Searcharxiv⌕ Search

arXiv subjects

Chen Xu

Publications and source records attributed to Chen Xu.

At least 235 records · Page 13Linked to original sources
arXiv2019

Realistic Channel Models Pre-training

In this paper, we propose a neural-network-based realistic channel model with both the similar accuracy as deterministic channel models and uniformity as stochastic channel models. To facilitate this realistic channel modeling, a multi-domain channel embedding method combined with self-attention mechanism is proposed to extract channel features from multiple domains simultaneously. This 'one model to fit them all' solution employs available wireless channel data as the only data set for self-supervised pre-training. With the permission of users, network operators or other organizations can make use of some available user specific data to fine-tune this pre-trained realistic channel model for applications on channel-related downstream tasks. Moreover, even without fine-tuning, we show that the pre-trained realistic channel model itself is a great tool with its understanding of wireless channel.

eess.SP↗
arXiv2019

Lower bounds for fluctuations in first-passage percolation for general distributions

In first-passage percolation (FPP), one assigns i.i.d.~weights to the edges of the cubic lattice $\mathbb{Z}^d$ and analyzes the induced weighted graph metric. If $T(x,y)$ is the distance between vertices $x$ and $y$, then a primary question in the model is: what is the order of the fluctuations of $T(0,x)$? It is expected that the variance of $T(0,x)$ grows like the norm of $x$ to a power strictly less than 1, but the best lower bounds available are (only in two dimensions) of order $\log \|x\|$. This result was found in the '90s and there has not been any improvement since. In this paper, we address the problem of getting stronger fluctuation bounds: to show that $T(0,x)$ is with high probability not contained in an interval of size $o(\log \|x\|)^{1/2}$, and similar statements for FPP in thin cylinders. Such statements have been proved for special edge-weight distributions, and here we obtain such bounds for general edge-weight distributions. The methods involve inducing a fluctuation in the number of edges in a box whose weights are of "hi-mode" (large).

math.PR↗
arXiv2019

Predicting the Mumble of Wireless Channel with Sequence-to-Sequence Models

Accurate prediction of fading channel in the upcoming transmission frame is essential to realize adaptive transmission for transmitters, and receivers with the ability of channel prediction can also save some computations of channel estimation. However, due to the rapid channel variation and channel estimation error, reliable prediction is hard to realize. In this situation, an appropriate channel model should be selected, which can cover both the statistical model and small scale fading of channel, this reminds us the natural languages, which also have statistical word frequency and specific sentences. Accordingly, in this paper, we take wireless channel model as a language model, and the time-varying channel as talking in this language, while the realistic noisy estimated channel can be compared with mumbling. Furthermore, in order to utilize as much as possible the information a channel coefficient takes, we discard the conventional two features of absolute value and phase, replacing with hundreds of features which will be learned by our channel model, to do this, we use a vocabulary to map a complex channel coefficient into an ID, which is represented by a vector of real numbers. Recurrent neural networks technique is used as its good balance between memorization and generalization, moreover, we creatively introduce sequence-to-sequence (seq2seq) models in time series channel prediction, which can translates past channel into future channel. The results show that realistic channel prediction with superior performance relative to channel estimation is attainable.

eess.SP↗
arXiv2019

Distributed Feature Screening via Componentwise Debiasing

Feature screening is a powerful tool in the analysis of high dimensional data. When the sample size $N$ and the number of features $p$ are both large, the implementation of classic screening methods can be numerically challenging. In this paper, we propose a distributed screening framework for big data setup. In the spirit of "divide-and-conquer", the proposed framework expresses a correlation measure as a function of several component parameters, each of which can be distributively estimated using a natural U-statistic from data segments. With the component estimates aggregated, we obtain a final correlation estimate that can be readily used for screening features. This framework enables distributed storage and parallel computing and thus is computationally attractive. Due to the unbiased distributive estimation of the component parameters, the final aggregated estimate achieves a high accuracy that is insensitive to the number of data segments $m$ specified by the problem itself or to be chosen by users. Under mild conditions, we show that the aggregated correlation estimator is as efficient as the classic centralized estimator in terms of the probability convergence bound; the corresponding screening procedure enjoys sure screening property for a wide range of correlation measures. The promising performances of the new method are supported by extensive numerical examples.

stat.ME↗
arXiv2018

Beam-controlled spectral-selective architecture with planar polydimethylsiloxane/metal-films for all-day radiative cooling

Radiative cooling is a passive cooling strategy with zero consumption of electricity. Although this technology can work well during optimal atmospheric conditions at nighttime, it is essential to achieve efficient radiative cooling during daytime when peak cooling demand actually occurs. In this article, we report an inexpensive planar polydimethylsiloxane (PDMS)/metal thermal emitter, i.e., a thin film structure fabricated using fast solution coating process that is scalable for large area manufacturing. By manipulating the beaming effect of the thermal radiation, temperature reduction of 9.5 °C and 11.0 °C were demonstrated in the laboratory and out-door environment, respectively. In addition, a spectral-selective solar shelter architecture was designed and implemented to suppress the solar input during the daytime. Due to the enhanced directionality of the thermal emission, the dependence of the radiative cooling performance on the surrounding environment was minimized. Out-door experiments were performed in Buffalo NY, realizing continuous all-day radiative cooling with an average power of ~120 W/m2 on a typical clear sunny day at Northern United States latitudes. This practical passive cooling strategy that cools without any electricity input could have a significant impact on global energy consumption.

physics.app-ph↗
arXiv2018

A Flip-Syndrome-List Polar Decoder Architecture for Ultra-Low-Latency Communications

We consider practical hardware implementation of Polar decoders. To reduce latency due to the serial nature of successive cancellation (SC), existing optimizations improve parallelism with two approaches, i.e., multi-bit decision or reduced path splitting. In this paper, we combine the two procedures into one with an error-pattern-based architecture. It simultaneously generates a set of candidate paths for multiple bits with pre-stored patterns. For rate-1 (R1) or single parity-check (SPC) nodes, we prove that a small number of deterministic patterns are required to guarantee performance preservation. For general nodes, low-weight error patterns are indexed by syndrome in a look-up table and retrieved in O(1) time. The proposed flip-syndrome-list (FSL) decoder fully parallelizes all constituent code blocks without sacrificing performance, thus is suitable for ultra-low-latency applications. Meanwhile, two code construction optimizations are presented to further reduce complexity and improve performance, respectively.

cs.IT↗
arXiv2018

Independently tunable dual-spectral electromagnetically induced transparency in a terahertz metal-graphene metamaterial

We theoretically investigate the interaction between the conductive graphene layer with the dual-spectral electromagnetically induced transparency (EIT) metamaterial and achieve independent amplitude modulation of the transmission peaks in terahertz (THz) regime. The dual-spectral EIT resonance results from the strong near field coupling effects between the bright cut wire resonator (CWR) in the middle and two dark double-split ring resonators (DSRRs) on the two sides. By integrating monolayer graphene under the dark mode resonators, the two transmission peaks of the EIT resonance can exhibit independent amplitude modulation via shifting the Fermi level of the corresponding graphene layer. The physical mechanism of the modulation can be attributed to the variation of damping factors of the dark mode resonators arising from the tunable conductivity of graphene. This work shows great prospects in designing multiple-spectral THz functional devices with highly flexible tunability and implies promising applications in multi-channel selective switching, modulation and slow light.

physics.optics↗
arXiv2018

Actively tunable slow light in a terahertz hybrid metal-graphene metamaterial

We theoretically and numerically demonstrate an actively tunable slow light in a hybrid metal-graphene metamaterial in the terahertz (THz) regime. In the unit cell, the near field coupling between the metallic elements including the bright cut wire resonator and the dark double split-ring resonator gives rises to a pronounced transmission peak. By positioning a monolayer graphene under the dark mode resonator, an active modulation of the near field coupling is achieved via shifting the Fermi level of graphene. The physical origin can be attributed to the variation in the damping rate of the dark mode resonator arising from the conductive effect of graphene. Accompanied with the actively tunable near filed coupling effect is the dynamically controllable phase dispersion, allowing for the highly tunable slow light effect. This work offers an alternative way to design compact slow light devices in the THz regime for future optical signal processing applications.

physics.optics↗
arXiv2018

Alternating Multi-bit Quantization for Recurrent Neural Networks

Recurrent neural networks have achieved excellent performance in many applications. However, on portable devices with limited resources, the models are often too large to deploy. For applications on the server with large scale concurrent requests, the latency during inference can also be very critical for costly computing resources. In this work, we address these problems by quantizing the network, both weights and activations, into multiple binary codes {-1,+1}. We formulate the quantization as an optimization problem. Under the key observation that once the quantization coefficients are fixed the binary codes can be derived efficiently by binary search tree, alternating minimization is then applied. We test the quantization for two well-known RNNs, i.e., long short term memory (LSTM) and gated recurrent unit (GRU), on the language models. Compared with the full-precision counter part, by 2-bit quantization we can achieve ~16x memory saving and ~6x real inference acceleration on CPUs, with only a reasonable loss in the accuracy. By 3-bit quantization, we can achieve almost no loss in the accuracy or even surpass the original model, with ~10.5x memory saving and ~3x real inference acceleration. Both results beat the exiting quantization works with large margins. We extend our alternating quantization to image classification tasks. In both RNNs and feedforward neural networks, the method also achieves excellent performance.

cs.LG↗
arXiv2018

Concentration of Geodesics in Directed Bernoulli Percolation

For directed Bernoulli last passage percolation with i.i.d.~weights on vertices over a $n\times n$ grid and for $n$ large enough, the geodesics are shown to be concentrated in a cylinder, centered on the main diagonal and of width of order $n^{(2κ+2)/(2κ+3)}\sqrt{\ln n}$, where $1\leκ<\infty$ is the curvature power of the shape function at $(1,1)$. The methodology of proof is robust enough to also apply to directed Bernoulli first passage site percolation, and further to longest common subsequences in random words.

math.PR↗
arXiv2017

Tunable light trapping and absorption enhancement with graphene-based complementary metamaterials

Surface plasmon resonance (SPR) has been intensively investigated and widely exploited to trap the incident light and enhance absorption in the optoelectronic devices. The availability of graphene as a plasmonic material with strong half-metallicity and continuously tunable surface conductivity makes it promising to dynamically modulate the absorption enhancement with graphene-based metamaterials. Here we numerically demonstrate tunable light trapping and absorption enhancement can be realized with graphene-based complementary metamaterials. Furthermore, we also explore the polarization sensitivity in the proposed device, in which case either TM or TE plane wave at the specific wavelength can be efficiently absorbed by simply manipulating the Fermi energy of graphene. Therefore, this work can find potential applications in the next generation of photodetectors with tunable spectral and polarization selectivity in the mid-infrared and terahertz (THz) regimes.

physics.optics↗
arXiv2017

Active modulation of electromagnetically induced transparency analogue in terahertz hybrid metal-graphene metamaterials

Metamaterial analogues of electromagnetically induced transparency (EIT) have been intensively studied and widely employed for slow light and enhanced nonlinear effects. In particular, the active modulation of the EIT analogue and well-controlled group delay in metamaterials have shown great prospects in optical communication networks. Previous studies have focused on the optical control of the EIT analogue by integrating the photoactive materials into the unit cell, however, the response time is limited by the recovery time of the excited carriers in these bulk materials. Graphene has recently emerged as an exceptional optoelectronic material. It shows an ultrafast relaxation time on the order of picosecond and its conductivity can be tuned via manipulating the Fermi energy. Here we integrate a monolayer graphene into metal-based terahertz (THz) metamaterials, and realize a complete modulation in the resonance strength of the EIT analogue at the accessible Fermi energy. The physical mechanism lies in the active tuning the damping rate of the dark mode resonator through the recombination effect of the conductive graphene. Note that the monolayer morphology in our work is easier to fabricate and manipulate than isolated fashion. This work presents a novel modulation strategy of the EIT analogue in the hybrid metamaterials, and pave the way towards designing very compact slow light devices to meet future demand of ultrafast optical signal processing.

physics.optics↗
arXiv2017

A spectrally tunable plasmonic photosensor with an ultrathin semiconductor region

Surface plasmon resonance (SPR) has been widely utilized to improve the absorption performance in the photosensors. Graphene has emerged as a promising plasmonic material, which supports tunable SPR and shows significant flexibility over metals. In this letter, a hybrid photosensor based on the integration of periodic cross-shaped graphene arrays with an ultrathin light-absorbing semiconductor is proposed. A tenfold absorption enhancement over a large range of the incidence angle for both light polarizations as well as a considerably high photogeneration rate ($\sim10^{37}$) is demonstrated at the resonance. Compared with traditional metal-based plasmon-enhanced photosensors, the absorption enhancement here can be expediently tuned with manipulating the Fermi energy of graphene. The proposed photosensor can amplify the photoresponse to the incidence light at the selected wavelength and thus be utilized in photosensing with high efficiency and tunable spectral selectivity in the mid-infrared (mid-IR) and terahertz (THz) regime.

physics.optics↗
arXiv2017

Strong interaction between graphene layer and Fano resonance in terahertz metamaterials

Graphene has emerged as a promising building block in the modern optics and optoelectronics due to its novel optical and electrical properties. In the mid-infrared and terahertz (THz) regime, graphene behaves like metals and supports surface plasmon resonances (SPRs). Moreover, the continuously tunable conductivity of graphene enables active SPRs and gives rise to a range of active applications. However, the interaction between graphene and metal-based resonant metamaterials has not been fully understood. In this work, a simulation investigation on the interaction between the graphene layer and THz resonances supported by the two-gap split ring metamaterials is systematically conducted. The simulation results show that the graphene layer can substantially reduce the Fano resonance and even switch it off, while leave the dipole resonance nearly unaffected, which phenomenon is well explained with the high conductivity of graphene. With the manipulation of graphene conductivity via altering its Fermi energy or layer number, the amplitude of the Fano resonance can be modulated. The tunable Fano resonance here together with the underlying physical mechanism can be strategically important in designing active metal-graphene hybrid metamaterials. In addition, the "sensitivity" to the graphene layer of the Fano resonance is also highly appreciated in the field of ultrasensitive sensing, where the novel physical mechanism can be employed in sensing other graphene-like two-dimensional (2D) materials or biomolecules with the high conductivity.

physics.optics↗
arXiv2016

A Unified Convex Surrogate for the Schatten-$p$ Norm

The Schatten-$p$ norm ($0 0$ satisfying $1/p=1/p_1+1/p_2$, there is an equivalence between the Schatten-$p$ norm of one matrix and the Schatten-$p_1$ and the Schatten-$p_2$ norms of its two factor matrices. We further extend the equivalence to multiple factor matrices and show that all the factor norms can be convex and smooth for any $p>0$. In contrast, the original Schatten-$p$ norm for $0<p<1$ is non-convex and non-smooth. As an example we conduct experiments on matrix completion. To utilize the convexity of the factor matrix norms, we adopt the accelerated proximal alternating linearized minimization algorithm and establish its sequence convergence. Experiments on both synthetic and real datasets exhibit its superior performance over the state-of-the-art methods. Its speed is also highly competitive.

stat.ML↗
arXiv2016

Atypical viral dynamics from transport through popular places

The flux of visitors through popular places undoubtedly influences viral spreading -- from H1N1 and Zika viruses spreading through physical spaces such as airports, to rumors and ideas spreading though online spaces such as chatrooms and social media. However there is a lack of understanding of the types of viral dynamics that can result. Here we present a minimal dynamical model which focuses on the time-dependent interplay between the {\em mobility through} and the {\em occupancy of} such spaces. Our generic model permits analytic analysis while producing a rich diversity of infection profiles in terms of their shapes, durations, and intensities. The general features of these theoretical profiles compare well to real-world data of recent social contagion phenomena.

q-bio.PE↗