The PDB has been a cornerstone for irreversible thermodynamics and chemical kinetics for a long time, and its wide application in geochemistry has mostly been impli...