arXiv subjects

Cheng Cheng

Publications and source records attributed to Cheng Cheng.

At least 163 records · Page 9Linked to original sources

arXiv2022

Graph Fourier transforms on directed product graphs

Graph Fourier transform (GFT) is one of the fundamental tools in graph signal processing to decompose graph signals into different frequency components and to represent graph signals with strong correlation by different modes of variation effectively. The GFT on undirected graphs has been well studied and several approaches have been proposed to define GFTs on directed graphs. In this paper, based on the singular value decompositions of some graph Laplacians, we propose two GFTs on the Cartesian product graph of two directed graphs. We show that the proposed GFTs could represent spatial-temporal data sets on directed networks with strong correlation efficiently, and in the undirected graph setting they are essentially the joint GFT in the literature. In this paper, we also consider the bandlimiting procedure in the spectral domain of the proposed GFTs, and demonstrate its performance to denoise the temperature data set in the region of Brest (France) on January 2014.

cs.IT↗

arXiv2022

Likelihood-free Inference with Mixture Density Network

In this work, we propose using the mixture density network (MDN) to estimate cosmological parameters. We test the MDN method by constraining parameters of the $Λ$CDM and $w$CDM models using Type Ia supernovae and the power spectra of the cosmic microwave background. We find that the MDN method can achieve the same level of accuracy as the Markov Chain Monte Carlo method, with a slight difference of $\mathcal{O}(10^{-2}σ)$. Furthermore, the MDN method can provide accurate parameter estimates with $\mathcal{O}(10^3)$ forward simulation samples, which are useful for complex and resource-consuming cosmological models. This method can process either one data set or multiple data sets to achieve joint constraints on parameters, extendable for any parameter estimation of complicated models in a wider scientific field. Thus, the MDN provides an alternative way for likelihood-free inference of parameters.

astro-ph.CO↗

arXiv2022

A General End-to-end Diagnosis Framework for Manufacturing Systems

The manufacturing sector is envisioned to be heavily influenced by artificial intelligence-based technologies with the extraordinary increases in computational power and data volumes. A central challenge in manufacturing sector lies in the requirement of a general framework to ensure satisfied diagnosis and monitoring performances in different manufacturing applications. Here we propose a general data-driven, end-to-end framework for the monitoring of manufacturing systems. This framework, derived from deep learning techniques, evaluates fused sensory measurements to detect and even predict faults and wearing conditions. This work exploits the predictive power of deep learning to automatically extract hidden degradation features from noisy, time-course data. We have experimented the proposed framework on ten representative datasets drawn from a wide variety of manufacturing applications. Results reveal that the framework performs well in examined benchmark applications and can be applied in diverse contexts, indicating its potential use as a critical corner stone in smart manufacturing.

cs.LG↗

arXiv2022

A deep learning-based remaining useful life prediction approach for bearings

In industrial applications, nearly half the failures of motors are caused by the degradation of rolling element bearings (REBs). Therefore, accurately estimating the remaining useful life (RUL) for REBs are of crucial importance to ensure the reliability and safety of mechanical systems. To tackle this challenge, model-based approaches are often limited by the complexity of mathematical modeling. Conventional data-driven approaches, on the other hand, require massive efforts to extract the degradation features and construct health index. In this paper, a novel online data-driven framework is proposed to exploit the adoption of deep convolutional neural networks (CNN) in predicting the RUL of bearings. More concretely, the raw vibrations of training bearings are first processed using the Hilbert-Huang transform (HHT) and a novel nonlinear degradation indicator is constructed as the label for learning. The CNN is then employed to identify the hidden pattern between the extracted degradation indicator and the vibration of training bearings, which makes it possible to estimate the degradation of the test bearings automatically. Finally, testing bearings' RULs are predicted by using a $ε$-support vector regression model. The superior performance of the proposed RUL estimation framework, compared with the state-of-the-art approaches, is demonstrated through the experimental results. The generality of the proposed CNN model is also validated by transferring to bearings undergoing different operating conditions.

cs.LG↗

arXiv2022

Learning Clinical Concepts for Predicting Risk of Progression to Severe COVID-19

With COVID-19 now pervasive, identification of high-risk individuals is crucial. Using data from a major healthcare provider in Southwestern Pennsylvania, we develop survival models predicting severe COVID-19 progression. In this endeavor, we face a tradeoff between more accurate models relying on many features and less accurate models relying on a few features aligned with clinician intuition. Complicating matters, many EHR features tend to be under-coded, degrading the accuracy of smaller models. In this study, we develop two sets of high-performance risk scores: (i) an unconstrained model built from all available features; and (ii) a pipeline that learns a small set of clinical concepts before training a risk predictor. Learned concepts boost performance over the corresponding features (C-index 0.858 vs. 0.844) and demonstrate improvements over (i) when evaluated out-of-sample (subsequent time periods). Our models outperform previous works (C-index 0.844-0.872 vs. 0.598-0.810).

cs.LG↗

arXiv2022

Properties of host galaxies of submillimeter sources as revealed by JWST Early Release Observations in SMACS J0723.3-7327

Using the 0.9--4.4~$μ$m imaging data from the James Webb Space Telescope (JWST) Early Release Observation in the SMACS J0723.3-7327 galaxy cluster field, we discuss the properties of three submillimeter galaxies (SMGs) detected by the Atacama Large Millimeter Array. These sources are magnified by 1.4--2.1$\times$due to gravitational lensing. This is the first time that SMG host galaxies are resolved in the rest-frame near-infrared (NIR). One source was previous undetected by HST, while the remaining two are disk galaxies with Sérsic indices of $\sim 0.9$ and star formation rates on or just below the star formation "main sequence". Their submillimeter emission originates from the inner parts of the hosts, suggesting that their dust contents are concentrated towards the center. The host half-light radii measured in the rest-frame NIR are $\sim$1.5$\times$ smaller than those measured in the rest-frame optical, consistent with a concentrated dust distribution. The more severe extinction that optical light suffers towards the center makes it seemingly less concentrated. Therefore, we expect that the optically-based determination of the stellar mass distribution within host galaxies could still be severely biased by dust. Interestingly, these two disk galaxies are dramatically different in their outer regions, with one being star forming and the other being quiescent. Upcoming JWST observations of statistically significant samples of SMGs will allow us to understand the correlation between the dusty star forming regions and their hosts.

astro-ph.GA↗

arXiv2022

Streamwise inclination angle of wall-attached eddies in turbulent channel flows

We develop a new methodology to assess the streamwise inclination angles (SIAs) of the wall-attached eddies populating the logarithmic region with a given wall-normal height. To remove the influences originating from other scales on the SIA estimated via two-point correlation, the footprints of the targeted eddies in the vicinity of the wall and the corresponding streamwise velocity fluctuations carried by them are isolated simultaneously, by coupling the spectral stochastic estimation with the attached-eddy hypothesis. Datasets produced with direct numerical simulations spanning $Re_τ \sim O(10^2)-O(10^3)$ are dissected to study the Reynolds-number effect. The present results show, for the first time, that the SIAs of attached eddies are Reynolds-number dependent in low and medium Reynolds numbers and tend to saturate at $45^{\circ}$ as the Reynolds number increases. The mean SIA reported by vast previous experimental studies are demonstrated to be the outcomes of the additive effect contributed by multi-scale attached eddies. These findings clarify the long-term debate and perfect the picture of the attached-eddy model.

physics.flu-dyn↗

arXiv2022

FastAdaBelief: Improving Convergence Rate for Belief-based Adaptive Optimizers by Exploiting Strong Convexity

AdaBelief, one of the current best optimizers, demonstrates superior generalization ability compared to the popular Adam algorithm by viewing the exponential moving average of observed gradients. AdaBelief is theoretically appealing in that it has a data-dependent $O(\sqrt{T})$ regret bound when objective functions are convex, where $T$ is a time horizon. It remains however an open problem whether the convergence rate can be further improved without sacrificing its generalization ability. %on how to exploit strong convexity to further improve the convergence rate of AdaBelief. To this end, we make a first attempt in this work and design a novel optimization algorithm called FastAdaBelief that aims to exploit its strong convexity in order to achieve an even faster convergence rate. In particular, by adjusting the step size that better considers strong convexity and prevents fluctuation, our proposed FastAdaBelief demonstrates excellent generalization ability as well as superior convergence. As an important theoretical contribution, we prove that FastAdaBelief attains a data-dependant $O(\log T)$ regret bound, which is substantially lower than AdaBelief. On the empirical side, we validate our theoretical analysis with extensive experiments in both scenarios of strong and non-strong convexity on three popular baseline models. Experimental results are very encouraging: FastAdaBelief converges the quickest in comparison to all mainstream algorithms while maintaining an excellent generalization ability, in cases of both strong or non-strong convexity. FastAdaBelief is thus posited as a new benchmark model for the research community.

cs.LG↗

arXiv2022

Consistency between the attached-eddy model and the inner-outer interaction model: a study of streamwise wall-shear stress fluctuations in a turbulent channel flow

The inner-outer interaction model (Marusic, Mathis & Hutchins, Science, vol. 329, 2010, 193-196) and the attached-eddy model (Townsend, Cambridge University Press, 1976) are two fundamental models describing the multi-scale turbulence interactions and the organization of energy-containing motions in the logarithmic region of high-Reynolds number wall-bounded turbulence, respectively. In this paper, by coupling the additive description with the attached-eddy model, the generation process of streamwise wall-shear fluctuations, resulting from wall-attached eddies, is portrayed. Then, by resorting to the inner-outer interaction model, the streamwise wall-shear stress fluctuations generated by attached eddies in a turbulent channel flow are isolated. Direct comparison between the statistics from these two models demonstrates that they are consistent to and complement each other. Meanwhile, we further show that the superpositions of attached eddies follow an additive process strictly by verifying the validity of the strong and extended self similarity. Moreover, we propose a Gaussian model to characterize the instantaneous distribution of streamwise wall-shear stress, resulting from the attached-eddy superpositions. These findings are important for developing an advanced reduced-order wall model.

physics.flu-dyn↗

arXiv2022

Graph Fourier transform based on singular value decomposition of directed Laplacian

Graph Fourier transform (GFT) is a fundamental concept in graph signal processing. In this paper, based on singular value decomposition of Laplacian, we introduce a novel definition of GFT on directed graphs, and use singular values of Laplacian to carry the notion of graph frequencies. % of the proposed GFT. The proposed GFT is consistent with the conventional GFT in the undirected graph setting, and on directed circulant graphs, the proposed GFT is the classical discrete Fourier transform, up to some rotation, permutation and phase adjustment. We show that frequencies and frequency components of the proposed GFT can be evaluated by solving some constrained minimization problems with low computational cost. Numerical demonstrations indicate that the proposed GFT could represent graph signals with different modes of variation efficiently.

eess.SP↗

arXiv2022

Wiener filters on graphs and distributed polynomial approximation algorithms

In this paper, we consider Wiener filters to reconstruct deterministic and (wide-band) stationary graph signals from their observations corrupted by random noises, and we propose distributed algorithms to implement Wiener filters and inverse filters on networks in which agents are equipped with a data processing subsystem for limited data storage and computation power, and with a one-hop communication subsystem for direct data exchange only with their adjacent agents. The proposed distributed polynomial approximation algorithm is an exponential convergent quasi-Newton method based on Jacobi polynomial approximation and Chebyshev interpolation polynomial approximation to analytic functions on a cube. Our numerical simulations show that Wiener filtering procedure performs better on denoising (wide-band) stationary signals than the Tikhonov regularization approach does, and that the proposed polynomial approximation algorithms converge faster than the Chebyshev polynomial approximation algorithm and gradient decent algorithm do in the implementation of an inverse filtering procedure associated with a polynomial filter of commutative graph shifts.

eess.SP↗

arXiv2022

OneFlow: Redesign the Distributed Deep Learning Framework from Scratch

Deep learning frameworks such as TensorFlow and PyTorch provide a productive interface for expressing and training a deep neural network (DNN) model on a single device or using data parallelism. Still, they may not be flexible or efficient enough in training emerging large models on distributed devices, which require more sophisticated parallelism beyond data parallelism. Plugins or wrappers have been developed to strengthen these frameworks for model or pipeline parallelism, but they complicate the usage and implementation of distributed deep learning. Aiming at a simple, neat redesign of distributed deep learning frameworks for various parallelism paradigms, we present OneFlow, a novel distributed training framework based on an SBP (split, broadcast and partial-value) abstraction and the actor model. SBP enables much easier programming of data parallelism and model parallelism than existing frameworks, and the actor model provides a succinct runtime mechanism to manage the complex dependencies imposed by resource constraints, data movement and computation in distributed deep learning. We demonstrate the general applicability and efficiency of OneFlow for training various large DNN models with case studies and extensive experiments. The results show that OneFlow outperforms many well-known customized libraries built on top of the state-of-the-art frameworks. The code of OneFlow is available at: https://github.com/Oneflow-Inc/oneflow.

cs.DC↗

arXiv2022

Snowmass 2021 CMB-S4 White Paper

This Snowmass 2021 White Paper describes the Cosmic Microwave Background Stage 4 project CMB-S4, which is designed to cross critical thresholds in our understanding of the origin and evolution of the Universe, from the highest energies at the dawn of time through the growth of structure to the present day. We provide an overview of the science case, the technical design, and project plan.

astro-ph.CO↗

arXiv2022

MetaBalance: Improving Multi-Task Recommendations via Adapting Gradient Magnitudes of Auxiliary Tasks

In many personalized recommendation scenarios, the generalization ability of a target task can be improved via learning with additional auxiliary tasks alongside this target task on a multi-task network. However, this method often suffers from a serious optimization imbalance problem. On the one hand, one or more auxiliary tasks might have a larger influence than the target task and even dominate the network weights, resulting in worse recommendation accuracy for the target task. On the other hand, the influence of one or more auxiliary tasks might be too weak to assist the target task. More challenging is that this imbalance dynamically changes throughout the training process and varies across the parts of the same network. We propose a new method: MetaBalance to balance auxiliary losses via directly manipulating their gradients w.r.t the shared parameters in the multi-task network. Specifically, in each training iteration and adaptively for each part of the network, the gradient of an auxiliary loss is carefully reduced or enlarged to have a closer magnitude to the gradient of the target loss, preventing auxiliary tasks from being so strong that dominate the target task or too weak to help the target task. Moreover, the proximity between the gradient magnitudes can be flexibly adjusted to adapt MetaBalance to different scenarios. The experiments show that our proposed method achieves a significant improvement of 8.34% in terms of NDCG@10 upon the strongest baseline on two real-world datasets. The code of our approach can be found at here: https://github.com/facebookresearch/MetaBalance

cs.LG↗

arXiv2022

Learning to Adapt to Light

Light adaptation or brightness correction is a key step in improving the contrast and visual appeal of an image. There are multiple light-related tasks (for example, low-light enhancement and exposure correction) and previous studies have mainly investigated these tasks individually. However, it is interesting to consider whether these light-related tasks can be executed by a unified model, especially considering that our visual system adapts to external light in such way. In this study, we propose a biologically inspired method to handle light-related image-enhancement tasks with a unified network (called LA-Net). First, a frequency-based decomposition module is designed to decouple the common and characteristic sub-problems of light-related tasks into two pathways. Then, a new module is built inspired by biological visual adaptation to achieve unified light adaptation in the low-frequency pathway. In addition, noise suppression or detail enhancement is achieved effectively in the high-frequency pathway regardless of the light levels. Extensive experiments on three tasks -- low-light enhancement, exposure correction, and tone mapping -- demonstrate that the proposed method almost obtains state-of-the-art performance compared with recent methods designed for these individual tasks.

cs.CV↗

arXiv2021

Simulation-Based Inference of Reionization Parameters From 3D Tomographic 21 cm Lightcone Images

Tomographic three-dimensional 21 cm images from the epoch of reionization contain a wealth of information about the reionization of the intergalactic medium by astrophysical sources. Conventional power spectrum analysis cannot exploit the full information in the 21 cm data because the 21 cm signal is highly non-Gaussian due to reionization patchiness. We perform a Bayesian inference of the reionization parameters where the likelihood is implicitly defined through forward simulations using density estimation likelihood-free inference (DELFI). We adopt a trained 3D Convolutional Neural Network (CNN) to compress the 3D image data into informative summaries (DELFI-3D CNN). We show that this method recovers accurate posterior distributions for the reionization parameters. Our approach outperforms earlier analysis based on two-dimensional 21 cm images. In contrast, an MCMC analysis of the 3D lightcone-based 21 cm power spectrum alone and using a standard explicit likelihood approximation results in less accurate credible parameter regions than inferred by the DELFI-3D CNN, both in terms of the location and shape of the contours. Our proof-of-concept study implies that the DELFI-3D CNN can effectively exploit more information in the 3D 21 cm images than a 2D CNN or power spectrum analysis. This technique can be readily extended to include realistic effects and is therefore a promising approach for the scientific interpretation of future 21 cm observation data.

astro-ph.CO↗

arXiv2021

A recurrent neural network approach for remaining useful life prediction utilizing a novel trend features construction method

Data-driven methods for remaining useful life (RUL) prediction normally learn features from a fixed window size of a priori of degradation, which may lead to less accurate prediction results on different datasets because of the variance of local features. This paper proposes a method for RUL prediction which depends on a trend feature representing the overall time sequence of degradation. Complete ensemble empirical mode decomposition, followed by a reconstruction procedure, is created to build the trend features. The probability distribution of sensors' measurement learned by conditional neural processes is used to evaluate the trend features. With the best trend feature, a data-driven model using long short-term memory is developed to predict the RUL. To prove the effectiveness of the proposed method, experiments on a benchmark C-MAPSS dataset are carried out and compared with other state-of-the-art methods. Comparison results show that the proposed method achieves the smallest root mean square values in prediction of all RUL.

eess.SP↗

arXiv2021

The In-Flight Realtime Trigger and Localization Software of GECAM

Realtime trigger and localization of bursts are the key functions of GECAM, which is an all-sky gamma-ray monitor launched in Dec 10, 2020. We developed a multifunctional trigger and localization software operating on the CPU of the GECAM electronic box (EBOX). This onboard software has the following features: high trigger efficiency for real celestial bursts with a suppression of false triggers caused by charged particle bursts and background fluctuation, dedicated localization algorithm optimized for short and long bursts respetively, short time latency of the trigger information which is downlinked throught the BeiDou satellite navigation System (BDS). This paper presents the detailed design and deveopment of this trigger and localization software system of GECAM, including the main functions, general design, workflow and algorithms, as well as the verification and demonstration of this software, including the on-ground trigger tests with simulated gamma-ray bursts made by a dedicated X-ray tube and the in-flight performance to real gamma-ray bursts and magnetar bursts.

astro-ph.IM↗