BZPEERReading, trust and collaboration for researchDiscover
Workspaces
Network
Opportunities
Account
Researcher profile
Peng Luo contributes to research discovery and scholarly infrastructure.
Trust snapshot
Actions
Identity and collaboration
Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.
Log in to claimDirect collaboration
Claim this author entity first to unlock direct invitations.
Research graph
Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.
BZPEER is loading the nearby papers, people, topics and institutions for this page.
Published work
This work presents a 28nm 13.93mm2 CNN-Transformer accelerator for semantic segmentation, achieving 3.86-to-10.91x energy reduction over previous designs. It features a hybrid attention unit, layer-fusion scheduler, and cascaded feature-map pruner, with peak energy efficiency of 52.90TOPS/W (INT8).
Semantic place annotation can provide individual semantics, which can be of great help in the field of trajectory data mining. Most existing methods rely on annotated or external data and require retraining following a change of region, thus preventing their large-scale applications. Herein, we propose an unsupervised method denoted as UPAPP for the semantic place annotation of trajectories using spatiotemporal information. The Bayesian Criterion is specifically employed to decompose the spatiotemporal probability of the candidate place into spatial probability, duration probability, and visiting time probability. Spatial information in ROI and POI data is subsequently adopted to calculate the spatial probability. In terms of the temporal probabilities, the Term Frequency Inverse Document Frequency weighting algorithm is used to count the potential visits to different place types in the trajectories, and generates the prior probabilities of the visiting time and duration. The spatiotemporal probability of the candidate place is then combined with the importance of the place category to annotate the visited places. Validation with a trajectory dataset collected by 709 volunteers in Beijing showed that our method achieved an overall and average accuracy of 0.712 and 0.720, respectively, indicating that the visited places can be annotated accurately without any external data.
This work investigates continuous time stochastic differential games with a large number of players, whose costs and dynamics interact through the empirical distribution of both their states and their controls. The control processes are assumed to be open-loop. We give regularity conditions guaranteeing that if the finite-player game admits a Nash equilibrium, then both the sequence of equilibria and the corresponding states processes satisfy a Sanov-type large deviation principle. The result requires existence of a Lipschitz continuous solution of the master equation of the corresponding mean field game, and is based on concentration inequalities for Lipschitz FBSDEs. The result carries over to cooperative (i.e. central planner) games. We study the linear-quadratic case of such games in details.
Let $Ω\subset\mathbb{R}^N$ be a smooth bounded domain with $N\ge2$ and $Ω_ε=Ω\backslash B(P,ε)$ where $B(P,ε)$ is the ball centered at $P\inΩ$ and radius $ε$. In this paper, we establish the number, location and non-degeneracy of critical points of the Robin function in $Ω_ε$ for $ε$ small enough. We will show that the location of $P$ plays a crucial role on the existence and multiplicity of the critical points. The proof of our result is a consequence of delicate estimates on the Green function near to $\partial B(P,ε)$. Some applications to compute the exact number of solutions of related well-studied nonlinear elliptic problems will be showed.
Human activity analysis based on sensor data plays a significant role in behavior sensing, human-machine interaction, health care, and so on. The current research focused on recognizing human activity and posture at the activity pattern level, neglecting the effective fusion of multi-sensor data and assessing different movement styles at the individual level, thus introducing the challenge to distinguish individuals in the same movement. In this study, the concept of RunnerDNA, consisting of five interpretable indicators, balance, stride, steering, stability, and amplitude, was proposed to describe human activity at the individual level. We collected smartphone multi-sensor data from 33 volunteers who engaged in physical activities such as walking, running, and bicycling and calculated the data into five indicators of RunnerDNA. The indicators were then used to build random forest models and recognize movement activities and the identity of users. The results show that the proposed model has high accuracy in identifying activities (accuracy of 0.679) and is also effective in predicting the identity of running users. Furthermore, the accuracy of the human activity recognition model has significant improved by combing RunnerDNA and two motion feature indicators, velocity, and acceleration. Results demonstrate that RunnerDNA is an effective way to describe an individual's physical activity and helps us understand individual differences in sports style, and the significant differences in balance and amplitude between men and women were found.
Name lookup is a key technology for the forwarding plane of content router in Named Data Networking (NDN). To realize the efficient name lookup, what counts is deploying a highperformance index in content routers. So far, the proposed indexes have shown good performance, most of which are optimized for or evaluated with URLs collected from the current Internet, as the large-scale NDN names are not available yet. Unfortunately, the performance of these indexes is always impacted in terms of lookup speed, memory consumption and false positive probability, as the distributions of URLs retrieved in memory may differ from those of real NDN names independently generated by content-centric applications online. Focusing on this gap, a smart mapping model named Pyramid-NN via neural networks is proposed to build an index called LNI for NDN forwarding plane. Through learning the distributions of the names retrieved in the static memory, LNI can not only reduce the memory consumption and the probability of false positive, but also ensure the performance of real NDN name lookup. Experimental results show that LNI-based FIB can reduce the memory consumption to 58.258 MB for 2 million names. Moreover, as it can be deployed on SRAMs, the throughput is about 177 MSPS, which well meets the current network requirement for fast packet processing.
We are concerned with the Lane-Emden problem \begin{equation*} \begin{cases} -Δu=u^{p} &{\text{in}~Ω},\\[0.5mm] u>0 &{\text{in}~Ω},\\[0.5mm] u=0 &{\text{on}~\partial Ω}, \end{cases} \end{equation*} where $Ω\subset \mathbb R^2$ is a smooth bounded domain and $p>1$ is sufficiently large. Improving some known asymptotic estimates on the solutions, we prove the non-degeneracy and local uniqueness of the multi-spikes positive solutions for general domains. Our methods mainly use ODE's theory, various local Pohozaev identities, blow-up analysis and the properties of Green's function.
Employing wide-angle neutron spin echo spectroscopy, we measured the Q-dependent coherent intermediate scattering function of the prototypical ionic glass former Ca0.4K0.6(NO3)1.4, in the equilibrium and supercooled liquid states beyond the hydrodynamic regime. The data reveal a clear two-step relaxation: an exponential fast process, and a stretched exponential slow alpha process. de Gennes narrowing is observed in all characteristic variables of the alpha process: the relaxation time, amplitude, and stretching exponent. At all length scales probed, the relative amplitude of the alpha-relaxation decreases with increasing temperature and levels off in the normal liquid state. The temperature dependence of the stretching exponent and the relaxation time at different Q's indicate that modifications of the relaxation mechanisms at the local length scales, manifested as temperature independent dynamic heterogeneity and smaller deviations from Arrhenius behavior, have occurred even above the alpha-beta (Johari-Goldstein) bifurcation temperature.
In this paper we study the following torsion problem \begin{equation*} \begin{cases} -Δu=1~&\mbox{in}\ Ω,\\[1mm] u=0~&\mbox{on}\ \partialΩ. \end{cases} \end{equation*} Let $Ω\subset \mathbb{R}^2$ be a bounded, convex domain and $u_0(x)$ be the solution of above problem with its maximum $y_0\in Ω$. Steinerberger proved that there are universal constants $c_1, c_2>0$ satisfying \begin{equation*} λ_{\max}\left(D^2u_0(y_0)\right)\leq -c_1\mbox{exp}\left(-c_2\frac{\text{diam}(Ω)}{\mbox{inrad}(Ω)}\right). \end{equation*} And he proposed following open problem: "Does above result hold true on domains that are not convex but merely simply connected or perhaps only bounded? The proof uses convexity of the domain $Ω$ in a very essential way and it is not clear to us whether the statement remains valid in other settings." Here by some new idea involving the computations on Green's function, we compute the spectral gap $λ_{\max}D^2u(y_0)$ for some non-convex smooth bounded domains, which gives a negative answer to above open problem. Also some extensions are given.
In this paper we study the following problem \begin{equation} \begin{cases} -Δu=f(u)~&\mbox{in}\ Ω_\varepsilon,\\ u>0~&\mbox{in}\ Ω_\varepsilon,\\ u=0~&\mbox{on}\ \partialΩ_\varepsilon, \end{cases} \end{equation} where $Ω_\varepsilon=Ω\backslash B(P,\varepsilon)$, $Ω\subset R^N$ with $N\geq 2$ is a smooth bounded domain, $B(P,\varepsilon)$ is the ball centered at $P$ and radius $\varepsilon>0$ and $f$ is a smooth nonlinearity. By some computations involving the Green function and degree theory, we compute the number and location of critical points of solutions for small $\varepsilon>0$.
Channel estimation for the downlink of frequency division duplex (FDD) massive MIMO systems is well known to generate a large overhead as the amount of training generally scales with the number of transmit antennas in a MIMO system. In this paper, we consider the solution of extrapolating the channel frequency response from uplink pilot estimates to the downlink frequency band, which completely removes the training overhead. We first show that conventional estimators fail to achieve reasonable accuracy. We propose instead to use high-resolution channel estimation. We derive theoretical lower bounds (LB) for the mean squared error (MSE) of the extrapolated channel. Assuming that the paths are well separated, the LB is simplified in an expression that gives considerable physical insight. It is then shown that the MSE is inversely proportional to the number of receive antennas while the extrapolation performance penalty scales with the square of the ratio of the frequency offset and the training bandwidth. The channel extrapolation performance is validated through numeric simulations and experimental measurements taken in an anechoic chamber. Our main conclusion is that channel extrapolation is a viable solution for FDD massive MIMO systems if accurate system calibration is performed and favorable propagation conditions are present.
This paper investigates solvability of fully coupled systems of forward-backward stochastic differential equations (FBSDEs) with irregular coefficients. In particular, we assume that the coefficients of the FBSDEs are merely measurable and bounded in the forward process. We crucially use compactness results from the theory of Malliavin calculus to construct strong solutions. Despite the irregularity of the coefficients, the solutions turn out to be differentiable, at least in the Malliavin sense and, as functions of the initial variable, in the Sobolev sense.
In this paper we are concerned with the well-known Brezis-Nirenberg problem \begin{equation*} \begin{cases} -Δu= u^{\frac{N+2}{N-2}}+\varepsilon u, &{\text{in}~Ω},\\ u>0, &{\text{in}~Ω},\\ u=0, &{\text{on}~\partial Ω}. \end{cases} \end{equation*} The existence of multi-peak solutions to the above problem for small $\varepsilon>0$ was obtained by Musso and Pistoia. However, the uniqueness or the exact number of positive solutions to the above problem is still unknown. Here we focus on the local uniqueness of multi-peak solutions and the exact number of positive solutions to the above problem for small $\varepsilon>0$. By using various local Pohozaev identities and blow-up analysis, we first detect the relationship between the profile of the blow-up solutions and the Green's function of the domain $Ω$ and then obtain a type of local uniqueness results of blow-up solutions. At last we give a description of the number of positive solutions for small positive $\varepsilon$, which depends also on the Green's function.
We analyze a market impact game between $n$ risk averse agents who compete for liquidity in a market impact model with permanent price impact and additional slippage. Most market parameters, including volatility and drift, are allowed to vary stochastically. Our first main result characterizes the Nash equilibrium in terms of a fully coupled system of forward-backward stochastic differential equations (FBSDEs). Our second main result provides conditions under which this system of FBSDEs has indeed a unique solution, which in turn yields the unique Nash equilibrium. We furthermore obtain closed-form solutions in special situations and analyze them numerically
In the present paper, we establish the uniqueness and nondegeneracy of positive energy solutions to the Kirchhoff equation \begin{eqnarray*} -\left(a+b\int_{\mathbb{R}^{3}}|\nabla u|^{2}\right)Δu+u=|u|^{p-1}u & & \text{in }\mathbb{R}^{3}, \end{eqnarray*} where $a,b>0$, $1<p<5$ are constants. Then, as applications, we derive the existence and local uniqueness of solutions to the perturbed Kirchhoff problem \begin{eqnarray*} -\left(ε^2a+εb\int_{\mathbb{R}^{3}}|\nabla u|^{2}\right)Δu+V(x)u=|u|^{p-1}u & & \text{in }\mathbb{R}^{3} \end{eqnarray*} for $ε>0$ sufficiently small, under some mild assumptions on the potential function $V:\mathbb{R}^3\to \mathbb{R}$. The existence result is obtained by applying the Lyapunov-Schmidt reduction method. It seems to be the first time to study singularly perturbed Kirchhoff problems by reduction method, as all the previous results were obtained by various variational methods. Another advantage of this approach is that it gives a unified proof to the perturbation problem for all $p\in (1,5)$, which is quite different from using variational methods in the literature. The local uniqueness result is totally new. It is obtained by using a type of local Pohozaev identity, which is developed quite recently by Deng, Lin and Yan in their work "On the prescribed scalar curvature problem in $\mathbb{R}^N$, local uniqueness and periodicity." (see J. Math. Pures Appl. (9) {\bf 104}(2015), 1013-1044).