-
Classification of Solvable Lie algebras whose non-trivial Coadjoint Orbits of simply connected Lie groups are all of Codimension 2
Authors:
Hieu Van Ha,
Vu Anh Le,
Tu Thi Cam Nguyen,
Hoa Duong Quang
Abstract:
We give a classification of real solvable Lie algebras whose non-trivial coadjoint orbits of corresponding simply connected Lie groups are all of codimension 2. These Lie algebras belong to a well-known class, called the class of MD-algebras.
We give a classification of real solvable Lie algebras whose non-trivial coadjoint orbits of corresponding simply connected Lie groups are all of codimension 2. These Lie algebras belong to a well-known class, called the class of MD-algebras.
△ Less
Submitted 29 July, 2022;
originally announced July 2022.
-
Improving Domain Generalization by Learning without Forgetting: Application in Retail Checkout
Authors:
Thuy C. Nguyen,
Nam LH. Phan,
Son T. Nguyen
Abstract:
Designing an automatic checkout system for retail stores at the human level accuracy is challenging due to similar appearance products and their various poses. This paper addresses the problem by proposing a method with a two-stage pipeline. The first stage detects class-agnostic items, and the second one is dedicated to classify product categories. We also track the objects across video frames to…
▽ More
Designing an automatic checkout system for retail stores at the human level accuracy is challenging due to similar appearance products and their various poses. This paper addresses the problem by proposing a method with a two-stage pipeline. The first stage detects class-agnostic items, and the second one is dedicated to classify product categories. We also track the objects across video frames to avoid duplicated counting. One major challenge is the domain gap because the models are trained on synthetic data but tested on the real images. To reduce the error gap, we adopt domain generalization methods for the first-stage detector. In addition, model ensemble is used to enhance the robustness of the 2nd-stage classifier. The method is evaluated on the AI City challenge 2022 -- Track 4 and gets the F1 score $40\%$ on the test A set. Code is released at the link https://github.com/cybercore-co-ltd/aicity22-track4.
△ Less
Submitted 12 July, 2022;
originally announced July 2022.
-
Improving Object Detection by Label Assignment Distillation
Authors:
Chuong H. Nguyen,
Thuy C. Nguyen,
Tuan N. Tang,
Nam L. H. Phan
Abstract:
Label assignment in object detection aims to assign targets, foreground or background, to sampled regions in an image. Unlike labeling for image classification, this problem is not well defined due to the object's bounding box. In this paper, we investigate the problem from a perspective of distillation, hence we call Label Assignment Distillation (LAD). Our initial motivation is very simple, we u…
▽ More
Label assignment in object detection aims to assign targets, foreground or background, to sampled regions in an image. Unlike labeling for image classification, this problem is not well defined due to the object's bounding box. In this paper, we investigate the problem from a perspective of distillation, hence we call Label Assignment Distillation (LAD). Our initial motivation is very simple, we use a teacher network to generate labels for the student. This can be achieved in two ways: either using the teacher's prediction as the direct targets (soft label), or through the hard labels dynamically assigned by the teacher (LAD). Our experiments reveal that: (i) LAD is more effective than soft-label, but they are complementary. (ii) Using LAD, a smaller teacher can also improve a larger student significantly, while soft-label can't. We then introduce Co-learning LAD, in which two networks simultaneously learn from scratch and the role of teacher and student are dynamically interchanged. Using PAA-ResNet50 as a teacher, our LAD techniques can improve detectors PAA-ResNet101 and PAA-ResNeXt101 to $46 \rm AP$ and $47.5\rm AP$ on the COCO test-dev set. With a stronger teacher PAA-SwinB, we improve the students PAA-ResNet50 to $43.7\rm AP$ by only 1x schedule training and standard setting, and PAA-ResNet101 to $47.9\rm AP$, significantly surpassing the current methods. Our source code and checkpoints are released at https://git.io/JrDZo.
△ Less
Submitted 19 October, 2021; v1 submitted 24 August, 2021;
originally announced August 2021.
-
Classification of 7-dimensional solvable Lie algebras having 5-dimensional nilradicals
Authors:
Vu A. Le,
Tuan A. Nguyen,
Tu T. C. Nguyen,
Tuyen T. M. Nguyen,
Thieu N. Vo
Abstract:
This paper presents a classification of 7-dimensional real and complex indecomposable solvable Lie algebras having some 5-dimensional nilradicals. Afterwards, we combine our results with those of Rubin and Winternitz (1993), Ndogmo and Winternitz (1994), Snobl and Winternitz (2005, 2009), Snobl and Karásek (2010) to obtain a complete classification of 7-dimensional real and complex indecomposable…
▽ More
This paper presents a classification of 7-dimensional real and complex indecomposable solvable Lie algebras having some 5-dimensional nilradicals. Afterwards, we combine our results with those of Rubin and Winternitz (1993), Ndogmo and Winternitz (1994), Snobl and Winternitz (2005, 2009), Snobl and Karásek (2010) to obtain a complete classification of 7-dimensional real and complex indecomposable solvable Lie algebras with 5-dimensional nilradicals. In association with Gong (1998), Parry (2007), Hindeleh and Thompson (2008), we achieve a classification of 7-dimensional real and complex indecomposable solvable Lie algebras.
△ Less
Submitted 8 July, 2021;
originally announced July 2021.
-
1st Place Solution for YouTubeVOS Challenge 2021:Video Instance Segmentation
Authors:
Thuy C. Nguyen,
Tuan N. Tang,
Nam LH. Phan,
Chuong H. Nguyen,
Masayuki Yamazaki,
Masao Yamanaka
Abstract:
Video Instance Segmentation (VIS) is a multi-task problem performing detection, segmentation, and tracking simultaneously. Extended from image set applications, video data additionally induces the temporal information, which, if handled appropriately, is very useful to identify and predict object motions. In this work, we design a unified model to mutually learn these tasks. Specifically, we propo…
▽ More
Video Instance Segmentation (VIS) is a multi-task problem performing detection, segmentation, and tracking simultaneously. Extended from image set applications, video data additionally induces the temporal information, which, if handled appropriately, is very useful to identify and predict object motions. In this work, we design a unified model to mutually learn these tasks. Specifically, we propose two modules, named Temporally Correlated Instance Segmentation (TCIS) and Bidirectional Tracking (BiTrack), to take the benefit of the temporal correlation between the object's instance masks across adjacent frames. On the other hand, video data is often redundant due to the frame's overlap. Our analysis shows that this problem is particularly severe for the YoutubeVOS-VIS2021 data. Therefore, we propose a Multi-Source Data (MSD) training mechanism to compensate for the data deficiency. By combining these techniques with a bag of tricks, the network performance is significantly boosted compared to the baseline, and outperforms other methods by a considerable margin on the YoutubeVOS-VIS 2019 and 2021 datasets.
△ Less
Submitted 8 July, 2021; v1 submitted 11 June, 2021;
originally announced June 2021.
-
Representation of Real Solvable Lie Algebras Having 2-dimensional Derived Ideal and Geometry of Coadjoint Orbits of Corresponding Lie Groups
Authors:
Tu T. C Nguyen,
Vu A. Le
Abstract:
Let {\Lnk} be the class of all $n$-dimensional real solvable Lie algebras having $k$-dimensional derived ideals. In 2020 the authors et al. gave a classification of all non 2-step nilpotent Lie algebras of {\Li}. We propose in this paper to study representations of these Lie algebras as well as their corresponding connected and simply connected Lie groups. That is, for each algebra, we give an upp…
▽ More
Let {\Lnk} be the class of all $n$-dimensional real solvable Lie algebras having $k$-dimensional derived ideals. In 2020 the authors et al. gave a classification of all non 2-step nilpotent Lie algebras of {\Li}. We propose in this paper to study representations of these Lie algebras as well as their corresponding connected and simply connected Lie groups. That is, for each algebra, we give an upper bound of the minimal degree of a faithful representation. Then, we give a geometrical description of coadjoint orbits of corresponding groups. Moreover, we show that the characteristic property of the family of maximal dimensional coadjoint orbits of a MD-group studied by K. P. Shum and the second author et al. is still true for the Lie groups considered here. Namely, we prove that, for each considered group, the family of the maximal dimensional coadjoint orbits forms a measurable foliation in the sense of Connes. The topological classification of these foliations is also provided.
△ Less
Submitted 25 September, 2021; v1 submitted 2 May, 2021;
originally announced May 2021.
-
Single Stage Class Agnostic Common Object Detection: A Simple Baseline
Authors:
Chuong H. Nguyen,
Thuy C. Nguyen,
Anh H. Vo,
Yamazaki Masayuki
Abstract:
This paper addresses the problem of common object detection, which aims to detect objects of similar categories from a set of images. Although it shares some similarities with the standard object detection and co-segmentation, common object detection, recently promoted by \cite{Jiang2019a}, has some unique advantages and challenges. First, it is designed to work on both closed-set and open-set con…
▽ More
This paper addresses the problem of common object detection, which aims to detect objects of similar categories from a set of images. Although it shares some similarities with the standard object detection and co-segmentation, common object detection, recently promoted by \cite{Jiang2019a}, has some unique advantages and challenges. First, it is designed to work on both closed-set and open-set conditions, a.k.a. known and unknown objects. Second, it must be able to match objects of the same category but not restricted to the same instance, texture, or posture. Third, it can distinguish multiple objects. In this work, we introduce the Single Stage Common Object Detection (SSCOD) to detect class-agnostic common objects from an image set. The proposed method is built upon the standard single-stage object detector. Furthermore, an embedded branch is introduced to generate the object's representation feature, and their similarity is measured by cosine distance. Experiments are conducted on PASCAL VOC 2007 and COCO 2014 datasets. While being simple and flexible, our proposed SSCOD built upon ATSSNet performs significantly better than the baseline of the standard object detection, while still be able to match objects of unknown categories. Our source code can be found at \href{https://github.com/cybercore-co-ltd/Single-Stage-Common-Object-Detection}{(URL)}
△ Less
Submitted 25 April, 2021;
originally announced April 2021.
-
NLPBK at VLSP-2020 shared task: Compose transformer pretrained models for Reliable Intelligence Identification on Social network
Authors:
Thanh Chinh Nguyen,
Van Nha Nguyen
Abstract:
This paper describes our method for tuning a transformer-based pretrained model, to adaptation with Reliable Intelligence Identification on Vietnamese SNSs problem. We also proposed a model that combines bert-base pretrained models with some metadata features, such as the number of comments, number of likes, images of SNS documents,... to improved results for VLSP shared task: Reliable Intelligenc…
▽ More
This paper describes our method for tuning a transformer-based pretrained model, to adaptation with Reliable Intelligence Identification on Vietnamese SNSs problem. We also proposed a model that combines bert-base pretrained models with some metadata features, such as the number of comments, number of likes, images of SNS documents,... to improved results for VLSP shared task: Reliable Intelligence Identification on Vietnamese SNSs. With appropriate training techniques, our model is able to achieve 0.9392 ROC-AUC on public test set and the final version settles at top 2 ROC-AUC (0.9513) on private test set.
△ Less
Submitted 29 January, 2021;
originally announced January 2021.
-
On the classifying problem for the class of real solvable Lie algebras having 2-dimensional or 2-codimensional derived ideal
Authors:
Vu A. Le,
Tuan A. Nguyen,
Tu T. C. Nguyen,
Tuyen T. M. Nguyen,
Hoa Q. Duong
Abstract:
Let $\mathrm{Lie} \left(n, k\right)$ denote the class of all $n$-dimensional real solvable Lie algebras having $k$-dimensional derived ideal ($1 \leqslant k \leqslant n-1$). In 1993, the class $\mathrm{Lie} \left(n, 1\right)$ was completely classified by Schöbel \cite{Sch93}. In 2016, Vu A. Le et al. \cite{VHTHT16} considered the class $\mathrm{Lie} \left(n, n-1\right)$ and classified its subclass…
▽ More
Let $\mathrm{Lie} \left(n, k\right)$ denote the class of all $n$-dimensional real solvable Lie algebras having $k$-dimensional derived ideal ($1 \leqslant k \leqslant n-1$). In 1993, the class $\mathrm{Lie} \left(n, 1\right)$ was completely classified by Schöbel \cite{Sch93}. In 2016, Vu A. Le et al. \cite{VHTHT16} considered the class $\mathrm{Lie} \left(n, n-1\right)$ and classified its subclass containing all the algebras having 1-codimensional commutative derived ideal. One subclass in {\Li} was firstly considered and incompletely classified by Schöbel \cite{Sch93} in 1993. Later, Janisse also gave an incomplete classification of {\Li} and published as a scientific report \cite{Jan10} in 2010. In this paper, we set up a new approach to study the classifying problem of classes {\Li} as well as {\li} and present the new complete classification of {\Li} in the combination with the well-known Eberlein's result of 2-step nilpotent Lie algebras from \cite[p.\,37--72]{Ebe03}. The paper will also classify a subclass of {\li} and will point out missings in Schöbel \cite{Sch93}, Janisse \cite{Jan10}, Mubarakzyanov \cite{Mub63a} as well as revise an error of Morozov \cite{Mor58}.
△ Less
Submitted 20 July, 2018; v1 submitted 26 June, 2018;
originally announced June 2018.
-
Algebraic EHP sequences
Authors:
The Cuong Nguyen
Abstract:
The James fibrations give rise to the geometric EHP sequences of homotopy groups of spheres. Using techniques from the Lambda algebra, \cite{BCKQRS66} shows that there are similar long exact sequences of Ext groups defining the $E_{2}-$page of the Bousfield-Kan spectral sequence (also known as the unstable Adams spectral sequence) computing homotopy groups of spheres. In this paper, we give anothe…
▽ More
The James fibrations give rise to the geometric EHP sequences of homotopy groups of spheres. Using techniques from the Lambda algebra, \cite{BCKQRS66} shows that there are similar long exact sequences of Ext groups defining the $E_{2}-$page of the Bousfield-Kan spectral sequence (also known as the unstable Adams spectral sequence) computing homotopy groups of spheres. In this paper, we give another proof for this phenomenon.
△ Less
Submitted 31 December, 2015;
originally announced December 2015.
-
Polynomial approximations of a class of stochastic multiscale elasticity problems
Authors:
Viet Ha Hoang,
Thanh Chung Nguyen,
Bingxing Xia
Abstract:
We consider a class of elasticity equations in ${\mathbb R}^d$ whose elastic moduli depend on $n$ separated microscopic scales, are random and expressed as a linear expansion of a countable sequence of random variables which are independently and identically uniformly distributed in a compact interval. The multiscale Hellinger-Reissner problem that allows for computing the stress directly, and the…
▽ More
We consider a class of elasticity equations in ${\mathbb R}^d$ whose elastic moduli depend on $n$ separated microscopic scales, are random and expressed as a linear expansion of a countable sequence of random variables which are independently and identically uniformly distributed in a compact interval. The multiscale Hellinger-Reissner problem that allows for computing the stress directly, and the multiscale mixed problem for nearly incompressible isotropic materials are considered. The stochastic problems are studied via deterministic problems that depend on a countable number of real parameters. We study the multiscale homogenized problems that contain all the macroscopic and microscopic information, whose solutions are written as generalized polynomial chaos (gpc) expansions. We approximate these solutions by semidiscrete Galerkin approximating problems that project into the spaces of functions with only a finite number of $N$ gpc modes. We deduce bounds and summability properties for the solutions' gpc expansion coefficients, which imply explicit rates of convergence in terms of $N$ when the gpc modes used for the Galerkin approximation are chosen to correspond to the best $N$ terms in the gpc expansion. For the mixed problem for nearly incompressible materials, the rate of convergence for the best $N$ term approximation is independent of the Lamé constants' ratio. We establish parametric correctors in terms of the semidiscrete Galerkin approximations. For two scale problems, an explicit homogenization rate is deduced. Together with the best $N$ term rate, it provides an explicit convergence rate for the correctors of the parametric multiscale problems. For nearly incompressible materials, we obtain a homogenization rate that is independent of the ratio of the Lamé constants, so that the error for the corrector is also independent of this ratio.
△ Less
Submitted 28 November, 2015;
originally announced November 2015.
-
Sur la torsion de Frobenius de la catégorie des modules instables
Authors:
The Cuong Nguyen
Abstract:
In the category $\mathcal{P}_{d}$ of strict polynomial functors, the morphisms between extension groups induced by the Frobenius twist are injective. In \cite{Cuo14a}, the category $\mathcal{P}_{d}$ is proved to be a full sub-category of the category $\mathcal{U}$ of unstable modules \textit{via} Hai's functor. The Frobenius twist is extended to the category $\mathcal{U}$ but remains mysterious th…
▽ More
In the category $\mathcal{P}_{d}$ of strict polynomial functors, the morphisms between extension groups induced by the Frobenius twist are injective. In \cite{Cuo14a}, the category $\mathcal{P}_{d}$ is proved to be a full sub-category of the category $\mathcal{U}$ of unstable modules \textit{via} Hai's functor. The Frobenius twist is extended to the category $\mathcal{U}$ but remains mysterious there. This article aims to study the Frobenius twist and its effects on the extension groups of unstable modules. We compute explicitly several extension groups and show that in these cases, the morphisms induced by the Frobenius twist are injective. These results are obtained by constructing the minimal injective resolution of the free unstable module $F(1)$.
△ Less
Submitted 4 December, 2015; v1 submitted 12 June, 2015;
originally announced June 2015.
-
Solutions to the Einstein-scalar field constraint equations with a small TT-tensor
Authors:
Romain Gicquaud,
The Cang Nguyen
Abstract:
In this paper, we prove a far-from-CMC result similar to the ones obtained by Holst, Nagy, Tsogtgerel and Maxwell for the conformal Einstein-scalar field constraint equations on compact Riemannian manifolds with positive (modified) Yamabe invariant.
In this paper, we prove a far-from-CMC result similar to the ones obtained by Holst, Nagy, Tsogtgerel and Maxwell for the conformal Einstein-scalar field constraint equations on compact Riemannian manifolds with positive (modified) Yamabe invariant.
△ Less
Submitted 18 February, 2015;
originally announced February 2015.
-
Gamma-spectrometric determination of 232U in uranium-bearing materials
Authors:
Jozsef Zsigrai,
Tam Cong Nguyen,
Andrey Berlizov
Abstract:
The 232U content of various uranium-bearing items was measured using low-background gamma spectrometry. The method is independent of the measurement geometry, sample form and chemical composition. Since 232U is an artificially produced isotope, it carries information about previous irradiation of the material, which is relevant for nuclear forensics, nuclear safeguards and for nuclear reactor oper…
▽ More
The 232U content of various uranium-bearing items was measured using low-background gamma spectrometry. The method is independent of the measurement geometry, sample form and chemical composition. Since 232U is an artificially produced isotope, it carries information about previous irradiation of the material, which is relevant for nuclear forensics, nuclear safeguards and for nuclear reactor operations. A correlation between the 232U content and 235U enrichment of the investigated samples has been established, which is consistent with theoretical predictions. It is also shown how the correlation of the mass ratio 232U/235U vs. 235U content can be used to distinguish materials contaminated with reprocessed uranium from materials made of reprocessed uranium.
△ Less
Submitted 7 August, 2015; v1 submitted 19 January, 2015;
originally announced January 2015.
-
SPATA: A Seeding and Patching Algorithm for Hybrid Transcriptome Assembly
Authors:
Tin Chi Nguyen,
Zhiyu Zhao,
Dongxiao Zhu
Abstract:
Transcriptome assembly from RNA-Seq reads is an active area of bioinformatics research. The ever-declining cost and the increasing depth of RNA-Seq have provided unprecedented opportunities to better identify expressed transcripts. However, the nonlinear transcript structures and the ultra-high throughput of RNA-Seq reads pose significant algorithmic and computational challenges to the existing tr…
▽ More
Transcriptome assembly from RNA-Seq reads is an active area of bioinformatics research. The ever-declining cost and the increasing depth of RNA-Seq have provided unprecedented opportunities to better identify expressed transcripts. However, the nonlinear transcript structures and the ultra-high throughput of RNA-Seq reads pose significant algorithmic and computational challenges to the existing transcriptome assembly approaches, either reference-guided or de novo. While reference-guided approaches offer good sensitivity, they rely on alignment results of the splice-aware aligners and are thus unsuitable for species with incomplete reference genomes. In contrast, de novo approaches do not depend on the reference genome but face a computational daunting task derived from the complexity of the graph built for the whole transcriptome. In response to these challenges, we present a hybrid approach to exploit an incomplete reference genome without relying on splice-aware aligners. We have designed a split-and-align procedure to efficiently localize the reads to individual genomic loci, which is followed by an accurate de novo assembly to assemble reads falling into each locus. Using extensive simulation data, we demonstrate a high accuracy and precision in transcriptome reconstruction by comparing to selected transcriptome assembly tools. Our method is implemented in assemblySAM, a GUI software freely available at http://sammate.sourceforge.net.
△ Less
Submitted 6 June, 2013;
originally announced June 2013.
-
CMOS-Compatible Nanowire Biosensors
Authors:
Thanh C. Nguyen,
Wanzhi Qiu,
Matteo Altissimo,
Paul G. Spizzirri,
Laurens H. Willems van Beveren,
Efstratios Skafidas
Abstract:
In this chapter, silicon nanowires that are compatible with CMOS fabrication processes have been described. It has been shown that these nanowires can be functionalized by conjugating monoclonal antibodies to their surface in order to build sensitive biochemical sensors. It has also been shown that by using frequency-based signals, all the necessary components to interrogate these nanowires can be…
▽ More
In this chapter, silicon nanowires that are compatible with CMOS fabrication processes have been described. It has been shown that these nanowires can be functionalized by conjugating monoclonal antibodies to their surface in order to build sensitive biochemical sensors. It has also been shown that by using frequency-based signals, all the necessary components to interrogate these nanowires can be built on low-cost CMOS processes.
△ Less
Submitted 21 February, 2013;
originally announced February 2013.
-
SASeq: A Selective and Adaptive Shrinkage Approach to Detect and Quantify Active Transcripts using RNA-Seq
Authors:
Tin Chi Nguyen,
Nan Deng,
Dongxiao Zhu
Abstract:
Identification and quantification of condition-specific transcripts using RNA-Seq is vital in transcriptomics research. While initial efforts using mathematical or statistical modeling of read counts or per-base exonic signal have been successful, they may suffer from model overfitting since not all the reference transcripts in a database are expressed under a specific biological condition. Standa…
▽ More
Identification and quantification of condition-specific transcripts using RNA-Seq is vital in transcriptomics research. While initial efforts using mathematical or statistical modeling of read counts or per-base exonic signal have been successful, they may suffer from model overfitting since not all the reference transcripts in a database are expressed under a specific biological condition. Standard shrinkage approaches, such as Lasso, shrink all the transcript abundances to zero in a non-discriminative manner. Thus it does not necessarily yield the set of condition-specific transcripts. Informed shrinkage approaches, using the observed exonic coverage signal, are thus desirable. Motivated by ubiquitous uncovered exonic regions in RNA-Seq data, termed as "naked exons", we propose a new computational approach that first filters out the reference transcripts not supported by splicing and paired-end reads, then followed by fitting a new mathematical model of per-base exonic coverage signal and the underlying transcripts structure. We introduce a tuning parameter to penalize the specific regions of the selected transcripts that were not supported by the naked exons. Our approach compares favorably with the selected competing methods in terms of both time complexity and accuracy using simulated and real-world data. Our method is implemented in SAMMate, a GUI software suite freely available from http://sammate.sourceforge.net
△ Less
Submitted 22 February, 2013; v1 submitted 17 August, 2012;
originally announced August 2012.
-
Dynamics and Thermodynamics of the Low-Temperature Strongly Interacting Bose Gas
Authors:
Nir Navon,
Swann Piatecki,
Kenneth John Günter,
Benno Rem,
Trong Canh Nguyen,
Frédéric Chevy,
Werner Krauth,
C. Salomon
Abstract:
We measure the zero-temperature equation of state of a homogeneous Bose gas of $^7$Li atoms by analyzing the \emph{in-situ} density distributions of trapped samples. For increasing repulsive interactions our data shows a clear departure from mean-field theory and provides a quantitative test of the many-body corrections first predicted in 1957 by Lee, Huang and Yang. We further probe the dynamic r…
▽ More
We measure the zero-temperature equation of state of a homogeneous Bose gas of $^7$Li atoms by analyzing the \emph{in-situ} density distributions of trapped samples. For increasing repulsive interactions our data shows a clear departure from mean-field theory and provides a quantitative test of the many-body corrections first predicted in 1957 by Lee, Huang and Yang. We further probe the dynamic response of the Bose gas to a varying interaction strength and compare it to simple theoretical models. We deduce a lower bound for the value of the universal constant $ξ>0.44(8)$ that would characterize the universal Bose gas at the unitary limit.
△ Less
Submitted 23 March, 2011;
originally announced March 2011.
