Recent years have witnessed the fast growth in telecommunication (Telco) techniques from 2G to upcoming 5G. Precise outdoor localization is important for Telco operators to manage, operate and optimize Telco networks. Differing from GPS, Telco locali
zation is a technique employed by Telco operators to localize outdoor mobile devices by using measurement report (MR) data. When given MR samples containing noisy signals (e.g., caused by Telco signal interference and attenuation), Telco localization often suffers from high errors. To this end, the main focus of this paper is how to improve Telco localization accuracy via the algorithms to detect and repair outlier positions with high errors. Specifically, we propose a context-aware Telco localization technique, namely RLoc, which consists of three main components: a machine-learning-based localization algorithm, a detection algorithm to find flawed samples, and a repair algorithm to replace outlier localization results by better ones (ideally ground truth positions). Unlike most existing works to detect and repair every flawed MR sample independently, we instead take into account spatio-temporal locality of MR locations and exploit trajectory context to detect and repair flawed positions. Our experiments on the real MR data sets from 2G GSM and 4G LTE Telco networks verify that our work RLoc can greatly improve Telco location accuracy. For example, RLoc on a large 4G MR data set can achieve 32.2 meters of median errors, around 17.4% better than state-of-the-art.
Recent years have witnessed unprecedented amounts of data generated by telecommunication (Telco) cellular networks. For example, measurement records (MRs) are generated to report the connection states between mobile devices and Telco networks, e.g.,
received signal strength. MR data have been widely used to localize outdoor mobile devices for human mobility analysis, urban planning, and traffic forecasting. Existing works using first-order sequence models such as the Hidden Markov Model (HMM) attempt to capture spatio-temporal locality in underlying mobility patterns for lower localization errors. The HMM approaches typically assume stable mobility patterns of the underlying mobile devices. Yet real MR datasets exhibit heterogeneous mobility patterns due to mixed transportation modes of the underlying mobile devices and uneven distribution of the positions associated with MR samples. Thus, the existing solutions cannot handle these heterogeneous mobility patterns. we propose a multi-task learning-based deep neural network (DNN) framework, namely PRNet+, to incorporate outdoor position recovery and transportation mode detection. To make sure the framework work, PRNet+ develops a feature extraction module to precisely learn local-, short- and long-term spatio-temporal locality from heterogeneous MR samples. Extensive evaluation on eight datasets collected at three representative areas in Shanghai indicates that PRNet+ greatly outperforms state-of-the-arts.
As the precision frontier of astrophysics advances towards the one millimagnitude level, flux calibration of photometric instrumentation remains an ongoing challenge. We present the results of a lab-bench assessment of the viability of monocrystallin
e silicon solar cells to serve as large-aperture (up to 125mm diameter), high-precision photodetectors. We measure the electrical properties, spatial response uniformity, quantum efficiency (QE), and frequency response of 3$^{rd}$ generation C60 solar cells, manufactured by Sunpower. Our new results, combined with our previous study of these cells linearity, dark current, and noise characteristics, suggest that these devices hold considerable promise, with QE and linearity that rival those of traditional, small-aperture photodiodes. We argue that any photocalibration project that relies on precise knowledge of the intensity of a large-diameter optical beam should consider using solar cells as calibrating photodetectors.
Architecture performance predictors have been widely used in neural architecture search (NAS). Although they are shown to be simple and effective, the optimization objectives in previous arts (e.g., precise accuracy estimation or perfect ranking of a
ll architectures in the space) did not capture the ranking nature of NAS. In addition, a large number of ground-truth architecture-accuracy pairs are usually required to build a reliable predictor, making the process too computationally expensive. To overcome these, in this paper, we look at NAS from a novel point of view and introduce Learning to Rank (LTR) methods to select the best (ace) architectures from a space. Specifically, we propose to use Normalized Discounted Cumulative Gain (NDCG) as the target metric and LambdaRank as the training algorithm. We also propose to leverage weak supervision from weight sharing by pretraining architecture representation on weak labels obtained from the super-net and then finetuning the ranking model using a small number of architectures trained from scratch. Extensive experiments on NAS benchmarks and large-scale search spaces demonstrate that our approach outperforms SOTA with a significantly reduced search cost.
Textures contain a wealth of image information and are widely used in various fields such as computer graphics and computer vision. With the development of machine learning, the texture synthesis and generation have been greatly improved. As a very c
ommon element in everyday life, wallpapers contain a wealth of texture information, making it difficult to annotate with a simple single label. Moreover, wallpaper designers spend significant time to create different styles of wallpaper. For this purpose, this paper proposes to describe wallpaper texture images by using multi-label semantics. Based on these labels and generative adversarial networks, we present a framework for perception driven wallpaper texture generation and style transfer. In this framework, a perceptual model is trained to recognize whether the wallpapers produced by the generator network are sufficiently realistic and have the attribute designated by given perceptual description; these multi-label semantic attributes are treated as condition variables to generate wallpaper images. The generated wallpaper images can be converted to those with well-known artist styles using CycleGAN. Finally, using the aesthetic evaluation method, the generated wallpaper images are quantitatively measured. The experimental results demonstrate that the proposed method can generate wallpaper textures conforming to human aesthetics and have artistic characteristics.
A significant abundance of primordial black hole (PBH) dark matter can be produced by curvature perturbations with power spectrum $Delta_zeta^2(k_{mathrm{peak}})sim mathcal{O}(10^{-2})$ at small scales, associated with the generation of observable sc
alar induced gravitational waves (SIGWs). However, the primordial non-Gaussianity may play a non-negligible role, which is not usually considered. We propose two inflation models that predict double peaks of order $mathcal{O}(10^{-2})$ in the power spectrum and study the effects of primordial non-Gaussianity on PBHs and SIGWs. This model is driven by a power-law potential, and has a noncanonical kinetic term whose coupling function admits two peaks. By field-redefinition, it can be recast into a canonical inflation model with two quasi-inflection points in the potential. We find that the PBH abundance will be altered saliently if non-Gaussianity parameter satisfies $|f_{mathrm{NL}}(k_{text{peak}},k_{text{peak}},k_{text{peak}})|gtrsim Delta^2_{zeta}(k_{mathrm{peak}})/(23delta^3_c) sim mathcal{O}(10^{-2})$. Whether the PBH abundance is suppressed or enhanced depends on the $f_{mathrm{NL}}$ being positive or negative, respectively. In our model, non-Gaussianity parameter $f_{mathrm{NL}}(k_{mathrm{peak}},k_{mathrm{peak}},k_{mathrm{peak}})sim mathcal{O}(1)$ takes positive sign, thus PBH abundance is suppressed dramatically. On the contrary, SIGWs are insensitive to primordial non-Gaussianity and hardly affected, so they are still within the sensitivities of space-based GWs observatories and Square Kilometer Array.
In the context of lossy compression, Blau & Michaeli (2019) adopt a mathematical notion of perceptual quality and define the information rate-distortion-perception function, generalizing the classical rate-distortion tradeoff. We consider the notion
of universal representations in which one may fix an encoder and vary the decoder to achieve any point within a collection of distortion and perception constraints. We prove that the corresponding information-theoretic universal rate-distortion-perception function is operationally achievable in an approximate sense. Under MSE distortion, we show that the entire distortion-perception tradeoff of a Gaussian source can be achieved by a single encoder of the same rate asymptotically. We then characterize the achievable distortion-perception region for a fixed representation in the case of arbitrary distributions, identify conditions under which the aforementioned results continue to hold approximately, and study the case when the rate is not fixed in advance. This motivates the study of practical constructions that are approximately universal across the RDP tradeoff, thereby alleviating the need to design a new encoder for each objective. We provide experimental results on MNIST and SVHN suggesting that on image compression tasks, the operational tradeoffs achieved by machine learning models with a fixed encoder suffer only a small penalty when compared to their variable encoder counterparts.
We present StrobeNet, a method for category-level 3D reconstruction of articulating objects from one or more unposed RGB images. Reconstructing general articulating object categories % has important applications, but is challenging since objects can
have wide variation in shape, articulation, appearance and topology. We address this by building on the idea of category-level articulation canonicalization -- mapping observations to a canonical articulation which enables correspondence-free multiview aggregation. Our end-to-end trainable neural network estimates feature-enriched canonical 3D point clouds, articulation joints, and part segmentation from one or more unposed images of an object. These intermediate estimates are used to generate a final implicit 3D reconstruction.Our approach reconstructs objects even when they are observed in different articulations in images with large baselines, and animation of reconstructed shapes. Quantitative and qualitative evaluations on different object categories show that our method is able to achieve high reconstruction accuracy, especially as more views are added.
Existing remote sensing change detection methods are heavily affected by seasonal variation. Since vegetation colors are different between winter and summer, such variations are inclined to be falsely detected as changes. In this letter, we proposed
an image translation method to solve the problem. A style-based recalibration module is introduced to capture seasonal features effectively. Then, a new style discriminator is designed to improve the translation performance. The discriminator can not only produce a decision for the fake or real sample, but also return a style vector according to the channel-wise correlations. Extensive experiments are conducted on season-varying dataset. The experimental results show that the proposed method can effectively perform image translation, thereby consistently improving the season-varying image change detection performance. Our codes and data are available at https://github.com/summitgao/RSIT_SRM_ISD.
Expressivity plays a fundamental role in evaluating deep neural networks, and it is closely related to understanding the limit of performance improvement. In this paper, we propose a three-pipeline training framework based on critical expressivity, i
ncluding global model contraction, weight evolution, and links weight rewiring. Specifically, we propose a pyramidal-like skeleton to overcome the saddle points that affect information transfer. Then we analyze the reason for the modularity (clustering) phenomenon in network topology and use it to rewire potential erroneous weighted links. We conduct numerical experiments on node classification and the results confirm that the proposed training framework leads to a significantly improved performance in terms of fast convergence and robustness to potential erroneous weighted links. The architecture design on GNNs, in turn, verifies the expressivity of GNNs from dynamics and topological space aspects and provides useful guidelines in designing more efficient neural networks.
