Showing 1–50 of 182 results for author: Huili

Harnessing XGBoost for Robust Biomarker Selection of Obsessive-Compulsive Disorder (OCD) from Adolescent Brain Cognitive Development (ABCD) data

Authors: Xinyu Shen, Qimin Zhang, Huili Zheng, Weiwei Qi

Abstract: This study evaluates the performance of various supervised machine learning models in analyzing highly correlated neural signaling data from the Adolescent Brain Cognitive Development (ABCD) Study, with a focus on predicting obsessive-compulsive disorder scales. We simulated a dataset to mimic the correlation structures commonly found in imaging data and evaluated logistic regression, elastic netw… ▽ More This study evaluates the performance of various supervised machine learning models in analyzing highly correlated neural signaling data from the Adolescent Brain Cognitive Development (ABCD) Study, with a focus on predicting obsessive-compulsive disorder scales. We simulated a dataset to mimic the correlation structures commonly found in imaging data and evaluated logistic regression, elastic networks, random forests, and XGBoost on their ability to handle multicollinearity and accurately identify predictive features. Our study aims to guide the selection of appropriate machine learning methods for processing neuroimaging data, highlighting models that best capture underlying signals in high feature correlations and prioritize clinically relevant features associated with Obsessive-Compulsive Disorder (OCD). △ Less

Submitted 14 May, 2024; originally announced July 2024.

CU-Net: a U-Net architecture for efficient brain-tumor segmentation on BraTS 2019 dataset

Authors: Qimin Zhang, Weiwei Qi, Huili Zheng, Xinyu Shen

Abstract: Accurately segmenting brain tumors from MRI scans is important for develo** effective treatment plans and improving patient outcomes. This study introduces a new implementation of the Columbia-University-Net (CU-Net) architecture for brain tumor segmentation using the BraTS 2019 dataset. The CU-Net model has a symmetrical U-shaped structure and uses convolutional layers, max pooling, and upsampl… ▽ More Accurately segmenting brain tumors from MRI scans is important for develo** effective treatment plans and improving patient outcomes. This study introduces a new implementation of the Columbia-University-Net (CU-Net) architecture for brain tumor segmentation using the BraTS 2019 dataset. The CU-Net model has a symmetrical U-shaped structure and uses convolutional layers, max pooling, and upsampling operations to achieve high-resolution segmentation. Our CU-Net model achieved a Dice score of 82.41%, surpassing two other state-of-the-art models. This improvement in segmentation accuracy highlights the robustness and effectiveness of the model, which helps to accurately delineate tumor boundaries, which is crucial for surgical planning and radiation therapy, and ultimately has the potential to improve patient outcomes. △ Less

Submitted 18 June, 2024; originally announced June 2024.

Towards Next-Generation Steganalysis: LLMs Unleash the Power of Detecting Steganography

Authors: Minhao Bai. **shuai Yang, Kaiyi Pang, Huili Wang, Yongfeng Huang

Abstract: Linguistic steganography provides convenient implementation to hide messages, particularly with the emergence of AI generation technology. The potential abuse of this technology raises security concerns within societies, calling for powerful linguistic steganalysis to detect carrier containing steganographic messages. Existing methods are limited to finding distribution differences between stegano… ▽ More Linguistic steganography provides convenient implementation to hide messages, particularly with the emergence of AI generation technology. The potential abuse of this technology raises security concerns within societies, calling for powerful linguistic steganalysis to detect carrier containing steganographic messages. Existing methods are limited to finding distribution differences between steganographic texts and normal texts from the aspect of symbolic statistics. However, the distribution differences of both kinds of texts are hard to build precisely, which heavily hurts the detection ability of the existing methods in realistic scenarios. To seek a feasible way to construct practical steganalysis in real world, this paper propose to employ human-like text processing abilities of large language models (LLMs) to realize the difference from the aspect of human perception, addition to traditional statistic aspect. Specifically, we systematically investigate the performance of LLMs in this task by modeling it as a generative paradigm, instead of traditional classification paradigm. Extensive experiment results reveal that generative LLMs exhibit significant advantages in linguistic steganalysis and demonstrate performance trends distinct from traditional approaches. Results also reveal that LLMs outperform existing baselines by a wide margin, and the domain-agnostic ability of LLMs makes it possible to train a generic steganalysis model (Both codes and trained models are openly available in https://github.com/ba0z1/Linguistic-Steganalysis-with-LLMs). △ Less

Submitted 15 May, 2024; originally announced May 2024.

Accurate Chemical Reaction Modeling on Noisy Intermediate-Scale Quantum Computers Using a Noise-Resilient Wavefunction Ansatz

Authors: Xiongzhi Zeng, Huili Zhang, Shizheng Zhang, Pei Liu, Kehuan Linghu, Jiangyu Cui, Xiaoxia Cai, Jie Liu, Zhenyu Li, **long Yang

Abstract: Quantum computing is of great potential for chemical system simulations. In this study, we propose an efficient protocol of quantum computer based simulation of chemical systems which enables accurate chemical reaction modeling on noisy intermediate-scale quantum (NISQ) devices. In this protocol, we combine an correlation energy-based active orbital selection, an effective Hamiltonian from the dri… ▽ More Quantum computing is of great potential for chemical system simulations. In this study, we propose an efficient protocol of quantum computer based simulation of chemical systems which enables accurate chemical reaction modeling on noisy intermediate-scale quantum (NISQ) devices. In this protocol, we combine an correlation energy-based active orbital selection, an effective Hamiltonian from the driven similarity renormalization group (DSRG) method, and a noise-resilient wavefunction ansatz. Such a combination gives a quantum resource-efficient way to accurately simulate chemical systems. The power of this protocol is demonstrated by numerical results for systems with up to tens of atoms. Modeling of a Diels-Alder (DA) reaction is also performed on a cloud-based superconducting quantum computer. These results represent an important step forward in realizing quantum utility in the NISQ era. △ Less

Submitted 22 April, 2024; originally announced April 2024.

Coexisting steady-state solutions of a class of reaction-diffusion systems with different boundary conditions

Authors: Ningning Zhu, Dongpo Hu, Huili Bi

Abstract: In this work, we investigate a reaction-diffusion system in which both species are influenced by self-diffusion. Due to Hopf's boundary lemma, we obtain the boundedness of the classical solution of the system. By considering a particular function, we provide a complete characterization of the parameter ranges such that coexisting solutions of the system do not exist under three boundary conditions… ▽ More In this work, we investigate a reaction-diffusion system in which both species are influenced by self-diffusion. Due to Hopf's boundary lemma, we obtain the boundedness of the classical solution of the system. By considering a particular function, we provide a complete characterization of the parameter ranges such that coexisting solutions of the system do not exist under three boundary conditions. Then based on the maximum principle, a sufficient condition for the existence of constant coexisting solutions of the system under Neumann boundary conditions was derived. △ Less

Submitted 19 April, 2024; originally announced April 2024.

Semantic SQL -- Combining and optimizing semantic predicates in SQL

Authors: Akash Mittal, Anshul Bheemreddy, Huili Tao

Abstract: In recent years, the surge in unstructured data analysis, facilitated by advancements in Machine Learning (ML), has prompted diverse approaches for handling images, text documents, and videos. Analysts, leveraging ML models, can extract meaningful information from unstructured data and store it in relational databases, allowing the execution of SQL queries for further analysis. Simultaneously, vec… ▽ More In recent years, the surge in unstructured data analysis, facilitated by advancements in Machine Learning (ML), has prompted diverse approaches for handling images, text documents, and videos. Analysts, leveraging ML models, can extract meaningful information from unstructured data and store it in relational databases, allowing the execution of SQL queries for further analysis. Simultaneously, vector databases have emerged, embedding unstructured data for efficient top-k queries based on textual queries. This paper introduces a novel framework SSQL - Semantic SQL that utilizes these two approaches, enabling the incorporation of semantic queries within SQL statements. Our approach extends SQL queries with dedicated keywords for specifying semantic queries alongside predicates related to ML model results and metadata. Our experimental results show that using just semantic queries fails catastrophically to answer count and spatial queries in more than 60% of the cases. Our proposed method jointly optimizes the queries containing both semantic predicates and predicates on structured tables, such as those generated by ML models or other metadata. Further, to improve the query results, we incorporated human-in-the-loop feedback to determine the optimal similarity score threshold for returning results. △ Less

Submitted 5 April, 2024; originally announced April 2024.

Dualtronics: leveraging both faces of polar semiconductors

Authors: Len van Deurzen, Eungkyun Kim, Naomi Pieczulewski, Zexuan Zhang, Anna Feduniewicz-Zmuda, Mikolaj Chlipala, Marcin Siekacz, David Muller, Huili Grace Xing, Debdeep Jena, Henryk Turski

Abstract: Unlike non-polar semiconductors such as silicon, the broken inversion symmetry of the wide bandgap semiconductor gallium nitride leads to a large electronic polarization along a unique crystal axis. This makes the two surfaces of the semiconductor wafer perpendicular to the polar axis dramatically different in their physical and chemical properties. In the last three decades, the cation (gallium)… ▽ More Unlike non-polar semiconductors such as silicon, the broken inversion symmetry of the wide bandgap semiconductor gallium nitride leads to a large electronic polarization along a unique crystal axis. This makes the two surfaces of the semiconductor wafer perpendicular to the polar axis dramatically different in their physical and chemical properties. In the last three decades, the cation (gallium) face of gallium nitride has been used for photonic devices such as LEDs and lasers. Though the cation face has also been predominantly used for electronic devices, the anion (nitrogen) face has recently shown promise for high electron mobility transistors due to favorable polarization discontinuities. In this work we show that it is possible to make photonic devices on the cation face, and electronic devices on the anion face, of the same semiconductor wafer. This opens the possibility for dualtronics in polar semiconductors, where electronic, photonic, and acoustic properties can be implemented on opposite faces of the same wafer, dramatically enhancing the functional capabilities of this revolutionary semiconductor family. △ Less

Submitted 4 April, 2024; originally announced April 2024.

Nanoring Tamm Cavity in the Telecommunications O band

Authors: Huili Hou, David Dlaka, Jon Pugh, Ruth Oulton, Edmund Harbord

Abstract: Quantum and classical telecommunications require efficient sources of light. Semiconductorsources, owing to the high refractive index of the medium, often exploit photonic cavities to enhance the external emission of photons into a well-defined optical mode. Optical Tamm States (OTS), in which light is confined between a distributed Bragg reflector and a thin metal layer have attracted interest as… ▽ More Quantum and classical telecommunications require efficient sources of light. Semiconductorsources, owing to the high refractive index of the medium, often exploit photonic cavities to enhance the external emission of photons into a well-defined optical mode. Optical Tamm States (OTS), in which light is confined between a distributed Bragg reflector and a thin metal layer have attracted interest as confined Tamm structures are readily manufactureable broadband cavities. Their effficiency is limited however by the absorption inherent in the metal layer. We propose a nanoring Tamm structure in which a nanoscale patterned annular metasurface is exploited to reduce this absorption and thereby enhance emission efficiency. To this end, we present designs for a nanoring Tamm structure optimised for the telecommunications O band, and demonstrate a near doubling of output efficiency (35%) over an analogous solid disc confined Tamm structure (18%). Simulations of designs optimised for different wavelengths are suggestive of annular coupling between the Tamm state and surface plasmons. These designs are applicable to the design of single photon sources, nanoLEDs, and nanolasers for communications. △ Less

Submitted 16 February, 2024; originally announced February 2024.

Integrating Flow Theory and Adaptive Robot Roles: A Conceptual Model of Dynamic Robot Role Adaptation for the Enhanced Flow Experience in Long-term Multi-person Human-Robot Interactions

Authors: Huili Chen, Sharifa Alghowinem, Cynthia Breazeal, Hae Won Park

Abstract: In this paper, we introduce a novel conceptual model for a robot's behavioral adaptation in its long-term interaction with humans, integrating dynamic robot role adaptation with principles of flow experience from psychology. This conceptualization introduces a hierarchical interaction objective grounded in the flow experience, serving as the overarching adaptation goal for the robot. This objectiv… ▽ More In this paper, we introduce a novel conceptual model for a robot's behavioral adaptation in its long-term interaction with humans, integrating dynamic robot role adaptation with principles of flow experience from psychology. This conceptualization introduces a hierarchical interaction objective grounded in the flow experience, serving as the overarching adaptation goal for the robot. This objective intertwines both cognitive and affective sub-objectives and incorporates individual and group-level human factors. The dynamic role adaptation approach is a cornerstone of our model, highlighting the robot's ability to fluidly adapt its support roles - from leader to follower - with the aim of maintaining equilibrium between activity challenge and user skill, thereby fostering the user's optimal flow experiences. Moreover, this work delves into a comprehensive exploration of the limitations and potential applications of our proposed conceptualization. Our model places a particular emphasis on the multi-person HRI paradigm, a dimension of HRI that is both under-explored and challenging. In doing so, we aspire to extend the applicability and relevance of our conceptualization within the HRI field, contributing to the future development of adaptive social robots capable of sustaining long-term interactions with humans. △ Less

Submitted 5 January, 2024; originally announced January 2024.

Journal ref: Proceedings of the 2024 ACM/IEEE International Conference on Human - Robot Interaction (HRI24), March 11 - 14, 2024, Boulder, CO, USA

On the symmetric $q$-analog on the bi-univalent functions with respect to symmetric points

Authors: Pinhong Long, Huili Han, Halit Orhan, Huo Tang

Abstract: Our objective is to usher and investigate the subclass$\widetilde{\mathcal{S^{*}_{\sum}}}^η_{q}(μ,λ;φ)$ of the function class $\sum$ of analytic and bi-univalent functions related with the symmetric $q$-derivative operator and the generalized Bernardi integral operator. On the one hand, without the generalized Bernardi integral operator we estimate the second Hankel determinants for the reduced su… ▽ More Our objective is to usher and investigate the subclass$\widetilde{\mathcal{S^{*}_{\sum}}}^η_{q}(μ,λ;φ)$ of the function class $\sum$ of analytic and bi-univalent functions related with the symmetric $q$-derivative operator and the generalized Bernardi integral operator. On the one hand, without the generalized Bernardi integral operator we estimate the second Hankel determinants for the reduced subclasses $\widetilde{\mathcal{S^{*}_{\sum}}}_{q}(λ;φ)$ with respect to symmetric points. On the other hand, we also give the corresponding results of Fekete-Szegö functional inequalities and the upper bounds of the coefficients $a_2$ and $a_3$ for these subclasses. △ Less

Submitted 15 December, 2023; originally announced December 2023.

Comments: 19 pages

MSC Class: 30C45; 30C50

Ultrawide bandgap semiconductor heterojunction p-n diodes with distributed polarization doped p-type AlGaN layers on bulk AlN substrates

Authors: Shivali Agrawal, Len van Deurzen, Jimy Encomendero, Joseph E. Dill, Hsin Wei, Huang, Vladimir Protasenko, Huili, Xing, Debdeep Jena

Abstract: Ultrawide bandgap heterojunction p-n diodes with polarization-induced AlGaN p-type layers are demonstrated using plasma-assisted molecular beam epitaxy on bulk AlN substrates. Current-voltage characteristics show a turn on voltage of $V_{\text{bi}}\approx5.5$ V, a minimum room temperature ideality factor of $η\approx 1.63$, and more than 12 orders of current modulation at room temperature. Stable… ▽ More Ultrawide bandgap heterojunction p-n diodes with polarization-induced AlGaN p-type layers are demonstrated using plasma-assisted molecular beam epitaxy on bulk AlN substrates. Current-voltage characteristics show a turn on voltage of $V_{\text{bi}}\approx5.5$ V, a minimum room temperature ideality factor of $η\approx 1.63$, and more than 12 orders of current modulation at room temperature. Stable current operation of the ultrawide bandgap semiconductor diode is measured up to a temperature of 300$^\circ$C. The one-sided n$^{+}$-p heterojunction diode design enables a direct measurement of the spatial distribution of polarization-induced mobile hole density in the graded AlGaN layer from the capacitance-voltage profile. The measured average mobile hole density is $p \sim 5.7 \times 10^{17}$ cm$^{-3}$, in close agreement with what is theoretically expected from distributed polarization do**. Light emission peaked at 260 nm (4.78 eV) observed in electroluminescence corresponds to interband radiative recombination in the n$^{+}$ AlGaN layer. A much weaker deep-level emission band observed at 3.4 eV is attributed to cation-vacancy and silicon complexes in the heavily Si-doped AlGaN layer. These results demonstrate that distributed polarization do** enables ultrawide bandgap semiconductor heterojunction p-n diodes that have wide applications ranging from power electronics to deep-ultraviolet photonics. These devices can operate at high temperatures and in harsh environments. △ Less

Submitted 13 December, 2023; originally announced December 2023.

Semi-Supervised End-To-End Contrastive Learning For Time Series Classification

Authors: Huili Cai, Xiang Zhang, Xiaofeng Liu

Abstract: Time series classification is a critical task in various domains, such as finance, healthcare, and sensor data analysis. Unsupervised contrastive learning has garnered significant interest in learning effective representations from time series data with limited labels. The prevalent approach in existing contrastive learning methods consists of two separate stages: pre-training the encoder on unlab… ▽ More Time series classification is a critical task in various domains, such as finance, healthcare, and sensor data analysis. Unsupervised contrastive learning has garnered significant interest in learning effective representations from time series data with limited labels. The prevalent approach in existing contrastive learning methods consists of two separate stages: pre-training the encoder on unlabeled datasets and fine-tuning the well-trained model on a small-scale labeled dataset. However, such two-stage approaches suffer from several shortcomings, such as the inability of unsupervised pre-training contrastive loss to directly affect downstream fine-tuning classifiers, and the lack of exploiting the classification loss which is guided by valuable ground truth. In this paper, we propose an end-to-end model called SLOTS (Semi-supervised Learning fOr Time clasSification). SLOTS receives semi-labeled datasets, comprising a large number of unlabeled samples and a small proportion of labeled samples, and maps them to an embedding space through an encoder. We calculate not only the unsupervised contrastive loss but also measure the supervised contrastive loss on the samples with ground truth. The learned embeddings are fed into a classifier, and the classification loss is calculated using the available true labels. The unsupervised, supervised contrastive losses and classification loss are jointly used to optimize the encoder and classifier. We evaluate SLOTS by comparing it with ten state-of-the-art methods across five datasets. The results demonstrate that SLOTS is a simple yet effective framework. When compared to the two-stage framework, our end-to-end SLOTS utilizes the same input data, consumes a similar computational cost, but delivers significantly improved performance. We release code and datasets at https://anonymous.4open.science/r/SLOTS-242E. △ Less

Submitted 13 October, 2023; originally announced October 2023.

Comments: Submitted to NeurIPS 2023

ScaleNet: An Unsupervised Representation Learning Method for Limited Information

Authors: Huili Huang, M. Mahdi Roozbahani

Abstract: Although large-scale labeled data are essential for deep convolutional neural networks (ConvNets) to learn high-level semantic visual representations, it is time-consuming and impractical to collect and annotate large-scale datasets. A simple and efficient unsupervised representation learning method named ScaleNet based on multi-scale images is proposed in this study to enhance the performance of… ▽ More Although large-scale labeled data are essential for deep convolutional neural networks (ConvNets) to learn high-level semantic visual representations, it is time-consuming and impractical to collect and annotate large-scale datasets. A simple and efficient unsupervised representation learning method named ScaleNet based on multi-scale images is proposed in this study to enhance the performance of ConvNets when limited information is available. The input images are first resized to a smaller size and fed to the ConvNet to recognize the rotation degree. Next, the ConvNet learns the rotation-prediction task for the original size images based on the parameters transferred from the previous model. The CIFAR-10 and ImageNet datasets are examined on different architectures such as AlexNet and ResNet50 in this study. The current study demonstrates that specific image features, such as Harris corner information, play a critical role in the efficiency of the rotation-prediction task. The ScaleNet supersedes the RotNet by ~7% in the limited CIFAR-10 dataset. The transferred parameters from a ScaleNet model with limited data improve the ImageNet Classification task by about 6% compared to the RotNet model. This study shows the capability of the ScaleNet method to improve other cutting-edge models such as SimCLR by learning effective features for classification tasks. △ Less

Submitted 3 October, 2023; originally announced October 2023.

Comments: Accepted by DAGM GCPR 2021

Epitaxial lattice-matched Al$_{0.89}$Sc$_{0.11}$N/GaN distributed Bragg reflectors

Authors: Len van Deurzen, Thai-Son Nguyen, Joseph Casamento, Huili Grace Xing, Debdeep Jena

Abstract: We demonstrate epitaxial lattice-matched Al$_{0.89}$Sc$_{0.11}$N/GaN ten and twenty period distributed Bragg reflectors (DBRs) grown on c-plane bulk n-type GaN substrates by plasma-enhanced molecular beam epitaxy (PA-MBE). Resulting from a rapid increase of in-plane lattice coefficient as scandium is incorporated into AlScN, we measure a lattice-matched condition to $c$-plane GaN for a Sc content… ▽ More We demonstrate epitaxial lattice-matched Al$_{0.89}$Sc$_{0.11}$N/GaN ten and twenty period distributed Bragg reflectors (DBRs) grown on c-plane bulk n-type GaN substrates by plasma-enhanced molecular beam epitaxy (PA-MBE). Resulting from a rapid increase of in-plane lattice coefficient as scandium is incorporated into AlScN, we measure a lattice-matched condition to $c$-plane GaN for a Sc content of just 11\%, resulting in a large refractive index mismatch $\mathrm{Δn}$ greater than 0.3 corresponding to an index contrast of $\mathrm{Δn/n_{GaN}}$ = 0.12 with GaN. The DBRs demonstrated here are designed for a peak reflectivity at a wavelength of 400 nm reaching a reflectivity of 0.98 for twenty periods. It is highlighted that AlScN/GaN multilayers require fewer periods for a desired reflectivity than other lattice-matched Bragg reflectors such as those based on AlInN/GaN multilayers. △ Less

Submitted 28 September, 2023; originally announced September 2023.

Effects of $p$-wave Interactions on Borromean Efimov Trimers in Heavy-Light Fermi Systems

Authors: Cai-Yun Zhao, Hui-Li Han, Ting-Yun Shi

Abstract: We investigate the effects of $p$-wave interactions on Efimov trimers in systems comprising two identical heavy fermions and a light particle, with mass ratios larger than $13.6$. Our focus lies on the borromean regime where the ground-state trimer exists in the absence of dimers. Using pair-wise Lennard-Jones potentials and concentrating on the $L^π = 1^{-}$ symmetry, we explore the critical valu… ▽ More We investigate the effects of $p$-wave interactions on Efimov trimers in systems comprising two identical heavy fermions and a light particle, with mass ratios larger than $13.6$. Our focus lies on the borromean regime where the ground-state trimer exists in the absence of dimers. Using pair-wise Lennard-Jones potentials and concentrating on the $L^π = 1^{-}$ symmetry, we explore the critical value of the interspecies $s$-wave scattering length $a_{c}$ at which the borromean state appears in several two-component particle systems. We study the universal properties of $a_{c}$ and the influence of $p$-wave fermion-fermion interactions on its value. Our findings show that, in the absence of $p$-wave fermion-fermion interactions, $a_{c}$ is universally determined by the van der Waals radius and mass ratio. However, when attractive interactions between the two fermions are introduced, the formation of the borromean state becomes favored over the absence of $p$-wave fermion-fermion interaction. Furthermore, we demonstrate that the $p$-wave Efimov effects persist even when the fermion-fermion interaction is taken to the $p$-wave unitary limit. △ Less

Submitted 24 July, 2023; originally announced July 2023.

Phase-field based lattice Boltzmann model for simulating thermocapillary flows

Authors: Lei Wanga, Kun Hea, Huili Wang

Abstract: This paper proposes a simple and accurate lattice Boltzmann model for simulating thermocapillary flows, which is able to deal with thermophysical parameters contrasts. In this model, two lattice Boltzmann equations are utilized to solve the conservative Allen-Cahn equation and the incompressible Navier-Stokes equations, while another lattice Boltzmann equation is used for solving the temperature f… ▽ More This paper proposes a simple and accurate lattice Boltzmann model for simulating thermocapillary flows, which is able to deal with thermophysical parameters contrasts. In this model, two lattice Boltzmann equations are utilized to solve the conservative Allen-Cahn equation and the incompressible Navier-Stokes equations, while another lattice Boltzmann equation is used for solving the temperature field, where the collision term is delicately designed such that the influence of the thermophysical parameters contrasts is incorporated. In contrast to previous lattice Boltzmann model for thermocapillary flows, the most distinct feature of the current model is that the forcing term used in the present thermal lattice Boltzmann equation is not needed to calculate space derivatives of the heat capacitance or the order parameter, making the scheme much simpler and also possible to retain the main merits of the lattice Boltzmann method. The developed model is firstly validated by considering the thermocapillary flows in a heated microchannel with two superimposed planar fluids. It is then used to simulate thermocapillary migration of a two-dimensional deformable droplet, and its accuracy is once again consistent with the theoretical prediction when the Marangoni number approaches zero. Finally, we numerically study the motion of two recalcitrant bubbles in a two-dimensional channel where the relationship between surface tension and temperature is assumed to be a parabolic function. It is found that owing to the competing between the inertia and thermal effects, the bubbles are able to move against the liquid's bulk motion and towards areas with low surface tension. △ Less

Submitted 20 June, 2023; originally announced June 2023.

Comparative Analysis of Deep Learning Models for Brand Logo Classification in Real-World Scenarios

Authors: Qimao Yang, Huili Chen, Qiwei Dong

Abstract: This report presents a comprehensive study on deep learning models for brand logo classification in real-world scenarios. The dataset contains 3,717 labeled images of logos from ten prominent brands. Two types of models, Convolutional Neural Networks (CNN) and Vision Transformer (ViT), were evaluated for their performance. The ViT model, DaViT small, achieved the highest accuracy of 99.60%, while… ▽ More This report presents a comprehensive study on deep learning models for brand logo classification in real-world scenarios. The dataset contains 3,717 labeled images of logos from ten prominent brands. Two types of models, Convolutional Neural Networks (CNN) and Vision Transformer (ViT), were evaluated for their performance. The ViT model, DaViT small, achieved the highest accuracy of 99.60%, while the DenseNet29 achieved the fastest inference speed of 366.62 FPS. The findings suggest that the DaViT model is a suitable choice for offline applications due to its superior accuracy. This study demonstrates the practical application of deep learning in brand logo classification tasks. △ Less

Submitted 20 May, 2023; originally announced May 2023.

Excitonic and deep-level emission from N- and Al-polar homoepitaxial AlN grown by molecular beam epitaxy

Authors: Len van Deurzen, Jashan Singhal, Jimy Encomendero, Naomi Pieczulewski, Celesta Chang, Yong** Cho, David Anthony Muller, Huili Grace Xing, Debdeep Jena, Oliver Brandt, Jonas Lähnemann

Abstract: Using low-temperature cathodoluminescence spectroscopy, we study the properties of N- and Al-polar AlN layers grown by molecular beam epitaxy on bulk AlN{0001}. Compared to the bulk AlN substrate, layers of both polarities feature a suppression of deep-level luminescence, a total absence of the prevalent donor with an exciton binding energy of 28 meV, and a much increased intensity of the emission… ▽ More Using low-temperature cathodoluminescence spectroscopy, we study the properties of N- and Al-polar AlN layers grown by molecular beam epitaxy on bulk AlN{0001}. Compared to the bulk AlN substrate, layers of both polarities feature a suppression of deep-level luminescence, a total absence of the prevalent donor with an exciton binding energy of 28 meV, and a much increased intensity of the emission from free excitons. The dominant donor in these layers is characterized by an associated exciton binding energy of 13 meV. The observation of excited exciton states up to the exciton continuum allows us to directly extract the $Γ_{5}$ free exciton binding energy of 57 meV. △ Less

Submitted 23 July, 2023; v1 submitted 17 May, 2023; originally announced May 2023.

Journal ref: APL Materials 11, 081109 (2023)

Ore Extension of Group-cograded Hopf Coquasigroups

Authors: Lingli Zhu, Bingbing **, Huili Liu, Tao Yang

Abstract: The aim of this paper is the Ore extension of group-cograded Hopf coquasigroups. This paper first shows a categorical interpretation and some examples of group-cograded Hopf coquasigroups, and then gives a necessary and sufficient conditions for the Ore extensions of group-cograded Hopf coquasigroups to be group-cograded Hopf coquasigroups. Finally, a certain isomorphism between Ore extensions are… ▽ More The aim of this paper is the Ore extension of group-cograded Hopf coquasigroups. This paper first shows a categorical interpretation and some examples of group-cograded Hopf coquasigroups, and then gives a necessary and sufficient conditions for the Ore extensions of group-cograded Hopf coquasigroups to be group-cograded Hopf coquasigroups. Finally, a certain isomorphism between Ore extensions are considered. △ Less

Submitted 11 July, 2023; v1 submitted 12 April, 2023; originally announced April 2023.

Comments: 15pages

MSC Class: 16T05; 16S36

Defeating Broken Symmetry with Do**: Symmetric Resonant Tunneling in Noncentrosymetric Heterostructures

Authors: Jimy Encomendero, Vladimir Protasenko, Debdeep Jena, Huili Grace Xing

Abstract: Resonant tunneling transport in polar heterostructures is intimately connected to the polarization fields emerging from the geometric Berry-phase. In these structures, quantum confinement results not only in a discrete electronic spectrum, but also in built-in polarization charges exhibiting a broken inversion symmetry along the transport direction. Thus, electrons undergo highly asymmetric quantu… ▽ More Resonant tunneling transport in polar heterostructures is intimately connected to the polarization fields emerging from the geometric Berry-phase. In these structures, quantum confinement results not only in a discrete electronic spectrum, but also in built-in polarization charges exhibiting a broken inversion symmetry along the transport direction. Thus, electrons undergo highly asymmetric quantum interference effects with respect to the direction of current flow. By employing do** to counter the broken symmetry, we deterministically control the resonant transmission through GaN/AlN resonant tunneling diodes and experimentally demonstrate the recovery of symmetric resonant tunneling injection across the noncentrosymmetric double-barrier potential. △ Less

Submitted 15 March, 2023; originally announced March 2023.

Comments: 6 pages, 4 figures

Journal ref: Physical Review B 107, 125301 (2023)

Fighting Broken Symmetry with Do**: Toward Polar Resonant Tunneling Diodes with Symmetric Characteristics

Authors: Jimy Encomendero, Vladimir Protasenko, Farhan Rana, Debdeep Jena, Huili Grace Xing

Abstract: The recent demonstration of resonant tunneling transport in nitride semiconductors has led to an invigorated effort to harness this quantum transport regime for practical applications. In polar semiconductors, however, the interplay between fixed polarization charges and mobile free carriers leads to asymmetric transport characteristics. Here, we investigate the possibility of using degenerately d… ▽ More The recent demonstration of resonant tunneling transport in nitride semiconductors has led to an invigorated effort to harness this quantum transport regime for practical applications. In polar semiconductors, however, the interplay between fixed polarization charges and mobile free carriers leads to asymmetric transport characteristics. Here, we investigate the possibility of using degenerately doped contact layers to screen the built-in polarization fields and recover symmetric resonant injection. Thanks to a high do** density, negative differential conductance is observed under both bias polarities of GaN/AlN resonant tunneling diodes (RTDs). Moreover, our analytical model reveals a lower bound for the minimum resonant-tunneling voltage achieved via uniform do**, owing to the dopant solubility limit. Charge storage dynamics is also studied by impedance measurements, showing that at close-to-equilibrium conditions, polar RTDs behave effectively as parallel-plate capacitors. These mechanisms are completely reproduced by our analytical model, providing a theoretical framework useful in the design and analysis of polar resonant-tunneling devices. △ Less

Submitted 15 March, 2023; originally announced March 2023.

Comments: 9 pages, 4 figures

Journal ref: Physical Review Applied 13, 034048 (2020)

Broken Symmetry Effects due to Polarization on Resonant Tunneling Transport in Double-Barrier Nitride Heterostructures

Authors: Jimy Encomendero, Vladimir Protasenko, Berardi Sensale-Rodriguez, Patrick Fay, Farhan Rana, Debdeep Jena, Huili Grace Xing

Abstract: The phenomenon of resonant tunneling transport through polar double-barrier heterostructures is systematically investigated using a combined experimental and theoretical approach. On the experimental side, GaN/AlN RTDs are grown by MBE. In-situ electron diffraction is employed to monitor the number of monolayers incorporated into each tunneling barrier. Using this precise epitaxial control at the… ▽ More The phenomenon of resonant tunneling transport through polar double-barrier heterostructures is systematically investigated using a combined experimental and theoretical approach. On the experimental side, GaN/AlN RTDs are grown by MBE. In-situ electron diffraction is employed to monitor the number of monolayers incorporated into each tunneling barrier. Using this precise epitaxial control at the monolayer level, we demonstrate exponential modulation of the resonant tunneling current as a function of barrier thickness. Both the peak voltage and characteristic threshold bias exhibit a dependence on barrier thickness as a result of the intense electric fields present in the polar heterostructures. To get further insight into the asymmetric tunneling injection, we present an analytical theory for tunneling transport across polar heterostructures. A general expression for the resonant tunneling current with contributions from coherent and sequential tunneling processes is introduced. After applying this theory to the case of GaN/AlN RTDs, their experimental current-voltage characteristics are reproduced over both bias polarities, with tunneling currents spanning several orders of magnitude. This agreement allows us to elucidate the role played by the internal polarization fields on the magnitude of the tunneling current and broadening of the resonant line shape. Under reverse bias, we identify new tunneling features originating from highly attenuated resonant tunneling phenomena, which are completely captured by our model. Our analytical model, provides a simple expression which reveals the connection between the polar RTD design parameters and its current-voltage characteristics. This new theory paves the way for the design of polar resonant tunneling devices exhibiting efficient resonant current injection and enhanced tunneling dynamics, as required in various practical applications. △ Less

Submitted 15 March, 2023; originally announced March 2023.

Comments: 13 pages, 3 figures

Journal ref: Physical Review Applied 11, 034032 (2019)

Universal Three-Body Physics for Ultracold $^{87}$Rb-$^{40}$K Mixtures with Finite-range Interactions

Authors: Ning-Ning Gao, Hui-Li Han, Ting-Yun Shi

Abstract: The breakdown of Efimov universal relations between resonances and interference features in ultracold $^{40}$K-$^{87}$Rb mixtures has sparked much discussion. In this study, we investigate the atom-dimer elastic scattering and three-body recombination (TBR) rates with $J=0$ symmetry for a $^{87}$Rb-$^{87}$Rb-$^{40}$K system to understand the mechanisms underlying this breakdown. We use the R-matri… ▽ More The breakdown of Efimov universal relations between resonances and interference features in ultracold $^{40}$K-$^{87}$Rb mixtures has sparked much discussion. In this study, we investigate the atom-dimer elastic scattering and three-body recombination (TBR) rates with $J=0$ symmetry for a $^{87}$Rb-$^{87}$Rb-$^{40}$K system to understand the mechanisms underlying this breakdown. We use the R-matrix propagation method in a hyperspherical coordinate frame with the Lennard-Jones model potential describing the interactions between atoms. We extract the atom-dimer scattering length, cross sections and three-body recombination rates with a finite intraspecies scattering length. An atom-dimer elastic scattering resonance at negative interspecies scattering length is found, which is due to the presence of the near-threshold trimer state at the atom-dimer threshold ($^{40}$K + $^{87}$Rb$^{87}$Rb). In addition, we predict an Efimov recombination minima at positive interspecies scattering length, which is outside the range of previous measurements. Our calculations show that the breakdown of the universal relations predicted by zero-range theory can be attributed to the effective range corrections and finite intraspecies scattering length. The present results help elucidate the universal relations between Efimov features as well as the importance of finite-range corrections for mixed systems. △ Less

Submitted 19 February, 2023; v1 submitted 16 February, 2023; originally announced February 2023.

Comments: 20 pages,10 figures

Crowd3D: Towards Hundreds of People Reconstruction from a Single Image

Authors: Hao Wen, **g Huang, Huili Cui, Haozhe Lin, YuKun Lai, Lu Fang, Kun Li

Abstract: Image-based multi-person reconstruction in wide-field large scenes is critical for crowd analysis and security alert. However, existing methods cannot deal with large scenes containing hundreds of people, which encounter the challenges of large number of people, large variations in human scale, and complex spatial distribution. In this paper, we propose Crowd3D, the first framework to reconstruct… ▽ More Image-based multi-person reconstruction in wide-field large scenes is critical for crowd analysis and security alert. However, existing methods cannot deal with large scenes containing hundreds of people, which encounter the challenges of large number of people, large variations in human scale, and complex spatial distribution. In this paper, we propose Crowd3D, the first framework to reconstruct the 3D poses, shapes and locations of hundreds of people with global consistency from a single large-scene image. The core of our approach is to convert the problem of complex crowd localization into pixel localization with the help of our newly defined concept, Human-scene Virtual Interaction Point (HVIP). To reconstruct the crowd with global consistency, we propose a progressive reconstruction network based on HVIP by pre-estimating a scene-level camera and a ground plane. To deal with a large number of persons and various human sizes, we also design an adaptive human-centric crop** scheme. Besides, we contribute a benchmark dataset, LargeCrowd, for crowd reconstruction in a large scene. Experimental results demonstrate the effectiveness of the proposed method. The code and datasets will be made public. △ Less

Submitted 1 April, 2023; v1 submitted 23 January, 2023; originally announced January 2023.

Comments: Accepted by CVPR 2023

Joint Engagement Classification using Video Augmentation Techniques for Multi-person Human-robot Interaction

Authors: Yubin Kim, Huili Chen, Sharifa Alghowinem, Cynthia Breazeal, Hae Won Park

Abstract: Affect understanding capability is essential for social robots to autonomously interact with a group of users in an intuitive and reciprocal way. However, the challenge of multi-person affect understanding comes from not only the accurate perception of each user's affective state (e.g., engagement) but also the recognition of the affect interplay between the members (e.g., joint engagement) that p… ▽ More Affect understanding capability is essential for social robots to autonomously interact with a group of users in an intuitive and reciprocal way. However, the challenge of multi-person affect understanding comes from not only the accurate perception of each user's affective state (e.g., engagement) but also the recognition of the affect interplay between the members (e.g., joint engagement) that presents as complex, but subtle, nonverbal exchanges between them. Here we present a novel hybrid framework for identifying a parent-child dyad's joint engagement by combining a deep learning framework with various video augmentation techniques. Using a dataset of parent-child dyads reading storybooks together with a social robot at home, we first train RGB frame- and skeleton-based joint engagement recognition models with four video augmentation techniques (General Aug, DeepFake, CutOut, and Mixed) applied datasets to improve joint engagement classification performance. Second, we demonstrate experimental results on the use of trained models in the robot-parent-child interaction context. Third, we introduce a behavior-based metric for evaluating the learned representation of the models to investigate the model interpretability when recognizing joint engagement. This work serves as the first step toward fully unlocking the potential of end-to-end video understanding models pre-trained on large public datasets and augmented with data augmentation and visualization techniques for affect recognition in the multi-person human-robot interaction in the wild. △ Less

Submitted 28 December, 2022; originally announced December 2022.

Silicon-doped $β$-Ga$_2$O$_3$ films grown at 1 $μ$m/h by suboxide molecular-beam epitaxy

Authors: Kathy Azizie, Felix V. E. Hensling, Cameron A. Gorsak, Yunjo Kim, Daniel M. Dryden, M. K. Indika Senevirathna, Selena Coye, Shun-Li Shang, Jacob Steele, Patrick Vogt, Nicholas A. Parker, Yorick A. Birkhölzer, Jonathan P. McCandless, Debdeep Jena, Huili G. Xing, Zi-Kui Liu, Michael D. Williams, Andrew J. Green, Kelson Chabak, Adam T. Neal, Shin Mou, Michael O. Thompson, Hari P. Nair, Darrell G. Schlom

Abstract: We report the use of suboxide molecular-beam epitaxy (S-MBE) to grow $β$-Ga$_2$O$_3$ at a growth rate of ~1 $μ$m/h with control of the silicon do** concentration from 5x10$^{16}$ to 10$^{19}$ cm$^{-3}$. In S-MBE, pre-oxidized gallium in the form of a molecular beam that is 99.98\% Ga$_2$O, i.e., gallium suboxide, is supplied. Directly supplying Ga2O to the growth surface bypasses the rate-limiti… ▽ More We report the use of suboxide molecular-beam epitaxy (S-MBE) to grow $β$-Ga$_2$O$_3$ at a growth rate of ~1 $μ$m/h with control of the silicon do** concentration from 5x10$^{16}$ to 10$^{19}$ cm$^{-3}$. In S-MBE, pre-oxidized gallium in the form of a molecular beam that is 99.98\% Ga$_2$O, i.e., gallium suboxide, is supplied. Directly supplying Ga2O to the growth surface bypasses the rate-limiting first step of the two-step reaction mechanism involved in the growth of $β$-Ga$_2$O$_3$ by conventional MBE. As a result, a growth rate of ~1 $μ$m/h is readily achieved at a relatively low growth temperature (T$_{sub}$ = 525 $^\circ$C), resulting in films with high structural perfection and smooth surfaces (rms roughness of < 2 nm on ~1 $μ$m thick films). Silicon-containing oxide sources (SiO and SiO$_2$) producing an SiO suboxide molecular beam are used to dope the $β$-Ga$_2$O$_3$ layers. Temperature-dependent Hall effect measurements on a 1 $μ$m thick film with a mobile carrier concentration of 2.7x10$^{17}$ cm$^{-3}$ reveal a room-temperature mobility of 124 cm$^2$ V$^{-1}$ s$^{-1}$ that increases to 627 cm$^2$ V$^{-1}$ s$^{-1}$ at 76 K; the silicon dopants are found to exhibit an activation energy of 27 meV. We also demonstrate working MESFETs made from these silicon-doped $β$-Ga$_2$O$_3$ films grown by S-MBE at growth rates of ~1 $μ$m/h. △ Less

Submitted 22 December, 2022; originally announced December 2022.

Comments: 19 pages, 7 figures, 2 tables, 2 pages supplementary materials

Topological Hall Effect Driven by Short-Range Magnetic Orders in EuZn$_2$As$_2$

Authors: Enkui Yi, Dong Feng Zheng, Feihao Pan, Hongxia Zhang, Bin Wang, Bowen Chen, Detong Wu, Huili Liang, Zeng Xia Mei, Hao Wu, Shengyuan A. Yang, Peng Cheng, Meng Wang, Bing Shen

Abstract: Short-range (SR) magnetic orders such as magnetic glass orders or fluctuations in a quantum system usually host exotic states or critical behaviors. As the long-range (LR) magnetic orders, SR magnetic orders can also break time-reversal symmetry and drive the non-zero Berry curvature leading to novel transport properties. In this work, we report that in EuZn$_2$As$_2$ compound, besides the LR A-ty… ▽ More Short-range (SR) magnetic orders such as magnetic glass orders or fluctuations in a quantum system usually host exotic states or critical behaviors. As the long-range (LR) magnetic orders, SR magnetic orders can also break time-reversal symmetry and drive the non-zero Berry curvature leading to novel transport properties. In this work, we report that in EuZn$_2$As$_2$ compound, besides the LR A-type antiferromagnetic (AF) order, the SR magnetic order is observed in a wide temperature region. The magnetization measurements and electron spin resonance (ESR) measurements reveal the ferromagnetic (FM) correlations for this SR magnetic order which results in an obvious anomalous Hall effect above the AF transition. Moreover the ESR results reveal that this FM SR order coexists with LR AF order exhibiting anisotropic magnetic correlations below the AF transition. The interactions of LR and SR magnetism evolving with temperature and field can host non-zero spin charility and berry curvature leading the additional topological Hall contribution even in a centrosymmetric simple AF system. Our results indicate that EuZn$_2$As$_2$ is a fertile platform to investigate exotic magnetic and electronic states. △ Less

Submitted 24 January, 2023; v1 submitted 10 December, 2022; originally announced December 2022.

Comments: 6 pages, 4 figures

Journal ref: Phys. Rev. B 107, 035142 (2023)

Braided crossed category over crossed group-cograded weak Hopf quasigroups

Authors: Huili Liu, Lingli Zhu, Tao Yang

Abstract: In this paper, we generalizing the main result in Liu[10] to weak Hopf coquasigroups case. We first define and study group-cograded weak Hopf quasigroups, which generalize both group-cograded Hopf quasigroups and weak Hopf group-coalgebras. Then we introduce the notion of p-Yetter-Drinfeld weak quasimodule over group-cograded weak Hopf quasigroups H. If the antipode of H is bijective, we show that… ▽ More In this paper, we generalizing the main result in Liu[10] to weak Hopf coquasigroups case. We first define and study group-cograded weak Hopf quasigroups, which generalize both group-cograded Hopf quasigroups and weak Hopf group-coalgebras. Then we introduce the notion of p-Yetter-Drinfeld weak quasimodule over group-cograded weak Hopf quasigroups H. If the antipode of H is bijective, we show that the category YDWQ(H) of Yetter-Drinfeld weak quasimodules over H is a crossed category, and the subcategory YD(H) of Yetter-Drinfeld modules is a braided crossed category. △ Less

Submitted 1 December, 2022; originally announced December 2022.

Comments: 19pages

MSC Class: 16T05; 17A01; 18M15

Neutron scattering studies on ionic diffusion behaviors of superionic $α$-Cu$_{2-δ}$Se

Authors: Lisi Li, Huili Liu, Maxim Avdeev, Dehong Yu, Sergey Danilkin, Meng Wang

Abstract: We present studies on crystal structure and ionic diffusion behaviors of superionic Cu$_{2-δ}$Se ($δ$=0, 0.04, and 0.2) by utilizing neutron powder diffraction and quasi-elastic neutron scattering. In the superionic phase, the structural model with Cu ions occupying the Wyckoff sites of 8$c$ and 32$f$ provides the best description of the structure. As the content of Cu increasing in Cu$_{2-δ}$Se,… ▽ More We present studies on crystal structure and ionic diffusion behaviors of superionic Cu$_{2-δ}$Se ($δ$=0, 0.04, and 0.2) by utilizing neutron powder diffraction and quasi-elastic neutron scattering. In the superionic phase, the structural model with Cu ions occupying the Wyckoff sites of 8$c$ and 32$f$ provides the best description of the structure. As the content of Cu increasing in Cu$_{2-δ}$Se, the Cu occupancy increases on the 32$f$ site, but decreases on the 8$c$ site . Fitting to the quasi-elastic neutron scattering spectra reveals two diffusion modes, the localized diffusion between the 8$c$ and 32$f$ sites and the long-range diffusion between the adjacent 8$c$ sites using the 32$f$ site as a bypass, respectively. Between 430 and 650 K, we measured that the compound with more Cu content exhibits a larger long-range diffusion coefficient. Temperature in this range does not affect the long-range diffusion process obviously. Our results suggest the two diffusion modes cooperative and thus provide a microscopic understanding of the ionic diffusion of the Cu ions in superionic Cu$_{2-δ}$Se. △ Less

Submitted 10 October, 2022; originally announced October 2022.

Comments: 6 pages, 3 figures

Journal ref: Appl. Phys. Lett. 121, 151901 (2022)

An automated approach for consecutive tuning of quantum dot arrays

Authors: Hanwei Liu, Baochuan Wang, Ning Wang, Zhonghai Sun, Huili Yin, Haiou Li, Gang Cao, Guo** Guo

Abstract: Recent progress has shown that the dramatically increased number of parameters has become a major issue in tuning of multi-quantum dot devices. The complicated interactions between quantum dots and gate electrodes cause the manual tuning process to no longer be efficient. Fortunately, machine learning techniques can automate and speed up the tuning of simple quantum dot systems. In this letter, we… ▽ More Recent progress has shown that the dramatically increased number of parameters has become a major issue in tuning of multi-quantum dot devices. The complicated interactions between quantum dots and gate electrodes cause the manual tuning process to no longer be efficient. Fortunately, machine learning techniques can automate and speed up the tuning of simple quantum dot systems. In this letter, we extend the techniques to tune multi-dot devices. We propose an automated approach that combines machine learning, virtual gates and a local-to-global method to realize the consecutive tuning of quantum dot arrays by dividing them into subsystems. After optimizing voltage configurations and establishing virtual gates to control each subsystem independently, a quantum dot array can be efficiently tuned to the few-electron regime with appropriate interdot tunnel coupling strength. Our experimental results show that this approach can consecutively tune quantum dot arrays into an appropriate voltage range without human intervention and possesses broad application prospects in large-scale quantum dot devices. △ Less

Submitted 25 August, 2022; originally announced August 2022.

Comments: 11 pages,4 figures

Journal ref: Appl. Phys. Lett. 121, 084002 (2022)

HZO-based FerroNEMS MAC for In-Memory Computing

Authors: Shubham Jadhav, Ved Gund, Benyamin Davaji, Debdeep Jena, Huili, Xing, Amit Lal

Abstract: This paper demonstrates a hafnium zirconium oxide (HZO)-based ferroelectric NEMS unimorph as the fundamental building block for very low-energy capacitive readout in-memory computing. The reported device consists of a 250 $μ$m $\times$ 30 $μ$m unimorph cantilever with 20 nm thick ferroelectric HZO on 1 $μ$m $SiO_2$.Partial ferroelectric switching in HZO achieves analog programmable control of the… ▽ More This paper demonstrates a hafnium zirconium oxide (HZO)-based ferroelectric NEMS unimorph as the fundamental building block for very low-energy capacitive readout in-memory computing. The reported device consists of a 250 $μ$m $\times$ 30 $μ$m unimorph cantilever with 20 nm thick ferroelectric HZO on 1 $μ$m $SiO_2$.Partial ferroelectric switching in HZO achieves analog programmable control of the piezoelectric coefficient ($d_{31}$) which serves as the computational weight for multiply-accumulate (MAC) operations. The displacement of the piezoelectric unimorph was recorded by actuating the device with different input voltages $V_{in}$. The resulting displacement was measured as a function of the ferroelectric programming/poling voltage $V_p$. The slopes of central beam displacement ($δ_{max}$) vs $V_{in}$ were measured to be between 182.9nm/V (for -8 $V_p$) and -90.5nm/V (for 8 $V_p$), demonstrating that $V_p$ can be used to change the direction of motion of the beam. The resultant ($δ_{max}$) from AC actuation is in the range of -18 to 36 nm and is a scaled product of the input voltage and programmed $d_{31}$ (governed by the $V_p$). The multiplication function serves as the fundamental unit for MAC operations with the ferroelectric NEMS unimorph. The displacement from many such beams can be added by summing the capacitance changes, providing a pathway to implement a multi-input and multi-weight neuron. A scaling and fabrication analysis suggests that this device can be CMOS compatible, achieving high in-memory computational throughput. △ Less

Submitted 12 August, 2022; originally announced August 2022.

PerD: Perturbation Sensitivity-based Neural Trojan Detection Framework on NLP Applications

Authors: Diego Garcia-soto, Huili Chen, Farinaz Koushanfar

Abstract: Deep Neural Networks (DNNs) have been shown to be susceptible to Trojan attacks. Neural Trojan is a type of targeted poisoning attack that embeds the backdoor into the victim and is activated by the trigger in the input space. The increasing deployment of DNNs in critical systems and the surge of outsourcing DNN training (which makes Trojan attack easier) makes the detection of Trojan attacks nece… ▽ More Deep Neural Networks (DNNs) have been shown to be susceptible to Trojan attacks. Neural Trojan is a type of targeted poisoning attack that embeds the backdoor into the victim and is activated by the trigger in the input space. The increasing deployment of DNNs in critical systems and the surge of outsourcing DNN training (which makes Trojan attack easier) makes the detection of Trojan attacks necessary. While Neural Trojan detection has been studied in the image domain, there is a lack of solutions in the NLP domain. In this paper, we propose a model-level Trojan detection framework by analyzing the deviation of the model output when we introduce a specially crafted perturbation to the input. Particularly, we extract the model's responses to perturbed inputs as the `signature' of the model and train a meta-classifier to determine if a model is Trojaned based on its signature. We demonstrate the effectiveness of our proposed method on both a dataset of NLP models we create and a public dataset of Trojaned NLP models from TrojAI. Furthermore, we propose a lightweight variant of our detection method that reduces the detection time while preserving the detection rates. △ Less

Submitted 8 August, 2022; originally announced August 2022.

Covariant-Contravariant Refinement Modal $μ$-calculus

Authors: Huili Xing

Abstract: The notion of covariant-contravariant refinement (CC-refinement, for short) is a generalization of the notions of bisimulation, simulation and refinement. This paper introduces CC-refinement modal $μ$-calculus (CCRML$^μ$) obtained from the modal $μ$-calculus system K$^μ$ by adding CC-refinement quantifiers, establishes an axiom system for CCRML$^μ$ and explores the important properties: soundness,… ▽ More The notion of covariant-contravariant refinement (CC-refinement, for short) is a generalization of the notions of bisimulation, simulation and refinement. This paper introduces CC-refinement modal $μ$-calculus (CCRML$^μ$) obtained from the modal $μ$-calculus system K$^μ$ by adding CC-refinement quantifiers, establishes an axiom system for CCRML$^μ$ and explores the important properties: soundness, completeness and decidability of this axiom system. The language of CCRML$^μ$ may be considered as a specification language for describing the properties of a system referring to reactive and generative actions. It may be used to formalize some interesting problems in the field of formal methods. △ Less

Submitted 5 August, 2022; originally announced August 2022.

An extension of process calculus for asynchronous communications between agents with epistemic states

Authors: Huili Xing

Abstract: It plays a central role in intelligent agent systems to model agent's epistemic state and its change. Asynchrony plays a key role in distributed systems, in which the messages transmitted may not be received instantly by the agents. To characterize asynchronous communications, asynchronous announcement logic (AAL) has been presented, which focuses on the logic laws of the change of epistemic state… ▽ More It plays a central role in intelligent agent systems to model agent's epistemic state and its change. Asynchrony plays a key role in distributed systems, in which the messages transmitted may not be received instantly by the agents. To characterize asynchronous communications, asynchronous announcement logic (AAL) has been presented, which focuses on the logic laws of the change of epistemic state after receiving information. However AAL does not involve the interactive behaviours between an agent and its environment. Through enriching the well-known pi-calculus by adding the operators for passing basic facts and applying the well-known action model logic to describe agents' epistemic states, this paper presents the e-calculus to model epistemic interactions between agents with epistemic states. The e-calculus can be adopted to characterize synchronous and asynchronous communications between agents. To capture the asynchrony, a buffer pools is constructed to store the basic facts announced and each agent reads these facts from this buffer pool in some order. Based on the transmission of link names, the e-calculus is able to realize reading from this buffer pool in different orders. This paper gives two examples: one is to read in the order in which the announced basic facts are sent (First-in-first-out, FIFO), and the other is in an arbitrary order. △ Less

Submitted 24 February, 2023; v1 submitted 29 June, 2022; originally announced June 2022.

Comments: 22 pages and 2 figures

Decoherence by Optical Phonons in GaN Defect Single-Photon Emitters

Authors: Yifei Geng, Jialun Luo, Len van Deurzen, Huili, Xing, Debdeep Jena, Gregory David Fuchs, Farhan Rana

Abstract: In most single-photon defect emitters, such as those in SiC and diamond, interaction with low-energy acoustic phonons determines the temperature dependence of the decoherence rate and the resulting broadening of the ZPL with the temperature obeys a power law. GaN hosts bright and stable single-photon emitters in the 600 nm to 700 nm wavelength range with strong ZPLs even at room temperature. In th… ▽ More In most single-photon defect emitters, such as those in SiC and diamond, interaction with low-energy acoustic phonons determines the temperature dependence of the decoherence rate and the resulting broadening of the ZPL with the temperature obeys a power law. GaN hosts bright and stable single-photon emitters in the 600 nm to 700 nm wavelength range with strong ZPLs even at room temperature. In this work, we study the temperature dependence of the ZPL spectra of GaN SPEs integrated with solid immersion lenses with the goal of understanding the relevant decoherence mechanisms. At temperatures below ~50 K, the ZPL lineshape is found to be Gaussian and the ZPL linewidth is temperature independent and dominated by spectral diffusion. Above ~50 K, the linewidth increases monotonically with the temperature and the lineshape evolves into a Lorentzian. Quite remarkably, the temperature dependence of the linewidth does not follow a power law. We propose a model in which decoherence caused by absorption/emission of optical phonons in an elastic Raman process determines the temperature dependence of the lineshape and the linewidth. Our model explains the temperature dependence of the ZPL linewidth and lineshape in the entire 10 K to 270 K temperature range explored in this work. The ~19 meV optical phonon energy extracted by fitting the model to the data matches remarkably well the ~18 meV zone center energy of the lowest optical phonon band (E2(low)) in GaN. Our work sheds light on the mechanisms responsible for linewidth broadening in GaN SPEs. Since a low energy optical phonon band (E2(low)) is a feature of most group III-V nitrides with a wurtzite crystal structure, including hBN and AlN, we expect our proposed mechanism to play an important role in defect emitters in these materials as well. △ Less

Submitted 29 January, 2023; v1 submitted 25 June, 2022; originally announced June 2022.

Comments: 11 pages, 7 figures

Molecular Beam Homoepitaxy of N-polar AlN on bulk AlN substrates

Authors: Jashan Singhal, Jimy Encomendero, Yong** Cho, Len van Deurzen, Zexuan Zhang, Kazuki Nomoto, Masato Toita, Huili Grace Xing, Debdeep Jena

Abstract: N-polar AlN epilayers were grown on the N-face of single crystal bulk AlN substrates by plasma assisted molecular beam epitaxy (PA-MBE). A combination of in situ thermal deoxidation and Al-assisted thermal desorption at high temperature helped in removing native surface oxides and impurities from the N-polar surface of the substrate enabling successful homoepitaxy. Subsequent epitaxial growth of A… ▽ More N-polar AlN epilayers were grown on the N-face of single crystal bulk AlN substrates by plasma assisted molecular beam epitaxy (PA-MBE). A combination of in situ thermal deoxidation and Al-assisted thermal desorption at high temperature helped in removing native surface oxides and impurities from the N-polar surface of the substrate enabling successful homoepitaxy. Subsequent epitaxial growth of AlN layer on the in situ cleaned substrates, grown in sufficiently high Al droplet regime, exhibited smooth surface morphologies with clean and wide atomic steps. KOH etch studies confirmed the N-polarity of the homoepitaxial films. Secondary ion mass spectrometry profiles show Si and H impurity concentrations below the noise levels, whereas O and C impurities concentrations of ~ 8x10^{17} atoms/cm^3 and ~ 2x10^{17} atoms/cm^3 are observed respectively. Though the structural defect densities are low, they interestingly appear as inversion domains of different dimensionalities. △ Less

Submitted 22 June, 2022; originally announced June 2022.

Exact modulus of continuities for $Λ$-Fleming-Viot processes with Brownian spatial motion

Authors: Huili Liu, Xiaowen Zhou

Abstract: For a class of $Λ$-Fleming-Viot processes with Brownian spatial motion in $\mathbb{R}^d$ whose associated $Λ$-coalescents come down from infinity, we obtain sharp global and local modulus of continuities for the ancestry processes recovered from the lookdown representations. As applications, we prove both global and local modulus of continuities for the $Λ$-Fleming-Viot support processes. In parti… ▽ More For a class of $Λ$-Fleming-Viot processes with Brownian spatial motion in $\mathbb{R}^d$ whose associated $Λ$-coalescents come down from infinity, we obtain sharp global and local modulus of continuities for the ancestry processes recovered from the lookdown representations. As applications, we prove both global and local modulus of continuities for the $Λ$-Fleming-Viot support processes. In particular, if the $Λ$-coalescent is the Beta$(2-β,β)$ coalescent for $β\in(1,2]$ with $β=2$ corresponding to Kingman's coalescent, then for $h(t)=\sqrt{t\log (1/t)}$, the global modulus of continuity holds for the support process with modulus function $\sqrt{2β/(β-1)}h(t)$, and both the left and right local modulus of continuities hold for the support process with modulus function $\sqrt{2/(β-1)}h(t)$. △ Less

Submitted 30 January, 2023; v1 submitted 17 June, 2022; originally announced June 2022.

Comments: 27 pages

MSC Class: 2020 subject classifications: Primary 60J68; 60G17; secondary 60G57; 60J95

Interfacial properties of 2D WS2 on SiO2 substrate from x-ray photoelectron spectroscopy and first-principles calculations

Authors: Changjie Zhou, Huili Zhu, Weifeng Yang, Qiubao Lin, Tongchang Zheng, Lan Yang, Shuqiong Lan

Abstract: Two-dimensional (2D) WS2 films were deposited on SiO2 wafers, and the related interfacial properties were investigated by high-resolution x-ray photoelectron spectroscopy (XPS) and first-principles calculations. Using the direct (indirect) method, the valence band offset (VBO) at monolayer WS2/SiO2 interface was found to be 3.97 eV (3.86 eV), and the conduction band offset (CBO) was 2.70 eV (2.81… ▽ More Two-dimensional (2D) WS2 films were deposited on SiO2 wafers, and the related interfacial properties were investigated by high-resolution x-ray photoelectron spectroscopy (XPS) and first-principles calculations. Using the direct (indirect) method, the valence band offset (VBO) at monolayer WS2/SiO2 interface was found to be 3.97 eV (3.86 eV), and the conduction band offset (CBO) was 2.70 eV (2.81 eV). Furthermore, the VBO (CBO) at bulk WS2/SiO2 interface is found to be about 0.48 eV (0.33 eV) larger due to the interlayer orbital coupling and splitting of valence and conduction band edges. Therefore, the WS2/SiO2 heterostructure has a Type I energy-band alignment. The band offsets obtained experimentally and theoretically are consistent except the narrower theoretical bandgap of SiO2. The theoretical calculations further reveal a binding energy of 75 meV per S atom and the totally separated partial density of states, indicating a weak interaction and negligible Fermi level pinning effect between WS2 monolayer and SiO2 surface. Our combined experimental and theoretical results provide proof of the sufficient VBOs and CBOs and weak interaction in 2D WS2/SiO2 heterostructures. △ Less

Submitted 18 May, 2022; originally announced May 2022.

Antiferromagnetic Spin Orientation and Magnetic Domain Structure in Epitaxially Grown MnN Studied using Optical Second Harmonic Generation

Authors: Joongwon Lee, Zexuan Zhang, Huili Xing, Debdeep Jena, Farhan Rana

Abstract: MnN is a centrosymmetric collinear antiferromagnet belonging to the transition metal nitride family with a high Neel temperature, a low anisotropy field, and a large magnetic moment per Mn atom. Despite several recent experimental and theoretical studies, the spin symmetry (magnetic point group) and magnetic domain structure of the material remain unknown. In this work, we use optical second harmo… ▽ More MnN is a centrosymmetric collinear antiferromagnet belonging to the transition metal nitride family with a high Neel temperature, a low anisotropy field, and a large magnetic moment per Mn atom. Despite several recent experimental and theoretical studies, the spin symmetry (magnetic point group) and magnetic domain structure of the material remain unknown. In this work, we use optical second harmonic generation (SHG) to study the magnetic structure of thin epitaxially-grown single-crystal (001) MnN films. Our work shows that spin moments in MnN are tilted away from the [001] direction and the components of the spin moments in the (001) plane are aligned along one of the two possible in-plane symmetry axes ([100] or [110]) resulting in a magnetic point group symmetry of 2/m1'. Our work rules out magnetic point group symmetries 4/mmm1' and mmm1' that have been previously discussed in the literature. Four different spin domains consistent with the 2/m1' magnetic point group symmetry are possible in MnN. A statistical model based on the observed variations in the polarization-dependent intensity of the second harmonic signal collected over large sample areas puts an upper bound of 0.65 microns on the mean domain size. Our results show that SHG can be used to probe the magnetic order in metallic antiferromagnets. This work is expected to contribute to the recent efforts in using antiferromagnets for spintronic applications. △ Less

Submitted 25 June, 2022; v1 submitted 25 April, 2022; originally announced April 2022.

Comments: 7 pages, 4 figures

N-polar GaN p-n junction diodes with low ideality factors

Authors: Kazuki Nomoto, Huili Grace Xing, Debdeep Jena, Yong** Cho

Abstract: High-quality N-polar GaN p-n diodes are realized on single-crystal N-polar GaN bulk substrate by plasma-assisted molecular beam epitaxy. The room-temperature current-voltage characteristics reveal a high on/off current ratio of 10^11 at 4 V and an ideality factor of 1.6. As the temperature increases to 200 C, the apparent ideality factor gradually approaches 2. At such high temperatures, Shockley-… ▽ More High-quality N-polar GaN p-n diodes are realized on single-crystal N-polar GaN bulk substrate by plasma-assisted molecular beam epitaxy. The room-temperature current-voltage characteristics reveal a high on/off current ratio of 10^11 at 4 V and an ideality factor of 1.6. As the temperature increases to 200 C, the apparent ideality factor gradually approaches 2. At such high temperatures, Shockley-Read-Hall recombination times of 0.32-0.46 ns are estimated. The measured electroluminescence spectrum is dominated by a strong near-band edge emission, while deep level and acceptor-related luminescence is greatly suppressed. A relatively high reverse breakdown field of 2.4 MV/cm without field-plates is achieved. This work indicates that the quality of N-polar GaN diodes is now approaching to that of their state-of-the-art Ga-polar counterparts. △ Less

Submitted 4 May, 2022; v1 submitted 24 April, 2022; originally announced April 2022.

Comments: 11 pages, 3 figures

Journal ref: Applied Physics Express 15, 064004 (2022)

Molecular beam homoepitaxy of N-polar AlN: enabling role of Al-assisted surface cleaning

Authors: Zexuan Zhang, Yusuke Hayashi, Tetsuya Tohei, Akira Sakai, Vladimir Protasenko, Jashan Singhal, Hideto Miyake, Huili Grace Xing, Debdeep Jena, Yong** Cho

Abstract: N-polar aluminum nitride (AlN) is an important building block for next-generation high-power RF electronics. We report successful homoepitaxial growth of N-polar AlN by molecular beam epitaxy (MBE) on large-area cost-effective N-polar AlN templates. Direct growth without any in-situ surface cleaning leads to films with inverted Al-polarity. It is found that Al-assisted cleaning before growth enabl… ▽ More N-polar aluminum nitride (AlN) is an important building block for next-generation high-power RF electronics. We report successful homoepitaxial growth of N-polar AlN by molecular beam epitaxy (MBE) on large-area cost-effective N-polar AlN templates. Direct growth without any in-situ surface cleaning leads to films with inverted Al-polarity. It is found that Al-assisted cleaning before growth enables the epitaxial film to maintain N-polarity. The grown N-polar AlN epilayer with its smooth, pit-free surface duplicates the structural quality of the substrate as evidenced by a clean and smooth growth interface with no noticeable extended defects generation. Near band-edge photoluminescence peaks are observed at room temperature on samples with MBE-grown layers but not on the bare AlN substrates, implying the suppression of non-radiative recombination centers in the epitaxial N-polar AlN. These results are pivotal steps towards future high-power RF electronics and deep ultraviolet photonics based on the N-polar AlN platform. △ Less

Submitted 18 April, 2022; originally announced April 2022.

Self-Aware Personalized Federated Learning

Authors: Huili Chen, Jie Ding, Eric Tramel, Shuang Wu, Anit Kumar Sahu, Salman Avestimehr, Tao Zhang

Abstract: In the context of personalized federated learning (FL), the critical challenge is to balance local model improvement and global model tuning when the personal and global objectives may not be exactly aligned. Inspired by Bayesian hierarchical models, we develop a self-aware personalized FL method where each client can automatically balance the training of its local personal model and the global mo… ▽ More In the context of personalized federated learning (FL), the critical challenge is to balance local model improvement and global model tuning when the personal and global objectives may not be exactly aligned. Inspired by Bayesian hierarchical models, we develop a self-aware personalized FL method where each client can automatically balance the training of its local personal model and the global model that implicitly contributes to other clients' training. Such a balance is derived from the inter-client and intra-client uncertainty quantification. A larger inter-client variation implies more personalization is needed. Correspondingly, our method uses uncertainty-driven local training steps and aggregation rule instead of conventional local fine-tuning and sample size-based aggregation. With experimental studies on synthetic data, Amazon Alexa audio data, and public datasets such as MNIST, FEMNIST, CIFAR10, and Sent140, we show that our proposed method can achieve significantly improved personalization performance compared with the existing counterparts. △ Less

Submitted 17 April, 2022; originally announced April 2022.

AdaTest:Reinforcement Learning and Adaptive Sampling for On-chip Hardware Trojan Detection

Authors: Huili Chen, Xinqiao Zhang, Ke Huang, Farinaz Koushanfar

Abstract: This paper proposes AdaTest, a novel adaptive test pattern generation framework for efficient and reliable Hardware Trojan (HT) detection. HT is a backdoor attack that tampers with the design of victim integrated circuits (ICs). AdaTest improves the existing HT detection techniques in terms of scalability and accuracy of detecting smaller Trojans in the presence of noise and variations. To achieve… ▽ More This paper proposes AdaTest, a novel adaptive test pattern generation framework for efficient and reliable Hardware Trojan (HT) detection. HT is a backdoor attack that tampers with the design of victim integrated circuits (ICs). AdaTest improves the existing HT detection techniques in terms of scalability and accuracy of detecting smaller Trojans in the presence of noise and variations. To achieve high trigger coverage, AdaTest leverages Reinforcement Learning (RL) to produce a diverse set of test inputs. Particularly, we progressively generate test vectors with high reward values in an iterative manner. In each iteration, the test set is evaluated and adaptively expanded as needed. Furthermore, AdaTest integrates adaptive sampling to prioritize test samples that provide more information for HT detection, thus reducing the number of samples while improving the sample quality for faster exploration. We develop AdaTest with a Software/Hardware co-design principle and provide an optimized on-chip architecture solution. AdaTest's architecture minimizes the hardware overhead in two ways:(i) Deploying circuit emulation on programmable hardware to accelerate reward evaluation of the test input; (ii) Pipelining each computation stage in AdaTest by automatically constructing auxiliary circuit for test input generation, reward evaluation, and adaptive sampling. We evaluate AdaTest's performance on various HT benchmarks and compare it with two prior works that use logic testing for HT detection. Experimental results show that AdaTest engenders up to two orders of test generation speedup and two orders of test set size reduction compared to the prior works while achieving the same level or higher Trojan detection rate. △ Less

Submitted 12 April, 2022; originally announced April 2022.

An Adaptive Black-box Backdoor Detection Method for Deep Neural Networks

Authors: Xinqiao Zhang, Huili Chen, Ke Huang, Farinaz Koushanfar

Abstract: With the surge of Machine Learning (ML), An emerging amount of intelligent applications have been developed. Deep Neural Networks (DNNs) have demonstrated unprecedented performance across various fields such as medical diagnosis and autonomous driving. While DNNs are widely employed in security-sensitive fields, they are identified to be vulnerable to Neural Trojan (NT) attacks that are controlled… ▽ More With the surge of Machine Learning (ML), An emerging amount of intelligent applications have been developed. Deep Neural Networks (DNNs) have demonstrated unprecedented performance across various fields such as medical diagnosis and autonomous driving. While DNNs are widely employed in security-sensitive fields, they are identified to be vulnerable to Neural Trojan (NT) attacks that are controlled and activated by stealthy triggers. In this paper, we target to design a robust and adaptive Trojan detection scheme that inspects whether a pre-trained model has been Trojaned before its deployment. Prior works are oblivious of the intrinsic property of trigger distribution and try to reconstruct the trigger pattern using simple heuristics, i.e., stimulating the given model to incorrect outputs. As a result, their detection time and effectiveness are limited. We leverage the observation that the pixel trigger typically features spatial dependency and propose the first trigger approximation based black-box Trojan detection framework that enables a fast and scalable search of the trigger in the input space. Furthermore, our approach can also detect Trojans embedded in the feature space where certain filter transformations are used to activate the Trojan. We perform extensive experiments to investigate the performance of our approach across various datasets and ML models. Empirical results show that our approach achieves a ROC-AUC score of 0.93 on the public TrojAI dataset. Our code can be found at https://github.com/xinqiaozhang/adatrojan △ Less

Submitted 8 April, 2022; originally announced April 2022.

Comments: arXiv admin note: substantial text overlap with arXiv:2102.01815

Growth windows of epitaxial $\textrm{Nb}_x\textrm{N}$ films on c-plane sapphire and their structural and superconducting properties

Authors: John G. Wright, Huili G. Xing, Debdeep Jena

Abstract: NbN films are grown on c-plane sapphire substrates by molecular beam epitaxy. The structural and superconducting properties of the film are characterized to demonstrate that growth parameters such as substrate temperature and active nitrogen flux effect the structural phase of films, and thereby the superconducting critical temperature. Four phases of NbN are identified for films grown in differen… ▽ More NbN films are grown on c-plane sapphire substrates by molecular beam epitaxy. The structural and superconducting properties of the film are characterized to demonstrate that growth parameters such as substrate temperature and active nitrogen flux effect the structural phase of films, and thereby the superconducting critical temperature. Four phases of NbN are identified for films grown in different conditions. In a novel finding, we demonstrate that atomically flat and highly crystalline $β$-$\textrm{Nb}_2\textrm{N}$ films can be grown at substrate temperatures of 1100 \degree C or higher, and that the superconducting critical temperature of phase pure $β$-$\textrm{Nb}_2\textrm{N}$ films is $0.35~K<T_c<0.6~K$, based on measurements of films grown at different substrate temperatures. △ Less

Submitted 26 March, 2022; originally announced March 2022.

Backdoor Defense in Federated Learning Using Differential Testing and Outlier Detection

Authors: Yein Kim, Huili Chen, Farinaz Koushanfar

Abstract: The goal of federated learning (FL) is to train one global model by aggregating model parameters updated independently on edge devices without accessing users' private data. However, FL is susceptible to backdoor attacks where a small fraction of malicious agents inject a targeted misclassification behavior in the global model by uploading polluted model updates to the server. In this work, we pro… ▽ More The goal of federated learning (FL) is to train one global model by aggregating model parameters updated independently on edge devices without accessing users' private data. However, FL is susceptible to backdoor attacks where a small fraction of malicious agents inject a targeted misclassification behavior in the global model by uploading polluted model updates to the server. In this work, we propose DifFense, an automated defense framework to protect an FL system from backdoor attacks by leveraging differential testing and two-step MAD outlier detection, without requiring any previous knowledge of attack scenarios or direct access to local model parameters. We empirically show that our detection method prevents a various number of potential attackers while consistently achieving the convergence of the global model comparable to that trained under federated averaging (FedAvg). We further corroborate the effectiveness and generalizability of our method against prior defense techniques, such as Multi-Krum and coordinate-wise median aggregation. Our detection method reduces the average backdoor accuracy of the global model to below 4% and achieves a false negative rate of zero. △ Less

Submitted 21 February, 2022; originally announced February 2022.

Hyperspherical approach to atom--dimer collision with the Jacobi boundary condition

Authors: Cai-Yun Zhao, Yi Zhang, Hui-Li Han, Ting-Yun Shi

Abstract: In this study, we investigate atom--dimer scattering within the framework of the hyperspherical method. The coupled channel Schrödinger equation is solved using the R-matrix propagation technique combined with the smooth variable discretization method. In the matching procedure, the asymptotic wave functions are expressed in the rotated Jacobi coordinates. We apply this approach to the elastic sca… ▽ More In this study, we investigate atom--dimer scattering within the framework of the hyperspherical method. The coupled channel Schrödinger equation is solved using the R-matrix propagation technique combined with the smooth variable discretization method. In the matching procedure, the asymptotic wave functions are expressed in the rotated Jacobi coordinates. We apply this approach to the elastic scattering $^{3}$He(T$\uparrow$) + $^{4}$He$_{2}$ and H$\uparrow$ + H$\uparrow$Li processes. The convergence of the scattering length as a function of the propagation distance is studied. We find that the method is reliable and can provide considerable savings over previous propagators. △ Less

Submitted 25 September, 2022; v1 submitted 17 February, 2022; originally announced February 2022.

A polynomial-time approximation scheme for parallel two-stage flowshops under makespan constraint

Authors: Weitian Tong, Yao Xu, Huili Zhang

Abstract: As a hybrid of the Parallel Two-stage Flowshop problem and the Multiple Knapsack problem, we investigate the scheduling of parallel two-stage flowshops under makespan constraint, which was motivated by applications in cloud computing and introduced by Chen et al. [3] recently. A set of two-stage jobs are selected and scheduled on parallel two-stage flowshops to achieve the maximum total profit whi… ▽ More As a hybrid of the Parallel Two-stage Flowshop problem and the Multiple Knapsack problem, we investigate the scheduling of parallel two-stage flowshops under makespan constraint, which was motivated by applications in cloud computing and introduced by Chen et al. [3] recently. A set of two-stage jobs are selected and scheduled on parallel two-stage flowshops to achieve the maximum total profit while maintaining the given makespan constraint. We give a positive answer to an open question about its approximability proposed by Chen et al. [3]. More specifically, based on guessing strategies and rounding techniques for linear programs, we present a polynomial-time approximation scheme (PTAS) for the case when the number of flowshops is a fixed constant. △ Less

Submitted 18 May, 2022; v1 submitted 11 January, 2022; originally announced January 2022.

Comments: Theoretical Computer Science (2022)

Experimental unsupervised learning of non-Hermitian knotted phases with solid-state spins

Authors: Yefei Yu, Li-Wei Yu, Wengang Zhang, Huili Zhang, Xiaolong Ouyang, Yanqing Liu, Dong-Ling Deng, L. -M. Duan

Abstract: Non-Hermiticity has widespread applications in quantum physics. It brings about distinct topological phases without Hermitian counterparts, and gives rise to the fundamental challenge of phase classification from both theoretical and experimental aspects. Here we report the first experimental demonstration of unsupervised learning of non-Hermitian topological phases with the nitrogen-vacancy cente… ▽ More Non-Hermiticity has widespread applications in quantum physics. It brings about distinct topological phases without Hermitian counterparts, and gives rise to the fundamental challenge of phase classification from both theoretical and experimental aspects. Here we report the first experimental demonstration of unsupervised learning of non-Hermitian topological phases with the nitrogen-vacancy center platform. In particular, we implement the non-Hermitian twister model, which hosts peculiar knotted topological phases, with a solid-state quantum simulator consisting of an electron spin and a nearby $^{13}$C nuclear spin in a nitrogen-vacancy center in diamond. By tuning the microwave pulses, we efficiently generate a set of experimental data without phase labels. Furthermore, based on the diffusion map method, we cluster this set of experimental raw data into three different knotted phases in an unsupervised fashion without a priori knowledge of the system, which is in sharp contrast to the previously implemented supervised learning phases of matter. Our results showcase the intriguing potential for autonomous classification of exotic unknown topological phases with experimental raw data. △ Less

Submitted 27 December, 2021; originally announced December 2021.

Comments: 7+20 pages

The Explicit Formula and Parity for Some Generalized Euler Functions

Authors: Shichun Yang, Qunying Liao, Shan Du, Huili Wang

Abstract: Based on elementary methods and techniques, the explicit formula for the generalized Euler function $\varphi_{e}(n)(e=8,12)$ is given, and then a sufficient and necessary condition for $\varphi_{8}(n)$ or $\varphi_{12}(n)$ to be odd is obtained, respectively. Based on elementary methods and techniques, the explicit formula for the generalized Euler function $\varphi_{e}(n)(e=8,12)$ is given, and then a sufficient and necessary condition for $\varphi_{8}(n)$ or $\varphi_{12}(n)$ to be odd is obtained, respectively. △ Less

Submitted 22 December, 2021; v1 submitted 22 December, 2021; originally announced December 2021.

Comments: 19pages,2 figures

MSC Class: 11A25(Primary); 11D45(Secondary)