arXiv subjects

Tao Yang

Publications and source records attributed to Tao Yang.

At least 271 records · Page 15Linked to original sources

arXiv2022

Learning Disentangled Representation by Exploiting Pretrained Generative Models: A Contrastive Learning View

From the intuitive notion of disentanglement, the image variations corresponding to different factors should be distinct from each other, and the disentangled representation should reflect those variations with separate dimensions. To discover the factors and learn disentangled representation, previous methods typically leverage an extra regularization term when learning to generate realistic images. However, the term usually results in a trade-off between disentanglement and generation quality. For the generative models pretrained without any disentanglement term, the generated images show semantically meaningful variations when traversing along different directions in the latent space. Based on this observation, we argue that it is possible to mitigate the trade-off by $(i)$ leveraging the pretrained generative models with high generation quality, $(ii)$ focusing on discovering the traversal directions as factors for disentangled representation learning. To achieve this, we propose Disentaglement via Contrast (DisCo) as a framework to model the variations based on the target disentangled representations, and contrast the variations to jointly discover disentangled directions and learn disentangled representations. DisCo achieves the state-of-the-art disentangled representation learning and distinct direction discovering, given pretrained non-disentangled generative models including GAN, VAE, and Flow. Source code is at https://github.com/xrenaa/DisCo.

cs.CV↗

arXiv2022

Low Gain Avalanche Detectors with Good Time Resolution Developed by IHEP and IME for ATLAS HGTD project

This paper shows the simulation and test results of 50um thick Low Gain Avalanche Detectors (LGAD) sensors designed by the Institute of High Energy Physics (IHEP) and fabricated by the Institute of Microelectronics of the Chinese Academy of Sciences (IME). Three wafers have been produced with four different gain layer implant doses each. Different production processes, including variation in the n++ layer implant energy and carbon co-implantation were used. Test results show that the IHEP-IME sensors with the higher dose of gain layer have lower breakdown voltages and higher gain layer voltages from capacitance-voltage properties, which are consistent with the TCAD simulation. Beta test results show that the time resolution of IHEP-IME sensors is better than 35ps when operated at high voltage and the collected charges of IHEP-IME sensors are larger than 15fC before irradiation, which fulfill the required specifications of sensors before irradiations for the ATLAS HGTD project.

physics.ins-det↗

arXiv2022

Space-borne atom interferometric gravitational wave detections. Part II. Dark sirens and finding the one

In this paper, we investigate the potential of dark sirens by the space-borne atom interferometric gravitational-wave detectors to probe the Hubble constant. In the mid-frequency band, the sources live a long time. The motion of a detector around the Sun as well as in Earth orbit would induce large Doppler and reorientation effects, providing a precise angular resolution. Such precise localization for the GW sources makes it possible to observe the dark sirens with only one potential host galaxy, which are dubbed "golden dark sirens". We construct the catalogs of golden dark sirens and estimate that there are around 79 and 35 golden dark sirens of binary neutron stars (BNS) and binary black holes (BBH) that would be pass the detection threshold of AEDGE in 5 years. Our results show that with 5, 10, and all 79 golden dark BNS tracked by AEDGE one can constrain $H_0$ at 5.5\%, 4.1\%, and 1.8\% precision levels. With 5, 10, and all 35 golden dark BBH one can constrain $H_0$ at 2.2\%, 1.8\%, and 1.5\% precision levels, respectively. It suggests that only 5-10 golden dark BBH by AEDGE are sufficient to arbitrate the current tension between local and high-$z$ measurements of $H_0$.

gr-qc↗

arXiv2022

A Primal-Dual SGD Algorithm for Distributed Nonconvex Optimization

The distributed nonconvex optimization problem of minimizing a global cost function formed by a sum of $n$ local cost functions by using local information exchange is considered. This problem is an important component of many machine learning techniques with data parallelism, such as deep learning and federated learning. We propose a distributed primal--dual stochastic gradient descent (SGD) algorithm, suitable for arbitrarily connected communication networks and any smooth (possibly nonconvex) cost functions. We show that the proposed algorithm achieves the linear speedup convergence rate $\mathcal{O}(1/\sqrt{nT})$ for general nonconvex cost functions and the linear speedup convergence rate $\mathcal{O}(1/(nT))$ when the global cost function satisfies the Polyak--Łojasiewicz (P--Ł) condition, where $T$ is the total number of iterations. We also show that the output of the proposed algorithm with constant parameters linearly converges to a neighborhood of a global optimum. We demonstrate through numerical experiments the efficiency of our algorithm in comparison with the baseline centralized SGD and recently proposed distributed SGD algorithms.

math.OC↗

arXiv2022

Composite Re-Ranking for Efficient Document Search with BERT

Although considerable efforts have been devoted to transformer-based ranking models for document search, the relevance-efficiency tradeoff remains a critical problem for ad-hoc ranking. To overcome this challenge, this paper presents BECR (BERT-based Composite Re-Ranking), a composite re-ranking scheme that combines deep contextual token interactions and traditional lexical term-matching features. In particular, BECR exploits a token encoding mechanism to decompose the query representations into pre-computable uni-grams and skip-n-grams. By applying token encoding on top of a dual-encoder architecture, BECR separates the attentions between a query and a document while capturing the contextual semantics of a query. In contrast to previous approaches, this framework does not perform expensive BERT computations during online inference. Thus, it is significantly faster, yet still able to achieve high competitiveness in ad-hoc ranking relevance. Finally, an in-depth comparison between BECR and other start-of-the-art neural ranking baselines is described using the TREC datasets, thereby further demonstrating the enhanced relevance and efficiency of BECR.

cs.IR↗

arXiv2022

Vortex Interaction in Triple Flickering Buoyant Diffusion Flames

Triple flickering buoyant diffusion flames as a nonlinear dynamical system of coupled oscillators were computationally investigated. The four distinct dynamical modes (in-phase, death, rotation, and partial in-phase) observed in the previous candle-flame experiments were computationally reproduced for jet diffusion flames of methane. The four modes were interpreted from the perspective of vortex interaction and particularly of vorticity reconnection and vortex-induced flow. Specifically, the in-phase mode is caused by the periodic shedding of the trefoil vortex formed by the reconnection of three toroidal vortices; the death mode is due to the suppression of vortex shedding at small Reynolds numbers; the rotation mode appears as three toroidal vortices alternatively shed off with a constant phase difference; the partial in-phase model is caused by the vorticity reconnection of two toroidal vortices leaving another one shedding off in anti-phase.

physics.flu-dyn↗

arXiv2022

Dynamical Mode Recognition of Triple Flickering Buoyant Diffusion Flames: from Physical Space to Phase Space and to Wasserstein Space

Triple flickering buoyant diffusion flames in an isosceles triangle arrangement were experimentally studied as a nonlinear dynamical system of coupled oscillators. The objective of the study is two-fold: to establish a well-controlled gas-fuel diffusion flame experiment that can remedy the deficiencies of previous candle-flame experiments, and to develop an objective methodology for dynamical mode recognition that is based on the Wasserstein distance between distribution functions of phase points and can be readily generalized to larger dynamical system consisting of more than three flames. By using the present experiment and methodology, we recognized seven distinct stable dynamical modes, such as in-phase mode, flickering death mode, partial flickering death mode, partial in-phase mode, rotation mode, partial decoupled mode, and decoupled mode. These results generalize the literature results for triple flickering flame system in a straight-line and equal-lateral triangle arrangement. The identification or discrimination of two dynamical modes can be quantified as the small or large Wasserstein distance, respectively.

physics.flu-dyn↗

arXiv2021

Yetter-Drinfeld modules for group-cograded Hopf quasigroups

Let $H$ be a crossed group-cograded Hopf quasigroup. We first introduce the notion of $p$-Yetter-Drinfeld quasimodule over $H$. If the antipode of $H$ is bijective, we show that the category $\mathscr Y\mathscr D\mathscr Q(H)$ of Yetter-Drinfeld quasimodules over $H$ is a crossed category, and the subcategory $\mathscr Y\mathscr D(H)$ of Yetter-Drinfeld modules is a braided crossed category.

math.RA↗

arXiv2021

New source for tuning the effective Rabi frequency discovered in multiphoton ionization

The Autler-Townes effect due to near resonance transition between 4s-4p states in potassium atoms is mapped out in the photo-electron-momentum distribution and manifests itself as a splitting in the photo-electron kinetic energy spectra. The energy splitting fits well with the calculated Rabi frequency at low laser intensities and shows clear deviation at laser intensities above 1.5x10^11 W/cm^2. An effective Rabi frequency formulae including the ionization process explains the observed results. Our results reveal the possibility to tune the effective coupling strength with the cost of the number of level-populations.

physics.atom-ph↗

arXiv2021

Zeroth-Order Algorithms for Stochastic Distributed Nonconvex Optimization

In this paper, we consider a stochastic distributed nonconvex optimization problem with the cost function being distributed over $n$ agents having access only to zeroth-order (ZO) information of the cost. This problem has various machine learning applications. As a solution, we propose two distributed ZO algorithms, in which at each iteration each agent samples the local stochastic ZO oracle at two points with a time-varying smoothing parameter. We show that the proposed algorithms achieve the linear speedup convergence rate $\mathcal{O}(\sqrt{p/(nT)})$ for smooth cost functions under the state-dependent variance assumptions which are more general than the commonly used bounded variance and Lipschitz assumptions, and $\mathcal{O}(p/(nT))$ convergence rate when the global cost function additionally satisfies the Polyak--Łojasiewicz (P--Ł) condition in addition, where $p$ and $T$ are the dimension of the decision variable and the total number of iterations, respectively. To the best of our knowledge, this is the first linear speedup result for distributed ZO algorithms, which enables systematic processing performance improvements by adding more agents. We also show that the proposed algorithms converge linearly under the relative bounded second moment assumptions and the P--Ł condition. We demonstrate through numerical experiments the efficiency of our algorithms on generating adversarial examples from deep neural networks in comparison with baseline and recently proposed centralized and distributed ZO algorithms.

math.OC↗

arXiv2021

Space-borne atom interferometric gravitational wave detections. Part I. The forecast of bright sirens on cosmology

Atom interferometers (AIs) as gravitational-wave (GW) detectors have been proposed a decade ago. Both ground and space-based projects will be in construction and preparation in the near future. In this paper, for the first time, we investigate the potential of the space-borne AIs on detecting GW standard sirens and hence the applications on cosmology. We consider AEDGE as our fiducial AI GW detector and estimate the number of bright sirens that would be obtained within a 5-years data-taking period of GW and with the follow-up observation of electromagnetic (EM) counterparts. We then construct the mock catalogue of bright sirens and predict their ability on constraining cosmological parameters such as the Hubble constant, dynamics of dark energy, and modified gravity theory. Our preliminary results show around order $\mathcal{O} (30)$ bright sirens can be obtained from a 5-years operation time of AEDGE and the follow-up observation of EM counterparts. The bright sirens alone can measure $H_0$ with a precision of 2.1\%, which is sufficient to arbitrate the Hubble tension. Combining current most precise electromagnetic experiments, the inclusion of AEDGE bright sirens can improve the measurement of the equation of state of dark energy, though marginally. Moreover, by modifying GW propagation on cosmological scales, the deviations from general relativity (modified gravity theory effects) can be constrained at 5.7\% precision level.

gr-qc↗

arXiv2021

Surface energy of the one-dimensional supersymmetric $t-J$ model with general integrable boundary terms in the antiferromagnetic sector

In this paper, we study the surface energy of the one-dimensional supersymmetric $t-J$ model with unparallel boundary magnetic fields, which is a typical $U(1)$-symmetry broken quantum integrable strongly correlated electron system. It is shown that at the ground state, the contribution of inhomogeneous term in the Bethe ansatz solution of eigenvalues of transfer matrix satisfies the finite size scaling law $L^β$ where $β<0$. Based on it, the physical quantities of the system in the thermodynamic limit are calculated. We obtain the patterns of Bethe roots and the analytical expressions of density of states, ground state energy and surface energy. We also find that there exist the stable boundary bound states if the boundary fields satisfy some constraints.

math-ph↗

arXiv2021

The Gravitational-Wave Physics II: Progress

It has been a half-decade since the first direct detection of gravitational waves, which signifies the coming of the era of the gravitational-wave astronomy and gravitational-wave cosmology. The increasing number of the detected gravitational-wave events has revealed the promising capability of constraining various aspects of cosmology, astronomy, and gravity. Due to the limited space in this review article, we will briefly summarize the recent progress over the past five years, but with a special focus on some of our own work for the Key Project ``Physics associated with the gravitational waves'' supported by the National Natural Science Foundation of China. In particular, (1) we have presented the mechanism of the gravitational-wave production during some physical processes of the early Universe, such as inflation, preheating and phase transition, and the cosmological implications of gravitational-wave measurements; (2) we have put constraints on the neutron star maximum mass according to GW170817 observations; (3) we have developed a numerical relativity algorithm based on the finite element method and a waveform model for the binary black hole coalescence along an eccentric orbit.

gr-qc↗

arXiv2021

On the tensor product of two oriented quantum algebras

In this paper, we give the oriented quantum algebra (abbr. OQA) structures on the tensor product of two different OQAs by using Chen's weak $\mathfrak{R}$-matrix in [J. Algebra 204(1998):504-531]. As a special case, the OQA structures on the tensor product of an OQA with itself are provided, which are different from Radford's results in [J. Knot Theory Ramifications 16(2007):929-957].

math.RA↗

arXiv2021

DeceFL: A Principled Decentralized Federated Learning Framework

Traditional machine learning relies on a centralized data pipeline, i.e., data are provided to a central server for model training. In many applications, however, data are inherently fragmented. Such a decentralized nature of these databases presents the biggest challenge for collaboration: sending all decentralized datasets to a central server raises serious privacy concerns. Although there has been a joint effort in tackling such a critical issue by proposing privacy-preserving machine learning frameworks, such as federated learning, most state-of-the-art frameworks are built still in a centralized way, in which a central client is needed for collecting and distributing model information (instead of data itself) from every other client, leading to high communication pressure and high vulnerability when there exists a failure at or attack on the central client. Here we propose a principled decentralized federated learning algorithm (DeceFL), which does not require a central client and relies only on local information transmission between clients and their neighbors, representing a fully decentralized learning framework. It has been further proven that every client reaches the global minimum with zero performance gap and achieves the same convergence rate $O(1/T)$ (where $T$ is the number of iterations in gradient descent) as centralized federated learning when the loss function is smooth and strongly convex. Finally, the proposed algorithm has been applied to a number of applications to illustrate its effectiveness for both convex and nonconvex loss functions, demonstrating its applicability to a wide range of real-world medical and industrial applications.

cs.LG↗

arXiv2021

Anomaly-Injected Deep Support Vector Data Description for Text Outlier Detection

Anomaly detection or outlier detection is a common task in various domains, which has attracted significant research efforts in recent years. Existing works mainly focus on structured data such as numerical or categorical data; however, anomaly detection on unstructured textual data is less attended. In this work, we target the textual anomaly detection problem and propose a deep anomaly-injected support vector data description (AI-SVDD) framework. AI-SVDD not only learns a more compact representation of the data hypersphere but also adopts a small number of known anomalies to increase the discriminative power. To tackle text input, we employ a multilayer perceptron (MLP) network in conjunction with BERT to obtain enriched text representations. We conduct experiments on three text anomaly detection applications with multiple datasets. Experimental results show that the proposed AI-SVDD is promising and outperforms existing works.

cs.CL↗

arXiv2021

Effects of shallow carbon and deep N++ layer on the radiation hardness of IHEP-IME LGAD sensors

Low Gain Avalanche Diode (LGAD) is applied for the High-Granularity Timing Detector (HGTD), and it will be used to upgrade the ATLAS experiment. The first batch IHEP-IME LGAD sensors were designed by the Institute of High Energy Physics (IHEP) and fabricated by the Institute of Microelectronics (IME). Three IHEP-IME sensors (W1, W7 and W8) were irradiated by the neutrons up to the fluence of 2.5 x 10^15 n_eq/cm^2 to study the effect of the shallow carbon and deep N++ layer on the irradiation hardness. Taking W7 as a reference, W1 has an extra shallow carbon applied, and W8 has a deeper N++ layer.

physics.ins-det↗

arXiv2021

Multiplex Graph Neural Network for Extractive Text Summarization

Extractive text summarization aims at extracting the most representative sentences from a given document as its summary. To extract a good summary from a long text document, sentence embedding plays an important role. Recent studies have leveraged graph neural networks to capture the inter-sentential relationship (e.g., the discourse graph) to learn contextual sentence embedding. However, those approaches neither consider multiple types of inter-sentential relationships (e.g., semantic similarity & natural connection), nor model intra-sentential relationships (e.g, semantic & syntactic relationship among words). To address these problems, we propose a novel Multiplex Graph Convolutional Network (Multi-GCN) to jointly model different types of relationships among sentences and words. Based on Multi-GCN, we propose a Multiplex Graph Summarization (Multi-GraS) model for extractive text summarization. Finally, we evaluate the proposed models on the CNN/DailyMail benchmark dataset to demonstrate the effectiveness of our method.

cs.CL↗