arXiv subjects

Yan Zhang

Publications and source records attributed to Yan Zhang.

At least 505 records · Page 28Linked to original sources

arXiv2021

Converged Reconfigurable Intelligent Surface and Mobile Edge Computing for Space Information Networks

Space information networks (SIN) are facing an ever-increasing thirst for high-speed and high-capacity seamless data transmission due to the integration of ground, air, and space communications. However, this imposes a new paradigm on the architecture design of the integrated SIN. Recently, reconfigurable intelligent surfaces (RISs) and mobile edge computing (MEC) are the most promising techniques, conceived to improve communication and computation capability by reconfiguring the wireless propagation environment and offloading. Hence, converging RISs and MEC in SIN is becoming an effort to reap the double benefits of computation and communication. In this article, we propose an RIS-assisted collaborative MEC architecture for SIN and discuss its implementation. Then we present its potential benefits, major challenges, and feasible applications. Subsequently, we study different cases to evaluate the system data rate and latency. Finally, we conclude with a list of open issues in this research area.

cs.NI↗

arXiv2021

Learning without Knowing: Unobserved Context in Continuous Transfer Reinforcement Learning

In this paper, we consider a transfer Reinforcement Learning (RL) problem in continuous state and action spaces, under unobserved contextual information. For example, the context can represent the mental view of the world that an expert agent has formed through past interactions with this world. We assume that this context is not accessible to a learner agent who can only observe the expert data. Then, our goal is to use the context-aware expert data to learn an optimal context-unaware policy for the learner using only a few new data samples. Such problems are typically solved using imitation learning that assumes that both the expert and learner agents have access to the same information. However, if the learner does not know the expert context, using the expert data alone will result in a biased learner policy and will require many new data samples to improve. To address this challenge, in this paper, we formulate the learning problem as a causal bound-constrained Multi-Armed-Bandit (MAB) problem. The arms of this MAB correspond to a set of basis policy functions that can be initialized in an unsupervised way using the expert data and represent the different expert behaviors affected by the unobserved context. On the other hand, the MAB constraints correspond to causal bounds on the accumulated rewards of these basis policy functions that we also compute from the expert data. The solution to this MAB allows the learner agent to select the best basis policy and improve it online. And the use of causal bounds reduces the exploration variance and, therefore, improves the learning rate. We provide numerical experiments on an autonomous driving example that show that our proposed transfer RL method improves the learner's policy faster compared to existing imitation learning methods and enjoys much lower variance during training.

cs.LG↗

arXiv2021

DynaEval: Unifying Turn and Dialogue Level Evaluation

A dialogue is essentially a multi-turn interaction among interlocutors. Effective evaluation metrics should reflect the dynamics of such interaction. Existing automatic metrics are focused very much on the turn-level quality, while ignoring such dynamics. To this end, we propose DynaEval, a unified automatic evaluation framework which is not only capable of performing turn-level evaluation, but also holistically considers the quality of the entire dialogue. In DynaEval, the graph convolutional network (GCN) is adopted to model a dialogue in totality, where the graph nodes denote each individual utterance and the edges represent the dependency between pairs of utterances. A contrastive loss is then applied to distinguish well-formed dialogues from carefully constructed negative samples. Experiments show that DynaEval significantly outperforms the state-of-the-art dialogue coherence model, and correlates strongly with human judgements across multiple dialogue evaluation aspects at both turn and dialogue level.

cs.CL↗

arXiv2021

Naked emergence of an anti-Hale active region I. Overall evolution and magnetic properties

{In order to understand the emergence of the active region, we investigate the emerging process and magnetic properties of a naked anti-Hale active region during the period between August 24 to 25, 2018.} {Using the data from Helioseismic and Magnetic Imager on board the Soar Dynamic Observatory and the New Vacuum Solar Telescope, we calculated different evolving parameters (such as pole separation, tilt angle) and magnetic parameters (such as vertical electric current, force-free parameter, relative magnetic helicity) during the emergence of the active region. With these calculated parameters and some reasonable assumptions, we use two different methods to estimate the twist of the active region.} {The magnetic flux and pole separation continue increasing while the tilt angle exhibits a decreasing pattern during the emergence of the active region. The increase of the pole separation is mainly contributed as a result of the enhancement in the longitude direction. A power-law relationship between pole separation and total flux is found during the emergence of the active region. On the other hand, it is found that both the positive and negative electric currents increased equivalently and the average flux-weighted force-free parameter $\tilde α$ remains almost consistently positive, on the order of $\sim$ 10$^{-8}$ m$^{-1}$. The relative magnetic helicity is mainly contributed by the shear term, while the relative magnetic helicity injection flux of the shear term changes its sign at the latter stage of the emergence. The twist number of the whole active region remains on the order of 10$^{-1}$ turns during the emergence of the active region.} {We find that the magnetic flux tube with low twist also could emerge into the solar atmosphere.}

astro-ph.SR↗

arXiv2021

PanGu-$α$: Large-scale Autoregressive Pretrained Chinese Language Models with Auto-parallel Computation

Large-scale Pretrained Language Models (PLMs) have become the new paradigm for Natural Language Processing (NLP). PLMs with hundreds of billions parameters such as GPT-3 have demonstrated strong performances on natural language understanding and generation with \textit{few-shot in-context} learning. In this work, we present our practice on training large-scale autoregressive language models named PanGu-$α$, with up to 200 billion parameters. PanGu-$α$ is developed under the MindSpore and trained on a cluster of 2048 Ascend 910 AI processors. The training parallelism strategy is implemented based on MindSpore Auto-parallel, which composes five parallelism dimensions to scale the training task to 2048 processors efficiently, including data parallelism, op-level model parallelism, pipeline model parallelism, optimizer model parallelism and rematerialization. To enhance the generalization ability of PanGu-$α$, we collect 1.1TB high-quality Chinese data from a wide range of domains to pretrain the model. We empirically test the generation ability of PanGu-$α$ in various scenarios including text summarization, question answering, dialogue generation, etc. Moreover, we investigate the effect of model scales on the few-shot performances across a broad range of Chinese NLP tasks. The experimental results demonstrate the superior capabilities of PanGu-$α$ in performing various tasks under few-shot or zero-shot settings.

cs.CL↗

arXiv2021

A Joint Energy and Latency Framework for Transfer Learning over 5G Industrial Edge Networks

In this paper, we propose a transfer learning (TL)-enabled edge-CNN framework for 5G industrial edge networks with privacy-preserving characteristic. In particular, the edge server can use the existing image dataset to train the CNN in advance, which is further fine-tuned based on the limited datasets uploaded from the devices. With the aid of TL, the devices that are not participating in the training only need to fine-tune the trained edge-CNN model without training from scratch. Due to the energy budget of the devices and the limited communication bandwidth, a joint energy and latency problem is formulated, which is solved by decomposing the original problem into an uploading decision subproblem and a wireless bandwidth allocation subproblem. Experiments using ImageNet demonstrate that the proposed TL-enabled edge-CNN framework can achieve almost 85% prediction accuracy of the baseline by uploading only about 1% model parameters, for a compression ratio of 32 of the autoencoder.

cs.NI↗

arXiv2021

LEAP: Learning Articulated Occupancy of People

Substantial progress has been made on modeling rigid 3D objects using deep implicit representations. Yet, extending these methods to learn neural models of human shape is still in its infancy. Human bodies are complex and the key challenge is to learn a representation that generalizes such that it can express body shape deformations for unseen subjects in unseen, highly-articulated, poses. To address this challenge, we introduce LEAP (LEarning Articulated occupancy of People), a novel neural occupancy representation of the human body. Given a set of bone transformations (i.e. joint locations and rotations) and a query point in space, LEAP first maps the query point to a canonical space via learned linear blend skinning (LBS) functions and then efficiently queries the occupancy value via an occupancy network that models accurate identity- and pose-dependent deformations in the canonical space. Experiments show that our canonicalized occupancy estimation with the learned LBS functions greatly improves the generalization capability of the learned occupancy representation across various human shapes and poses, outperforming existing solutions in all settings.

cs.CV↗

arXiv2021

Incommensurate smectic phase in close proximity to the high-Tc superconductor FeSe/SrTiO3

Superconductivity is significantly enhanced in monolayer FeSe grown on SrTiO3, but not for multilayer films, in which large strength of nematicity develops. However, the link between the high-transition temperature superconductivity in monolayer and the correlation related nematicity in multilayer FeSe films is not well understood. Here, we use low-temperature scanning tunneling microscopy to study few-layer FeSe thin films grown by molecular beam epitaxy. We observe an incommensurate long-range smectic phase, which solely appears in bilayer FeSe films. The smectic order still locally exists and gradually fades away with increasing film thickness, while it suddenly vanishes in monolayer FeSe, indicative of an abrupt smectic phase transition. Surface alkali-metal doping can suppress the smectic phase and induce high-Tc superconductivity in bilayer FeSe. Our observations provide evidence that the monolayer FeSe is in close proximity to the smectic phase, and its superconductivity is likely enhanced by this electronic instability as well.

cond-mat.supr-con↗

arXiv2021

Distinctive momentum dependent charge-density-wave gap observed in CsV$_3$Sb$_5$ superconductor with topological Kagome lattice

CsV$_3$Sb$_5$ is a newly discovered Kagome superconductor that attracts great interest due to its topological nontrivial band structure and the coexistence of superconductivity and charge-density-wave (CDW) with many exotic properties. Here, we report the detailed characterization of the CDW gap in high-quality CsV$_3$Sb$_5$ single crystals using high-resolution angle-resolved photoemission spectroscopy. We find that the CDW gap is strongly momentum dependent. While gapped around the $M$ point, the electronic states remain gapless around the $Γ$ point and along the $Γ$-$K$ direction. Such momentum dependence indicates that the CDW is driven by the scattering of electrons between neighboring $M$ points, where the band structure hosts multiple saddle points and the density of state diverges near the Fermi level. Our observations of the partially gapped Fermi surface and strongly momentum-dependent CDW gap not only provide a foundation for uncovering the mechanism of CDW in CsV$_3$Sb$_5$, but also shed light on the understanding of how the CDW coexists with superconductivity in this topological Kagome superconductor.

cond-mat.supr-con↗

arXiv2021

We are More than Our Joints: Predicting how 3D Bodies Move

A key step towards understanding human behavior is the prediction of 3D human motion. Successful solutions have many applications in human tracking, HCI, and graphics. Most previous work focuses on predicting a time series of future 3D joint locations given a sequence 3D joints from the past. This Euclidean formulation generally works better than predicting pose in terms of joint rotations. Body joint locations, however, do not fully constrain 3D human pose, leaving degrees of freedom undefined, making it hard to animate a realistic human from only the joints. Note that the 3D joints can be viewed as a sparse point cloud. Thus the problem of human motion prediction can be seen as point cloud prediction. With this observation, we instead predict a sparse set of locations on the body surface that correspond to motion capture markers. Given such markers, we fit a parametric body model to recover the 3D shape and pose of the person. These sparse surface markers also carry detailed information about human movement that is not present in the joints, increasing the naturalness of the predicted motions. Using the AMASS dataset, we train MOJO, which is a novel variational autoencoder that generates motions from latent frequencies. MOJO preserves the full temporal resolution of the input motion, and sampling from the latent frequencies explicitly introduces high-frequency components into the generated motion. We note that motion prediction methods accumulate errors over time, resulting in joints or markers that diverge from true human bodies. To address this, we fit SMPL-X to the predictions at each time step, projecting the solution back onto the space of valid bodies. These valid markers are then propagated in time. Experiments show that our method produces state-of-the-art results and realistic 3D body animations. The code for research purposes is at https://yz-cnsdqz.github.io/MOJO/MOJO.html

cs.CV↗

arXiv2021

Random Shifting Intelligent Reflecting Surface for OTP Encrypted Data Transmission

In this paper, we propose a novel encrypted data transmission scheme using an intelligent reflecting surface (IRS) to generate secret keys in wireless communication networks. We show that perfectly secure one-time pad (OTP) communications can be established by using a simple random phase shifting of the IRS elements. To maximize the secure transmission rate, we design an optimal time slot allocation algorithm for the IRS secret key generation and the encrypted data transmission phases. Moreover, a theoretical expression of the key generation rate is derived based on Poisson point process (PPP) for the practical scenario when eavesdroppers' channel state information (CSI) is unavailable. Simulation results show that employing our IRS-based scheme can significantly improve the encrypted data transmission performance for a wide-range of wireless channel gains and system parameters.

cs.IT↗

arXiv2021

Secret Key Generation for Intelligent Reflecting Surface Assisted Wireless Communication Networks

We propose and analyze secret key generation using intelligent reflecting surface (IRS) assisted wireless communication networks. To this end, we first formulate the minimum achievable secret key capacity for an IRS acting as a passive beamformer in the presence of multiple eavesdroppers. Next, we develop an optimization framework for the IRS reflecting coefficients based on the secret key capacity lower bound. To derive a tractable and efficient solution, we design and analyze a semidefinite relaxation (SDR) and successive convex approximation (SCA) based algorithm for the proposed optimization. Simulation results show that employing our IRS-based algorithm can significantly improve the secret key generation capacity for a wide-range of wireless channel parameters.

eess.SP↗

arXiv2021

Learning to Represent and Predict Sets with Deep Neural Networks

In this thesis, we develop various techniques for working with sets in machine learning. Each input or output is not an image or a sequence, but a set: an unordered collection of multiple objects, each object described by a feature vector. Their unordered nature makes them suitable for modeling a wide variety of data, ranging from objects in images to point clouds to graphs. Deep learning has recently shown great success on other types of structured data, so we aim to build the necessary structures for sets into deep neural networks. The first focus of this thesis is the learning of better set representations (sets as input). Existing approaches have bottlenecks that prevent them from properly modeling relations between objects within the set. To address this issue, we develop a variety of techniques for different scenarios and show that alleviating the bottleneck leads to consistent improvements across many experiments. The second focus of this thesis is the prediction of sets (sets as output). Current approaches do not take the unordered nature of sets into account properly. We determine that this results in a problem that causes discontinuity issues with many set prediction tasks and prevents them from learning some extremely simple datasets. To avoid this problem, we develop two models that properly take the structure of sets into account. Various experiments show that our set prediction techniques can significantly benefit over existing approaches.

cs.LG↗

arXiv2021

Formal Verification of Stochastic Systems with ReLU Neural Network Controllers

In this work, we address the problem of formal safety verification for stochastic cyber-physical systems (CPS) equipped with ReLU neural network (NN) controllers. Our goal is to find the set of initial states from where, with a predetermined confidence, the system will not reach an unsafe configuration within a specified time horizon. Specifically, we consider discrete-time LTI systems with Gaussian noise, which we abstract by a suitable graph. Then, we formulate a Satisfiability Modulo Convex (SMC) problem to estimate upper bounds on the transition probabilities between nodes in the graph. Using this abstraction, we propose a method to compute tight bounds on the safety probabilities of nodes in this graph, despite possible over-approximations of the transition probabilities between these nodes. Additionally, using the proposed SMC formula, we devise a heuristic method to refine the abstraction of the system in order to further improve the estimated safety bounds. Finally, we corroborate the efficacy of the proposed method with simulation results considering a robot navigation example and comparison against a state-of-the-art verification scheme.

eess.SY↗

arXiv2021

Revealing the intrinsic superconducting gap anisotropy in surface-neutralized BaFe$_2$(As$_{0.7}$P$_{0.3}$)$_2$

Alkaline-earth iron arsenide (122) is one of the most studied families of iron-based superconductors, especially for angle-resolved photoemission spectroscopy. While extensive photoemission results have been obtained, the surface complexity of 122 caused by its charge-non-neutral surface is rarely considered. Here, we show that the surface of 122 can be neutralized by potassium deposition. In potassium-coated BaFe$_2$(As$_{0.7}$P$_{0.3}$)$_2$, the surface-induced spectral broadening is strongly suppressed, and hence the coherent spectra that reflect the intrinsic bulk electronic state recover. This enables the measuring of superconducting gap with unpreceded precision. The result shows the existence of two pairing channels. While the gap anisotropy on the outer hole/electron pockets can be well fitted using an s$_\pm$ gap function, the gap anisotropy on the inner hole/electron shows a clear deviation. Our results provide quantitative constraints for refining theoretical models and also demonstrate an experimental method for revealing the intrinsic electronic properties of 122 in future studies.

cond-mat.supr-con↗

arXiv2021

An Unsupervised Sentence Embedding Method by Mutual Information Maximization

BERT is inefficient for sentence-pair tasks such as clustering or semantic search as it needs to evaluate combinatorially many sentence pairs which is very time-consuming. Sentence BERT (SBERT) attempted to solve this challenge by learning semantically meaningful representations of single sentences, such that similarity comparison can be easily accessed. However, SBERT is trained on corpus with high-quality labeled sentence pairs, which limits its application to tasks where labeled data is extremely scarce. In this paper, we propose a lightweight extension on top of BERT and a novel self-supervised learning objective based on mutual information maximization strategies to derive meaningful sentence embeddings in an unsupervised manner. Unlike SBERT, our method is not restricted by the availability of labeled data, such that it can be applied on different domain-specific corpus. Experimental results show that the proposed method significantly outperforms other unsupervised sentence embedding baselines on common semantic textual similarity (STS) tasks and downstream supervised tasks. It also outperforms SBERT in a setting where in-domain labeled data is not available, and achieves performance competitive with supervised methods on various tasks.

cs.CL↗

arXiv2021

A Survey on Concept Factorization: From Shallow to Deep Representation Learning

The quality of learned features by representation learning determines the performance of learning algorithms and the related application tasks (such as high-dimensional data clustering). As a relatively new paradigm for representation learning, Concept Factorization (CF) has attracted a great deal of interests in the areas of machine learning and data mining for over a decade. Lots of effective CF based methods have been proposed based on different perspectives and properties, but note that it still remains not easy to grasp the essential connections and figure out the underlying explanatory factors from exiting studies. In this paper, we therefore survey the recent advances on CF methodologies and the potential benchmarks by categorizing and summarizing the current methods. Specifically, we first re-view the root CF method, and then explore the advancement of CF-based representation learning ranging from shallow to deep/multilayer cases. We also introduce the potential application areas of CF-based methods. Finally, we point out some future directions for studying the CF-based representation learning. Overall, this survey provides an insightful overview of both theoretical basis and current developments in the field of CF, which can also help the interested researchers to understand the current trends of CF and find the most appropriate CF techniques to deal with particular applications.

cs.LG↗

arXiv2021

Quantifying the importance of firms by means of reputation and network control

The reputation of firms is largely channeled through their ownership structure. We use this relation to determine reputation spillovers between transnational companies and their participated companies in an ownership network core of 1318 firms. We then apply concepts of network controllability to identify minimum sets of driver nodes (MDS) of 314 firms in this network. The importance of these driver nodes is classified regarding their control contribution, their operating revenue, and their reputation. The latter two are also taken as proxies for the access costs when utilizing firms as driver nodes. Using an enrichment analysis, we find that firms with high reputation maintain the controllability of the network, but rarely become top drivers, whereas firms with medium reputation most likely become top driver nodes. We further show that MDSs with lower access costs can be used to control the reputation dynamics in the whole network.

physics.soc-ph↗