arXiv subjects

Jun Yuan

Publications and source records attributed to Jun Yuan.

At least 55 records · Page 3Linked to original sources

arXiv2022

Evaluation of Sampling Methods for Scatterplots

Given a scatterplot with tens of thousands of points or even more, a natural question is which sampling method should be used to create a small but "good" scatterplot for a better abstraction. We present the results of a user study that investigates the influence of different sampling strategies on multi-class scatterplots. The main goal of this study is to understand the capability of sampling methods in preserving the density, outliers, and overall shape of a scatterplot. To this end, we comprehensively review the literature and select seven typical sampling strategies as well as eight representative datasets. We then design four experiments to understand the performance of different strategies in maintaining: 1) region density; 2) class density; 3) outliers; and 4) overall shape in the sampling results. The results show that: 1) random sampling is preferred for preserving region density; 2) blue noise sampling and random sampling have comparable performance with the three multi-class sampling strategies in preserving class density; 3) outlier biased density based sampling, recursive subdivision based sampling, and blue noise sampling perform the best in keeping outliers; and 4) blue noise sampling outperforms the others in maintaining the overall shape of a scatterplot.

cs.HC↗

arXiv2021

Cyclic inequalities involving Cater cyclic function

By means of the mathematical analysis theory, inequality theory, mathematical induction and the dimension reduction method, under the proper hypotheses, we establish the following cyclic inequalities: \[\sum_{i=1}^{n} {a_i^{a_{n+1-i}}}\leq\sum_{\text{cyc}:~n}^{1\leq i\leq n} {a_i^{a_{i + 1}}}\leq \sum_{i=1}^{n} {a_i^{a_i}},~\forall n\geq2.\]

math.CA↗

arXiv2021

An Exploration And Validation of Visual Factors in Understanding Classification Rule Sets

Rule sets are often used in Machine Learning (ML) as a way to communicate the model logic in settings where transparency and intelligibility are necessary. Rule sets are typically presented as a text-based list of logical statements (rules). Surprisingly, to date there has been limited work on exploring visual alternatives for presenting rules. In this paper, we explore the idea of designing alternative representations of rules, focusing on a number of visual factors we believe have a positive impact on rule readability and understanding. We then presents a user study exploring their impact. The results show that some design factors have a strong impact on how efficiently readers can process the rules while having minimal impact on accuracy. This work can help practitioners employ more effective solutions when using rules as a communication strategy to understand ML models.

cs.HC↗

arXiv2021

AdViCE: Aggregated Visual Counterfactual Explanations for Machine Learning Model Validation

Rapid improvements in the performance of machine learning models have pushed them to the forefront of data-driven decision-making. Meanwhile, the increased integration of these models into various application domains has further highlighted the need for greater interpretability and transparency. To identify problems such as bias, overfitting, and incorrect correlations, data scientists require tools that explain the mechanisms with which these model decisions are made. In this paper we introduce AdViCE, a visual analytics tool that aims to guide users in black-box model debugging and validation. The solution rests on two main visual user interface innovations: (1) an interactive visualization design that enables the comparison of decisions on user-defined data subsets; (2) an algorithm and visual design to compute and visualize counterfactual explanations - explanations that depict model outcomes when data features are perturbed from their original values. We provide a demonstration of the tool through a use case that showcases the capabilities and potential limitations of the proposed approach.

cs.HC↗

arXiv2021

Maximizing the Set Cardinality of Users Scheduled for Ultra-dense uRLLC Networks

Ultra-reliability and low latency communication has long been an important but challenging task in the fifth and sixth generation wireless communication systems. Scheduling as many users as possible to serve on the limited time-frequency resource is one of a crucial topic, subjecting to the maximum allowable transmission power and the minimum rate requirement of each user. We address it by proposing a mixed integer programming model, with the goal of maximizing the set cardinality of users instead of maximizing the system sum rate or energy efficiency. Mathematical transformations and successive convex approximation are combined to solve the complex optimization problem. Numerical results show that the proposed method achieves a considerable performance compared with exhaustive search method, but with lower computational complexity.

cs.IT↗

arXiv2021

Bionic Optical Physical Unclonable Functions for Authentication and Encryption

Information security is of great importance for modern society with all things connected. Physical unclonable function (PUF) as a promising hardware primitive has been intensively studied for information security. However, the widely investigated silicon PUF with low entropy is vulnerable to various attacks. Herein, we introduce a concept of bionic optical PUFs inspired from unique biological architectures, and fabricate four types of bionic PUFs by molding the surface micro-nano structures of natural plant tissues with a simple, low-cost, green and environmentally friendly manufacturing process. The laser speckle responses of all bionic PUFs are statistically demonstrated to be random, unique, unpredictable and robust enough for cryptographic applications, indicating the broad applicability of bionic PUFs. On this ground, the feasibility of implementing bionic PUFs as cryptographic primitives in entity authentication and encrypted communication is experimentally validated, which shows its promising potential in the application of future information security.

cs.CR↗

arXiv2021

Visualizing Rule Sets: Exploration and Validation of a Design Space

Rule sets are often used in Machine Learning (ML) as a way to communicate the model logic in settings where transparency and intelligibility are necessary. Rule sets are typically presented as a text-based list of logical statements (rules). Surprisingly, to date there has been limited work on exploring visual alternatives for presenting rules. In this paper, we explore the idea of designing alternative representations of rules, focusing on a number of visual factors we believe have a positive impact on rule readability and understanding. The paper presents an initial design space for visualizing rule sets and a user study exploring their impact. The results show that some design factors have a strong impact on how efficiently readers can process the rules while having minimal impact on accuracy. This work can help practitioners employ more effective solutions when using rules as a communication strategy to understand ML models.

cs.HC↗

arXiv2020

Chirality-enabled optical dipole potential energy for two-level atoms

We consider the optical dipole potential energy, which arises from the interaction of a two-level atom with a circularly polarized Laguerre-Gaussian laser beam of small waist. The beam is characterized by the existence of a longitudinal electric field component which is responsible for the appearance of a chiral term in the optical dipole potential energy. This term reverses sign if either the winding number or the wave polarization of the beam reverses sign. We propose a scheme of a bi-chromatic vortex interaction with the two-level atom in which the resulting optical dipole potential is fully chiral.

physics.optics↗

arXiv2020

The role of charge recombination to spin-triplet excitons in non-fullerene acceptor organic solar cells

The power conversion efficiencies (PCEs) of organic solar cells (OSCs) using non-fullerene acceptors (NFAs) have now reached 18%. However, this is still lower than inorganic solar cells, for which PCEs >20% are commonplace. A key reason is that OSCs still show low open-circuit voltages (Voc) relative to their optical band gaps, attributed to non-radiative recombination. For OSCs to compete with inorganics in efficiency, all non-radiative loss pathways must be identified and where possible, removed. Here, we show that in most NFA OSCs, the majority of charge recombination at open-circuit proceeds via formation of non-emissive NFA triplet excitons (T1); in the benchmark PM6:Y6 blend, this fraction reaches 90%, contributing 60 mV to the reduction of Voc. We develop a new design to prevent recombination via this non-radiative channel through the engineering of significant hybridisation between the NFA T1 and the spin-triplet charge transfer exciton (3CTE). We model that the rate of the back charge transfer from 3CTE to T1 can be reduced by an order of magnitude, allowing re-dissociation of the 3CTE. We then demonstrate NFA systems where T1 formation is suppressed. This work therefore provides a clear design pathway for improved OSC performance to 20% PCE and beyond.

physics.app-ph↗

arXiv2020

Chirality and helicity of optical vortices of small beam waists

The chirality and helicity of linearly polarised Laguerre-Gaussian (LG) beams are examined. Such a type of light possesses a substantial longitudinal field when its beam waist is sufficiently small and so gives rise to non-zero chirality and helicity. In the simplest case of a doughnut beam of winding number $\ell=1$ and another identical to it but has $\ell= -1$, we obtain different chirality and helicity distributions at the focal plane $z=0$. We also show that this chiral behaviour persists and the patterns evolve so that on planes at $z<0$ and $z>0$ the beam convergence contributes differently to the changes in the chirality and helicity distributions.

physics.optics↗

arXiv2020

Mind the Pad -- CNNs can Develop Blind Spots

We show how feature maps in convolutional networks are susceptible to spatial bias. Due to a combination of architectural choices, the activation at certain locations is systematically elevated or weakened. The major source of this bias is the padding mechanism. Depending on several aspects of convolution arithmetic, this mechanism can apply the padding unevenly, leading to asymmetries in the learned weights. We demonstrate how such bias can be detrimental to certain tasks such as small object detection: the activation is suppressed if the stimulus lies in the impacted area, leading to blind spots and misdetection. We propose solutions to mitigate spatial bias and demonstrate how they can improve model accuracy.

cs.CV↗

arXiv2020

A Survey of Visual Analytics Techniques for Machine Learning

Visual analytics for machine learning has recently evolved as one of the most exciting areas in the field of visualization. To better identify which research topics are promising and to learn how to apply relevant techniques in visual analytics, we systematically review 259 papers published in the last ten years together with representative works before 2010. We build a taxonomy, which includes three first-level categories: techniques before model building, techniques during model building, and techniques after model building. Each category is further characterized by representative analysis tasks, and each task is exemplified by a set of recent influential works. We also discuss and highlight research challenges and promising potential future research opportunities useful for visual analytics researchers.

cs.HC↗

arXiv2020

ViCE: Visual Counterfactual Explanations for Machine Learning Models

The continued improvements in the predictive accuracy of machine learning models have allowed for their widespread practical application. Yet, many decisions made with seemingly accurate models still require verification by domain experts. In addition, end-users of a model also want to understand the reasons behind specific decisions. Thus, the need for interpretability is increasingly paramount. In this paper we present an interactive visual analytics tool, ViCE, that generates counterfactual explanations to contextualize and evaluate model decisions. Each sample is assessed to identify the minimal set of changes needed to flip the model's output. These explanations aim to provide end-users with personalized actionable insights with which to understand, and possibly contest or improve, automated decisions. The results are effectively displayed in a visual interface where counterfactual explanations are highlighted and interactive methods are provided for users to explore the data and model. The functionality of the tool is demonstrated by its application to a home equity line of credit dataset.

cs.HC↗

arXiv2020

OoDAnalyzer: Interactive Analysis of Out-of-Distribution Samples

One major cause of performance degradation in predictive models is that the test samples are not well covered by the training data. Such not well-represented samples are called OoD samples. In this paper, we propose OoDAnalyzer, a visual analysis approach for interactively identifying OoD samples and explaining them in context. Our approach integrates an ensemble OoD detection method and a grid-based visualization. The detection method is improved from deep ensembles by combining more features with algorithms in the same family. To better analyze and understand the OoD samples in context, we have developed a novel kNN-based grid layout algorithm motivated by Hall's theorem. The algorithm approximates the optimal layout and has $O(kN^2)$ time complexity, faster than the grid layout algorithm with overall best performance but $O(N^3)$ time complexity. Quantitative evaluation and case studies were performed on several datasets to demonstrate the effectiveness and usefulness of OoDAnalyzer.

cs.HC↗

arXiv2017

Inversion Domain Boundary Induced Stacking and Bandstructure Diversity in Bilayer MoSe2

Interlayer rotation and stacking were recently demonstrated as effective strategies for tuning physical properties of various two-dimensional materials. The latter strategy was mostly realized in hetero-structures with continuously varied stacking orders, which obscure the revelation of the intrinsic role of a certain stacking order in its physical properties. Here, we introduce inversion-domain-boundaries into molecular-beam-epitaxy grown MoSe2 homo-bilayers, which induce unusual-fractional lattice translations to their surrounding domains, accounting for the observed diversity of large-area and uniform stacking sequences. Unusual low-symmetric stacking orders were observed using transmission electron microscopy and detailed geometries were identified by density functional theory. A linear relation was also revealed between interlayer distance and stacking energy. These stacking sequences yield various energy alignments between the valence states at Γ and K, showing stacking dependent bandgaps and valence band tail states in the measured scanning tunneling spectroscopy. These results may benefit the design of two-dimensional multilayers with manipulable stacking orders.

cond-mat.mtrl-sci↗

arXiv2017

Direct identification of monolayer rhenium diselenide by an individual diffraction pattern

In the current extensive studies of transition metal dichalcogenides (TMDCs), compared to hexagonal layered materials, like graphene, hBN and MoS2, low symmetry layered 2D crystals showed great potential for applications in anisotropic devices. Rhenium diselenide (ReSe2) has the bulk space group P1 and belongs to triclinic crystal system with a deformed cadmium iodide type structure. Here we unambiguously determined monolayer and its vertical orientation of rhenium diselenide membrane with an individual electron diffraction pattern, which could be applicable to low symmetry crystal systems, including both triclinic and monoclinic lattices, as long as their third unit-cell basis vector is not perpendicular to basal plane. Atomically resolved image from probe corrected annular dark field scanning transmission electron microscope (ADF-STEM) was employed to validate layer number. Finally, experimental results were well explained by kinematical electron diffraction theory and corresponding simulations.

cond-mat.mtrl-sci↗

arXiv2017

Design of mechanically robust metasurface lenses for RGB colors

In recent years, dielectric rod based metasurface lenses have been particularly investigated for their potential applications in replacing the traditional bulky lens with high efficiency. However, the isolated granular structure may lead to robustness and chromatic aberration concerns. In this work, a low refractive-index embedding medium was applied to solve the structural stability problem that also enables the device to be transferable to desired position or surface. Based on this, a compound metasurface lens composed of randomly interleaved frequency-selective zone sectors is proposed to broaden the bandwidth of the two-dimensional device. The validities of the proposed method and the application potentials for the multifunctional lens that can manipulate RGB three colors have been numerically examined and verified. The current results are of essence in guiding the future design of metasurface lenses for real practice.

physics.optics↗

arXiv2017

Orbital angular momentum mode selection by rotationally symmetric superposition of chiral states with application to electron vortex beams

A general orbital angular momentum (OAM) mode selection principle is put forward involving the rotationally symmetric superposition of chiral states. This principle is not only capable of explaining the operation of spiral zone plate holograms and suggesting that naturally occurring rotationally symmetric patterns could be inadvertent sources of vortex beams, but more importantly, it enables the systematic and flexible generation of structured OAM waves in general. This is demonstrated both experimentally and theoretically in the context of electron vortex beams using rotationally symmetric binary amplitude chiral sieve masks.

quant-ph↗