arXiv subjects

Ling Li

Publications and source records attributed to Ling Li.

At least 109 records · Page 6Linked to original sources

arXiv2022

SaiT: Sparse Vision Transformers through Adaptive Token Pruning

While vision transformers have achieved impressive results, effectively and efficiently accelerating these models can further boost performances. In this work, we propose a dense/sparse training framework to obtain a unified model, enabling weight sharing across various token densities. Thus one model offers a range of accuracy and throughput tradeoffs for different applications. Besides, we introduce adaptive token pruning to optimize the patch token sparsity based on the input image. In addition, we investigate knowledge distillation to enhance token selection capability in early transformer modules. Sparse adaptive image Transformer (SaiT) offers varying levels of model acceleration by merely changing the token sparsity on the fly. Specifically, SaiT reduces the computation complexity (FLOPs) by 39% - 43% and increases the throughput by 67% - 91% with less than 0.5% accuracy loss for various vision transformer models. Meanwhile, the same model also provides the zero accuracy drop option by skipping the sparsification step. SaiT achieves better accuracy and computation tradeoffs than state-of-the-art transformer and convolutional models.

cs.CV↗

arXiv2022

MaiT: Leverage Attention Masks for More Efficient Image Transformers

Though image transformers have shown competitive results with convolutional neural networks in computer vision tasks, lacking inductive biases such as locality still poses problems in terms of model efficiency especially for embedded applications. In this work, we address this issue by introducing attention masks to incorporate spatial locality into self-attention heads. Local dependencies are captured efficiently with masked attention heads along with global dependencies captured by unmasked attention heads. With Masked attention image Transformer - MaiT, top-1 accuracy increases by up to 1.7% compared to CaiT with fewer parameters and FLOPs, and the throughput improves by up to 1.5X compared to Swin. Encoding locality with attention masks is model agnostic, and thus it applies to monolithic, hierarchical, or other novel transformer architectures.

cs.CV↗

arXiv2022

Coupled Metaball Discrete Element Lattice Boltzmann Method for Fluid-Particle Systems with non-spherical particle shapes: A sharp interface coupling scheme

Fluid-particle systems are very common in many natural processes and engineering applications. However, accurately and efficiently modelling fluid-particle systems with complex particle shapes is still a challenging task. Here, we present a numerical model that combines the advantages of Lattice Boltzmann Method (LBM) in solving complex flow problems and the capability of the recently introduced Metaball Discrete Element Method (MDEM) in handling non-spherical particle shapes. A sharp interface coupling scheme is developed and the numerical instability issues due to the discontinuity of interfaces are carefully addressed. A local refilling algorithm for new fluid nodes is proposed and special treatments are introduced to reduce numerical noises when two particles are close. The proposed model is validated by simulations of settling of a single sphere (with metaball representation) as well as a non-spherical particle in a viscous fluid. Good agreements are found comparing the simulations with experimental results which are also carried out in this study. The coupling scheme is also tested by multiple particle simulations which clearly illustrated the stability of the proposed model. Finally, numerical examples with complex particle shapes demonstrated that the proposed model can be a powerful tool for future applications such as shape-induced segregation in riverbeds, and phase transition of dense suspension.

physics.flu-dyn↗

arXiv2022

Automatic Facial Skin Feature Detection for Everyone

Automatic assessment and understanding of facial skin condition have several applications, including the early detection of underlying health problems, lifestyle and dietary treatment, skin-care product recommendation, etc. Selfies in the wild serve as an excellent data resource to democratize skin quality assessment, but suffer from several data collection challenges.The key to guaranteeing an accurate assessment is accurate detection of different skin features. We present an automatic facial skin feature detection method that works across a variety of skin tones and age groups for selfies in the wild. To be specific, we annotate the locations of acne, pigmentation, and wrinkle for selfie images with different skin tone colors, severity levels, and lighting conditions. The annotation is conducted in a two-phase scheme with the help of a dermatologist to train volunteers for annotation. We employ Unet++ as the network architecture for feature detection. This work shows that the two-phase annotation scheme can robustly detect the accurate locations of acne, pigmentation, and wrinkle for selfie images with different ethnicities, skin tone colors, severity levels, age groups, and lighting conditions.

cs.CV↗

arXiv2022

Cost-sensitive Boosting Pruning Trees for depression detection on Twitter

Depression is one of the most common mental health disorders, and a large number of depressed people commit suicide each year. Potential depression sufferers usually do not consult psychological doctors because they feel ashamed or are unaware of any depression, which may result in severe delay of diagnosis and treatment. In the meantime, evidence shows that social media data provides valuable clues about physical and mental health conditions. In this paper, we argue that it is feasible to identify depression at an early stage by mining online social behaviours. Our approach, which is innovative to the practice of depression detection, does not rely on the extraction of numerous or complicated features to achieve accurate depression detection. Instead, we propose a novel classifier, namely, Cost-sensitive Boosting Pruning Trees (CBPT), which demonstrates a strong classification ability on two publicly accessible Twitter depression detection datasets. To comprehensively evaluate the classification capability of the CBPT, we use additional three datasets from the UCI machine learning repository and the CBPT obtains appealing classification results against several state of the arts boosting algorithms. Finally, we comprehensively explore the influence factors of model prediction, and the results manifest that our proposed framework is promising for identifying Twitter users with depression.

cs.LG↗

arXiv2021

Ultra-Broadband Dispersion-Manipulated Dielectric Metalenses by Nonlinear Dispersive Phase Compensation

Dispersion decomposes compound light into monochromatic components at different spatial locations, which needs to be eliminated in imaging but utilized in spectral detection. Metasurfaces provide a unique path to modulate the dispersion only by adjusting the structural parameters without changing the material as required for refractive elements. However, the common linear phase compensation does not conform to the dispersion characteristics of the meta-unit limiting dispersion modulation in broader wavelength bands, which is desired for ultra-broadband or multiband imaging. Here, we propose a nonlinear dispersive phase compensation method to design polarization-insensitive achromatic metalenses from 400 nm to 1000 nm constructed with single-layer high aspect ratio nanostructures. This band matches the response spectrum of a typical CMOS sensor for both visible and near-infrared imaging applications without additional lens replacement. Moreover, the capability of the method in achieving arbitrary dispersion modulation is demonstrated for applications such as chromatography imaging and spectral detection.

physics.optics↗

arXiv2021

Semantically Contrastive Learning for Low-light Image Enhancement

Low-light image enhancement (LLE) remains challenging due to the unfavorable prevailing low-contrast and weak-visibility problems of single RGB images. In this paper, we respond to the intriguing learning-related question -- if leveraging both accessible unpaired over/underexposed images and high-level semantic guidance, can improve the performance of cutting-edge LLE models? Here, we propose an effective semantically contrastive learning paradigm for LLE (namely SCL-LLE). Beyond the existing LLE wisdom, it casts the image enhancement task as multi-task joint learning, where LLE is converted into three constraints of contrastive learning, semantic brightness consistency, and feature preservation for simultaneously ensuring the exposure, texture, and color consistency. SCL-LLE allows the LLE model to learn from unpaired positives (normal-light)/negatives (over/underexposed), and enables it to interact with the scene semantics to regularize the image enhancement network, yet the interaction of high-level semantic knowledge and the low-level signal prior is seldom investigated in previous methods. Training on readily available open data, extensive experiments demonstrate that our method surpasses the state-of-the-arts LLE models over six independent cross-scenes datasets. Moreover, SCL-LLE's potential to benefit the downstream semantic segmentation under extremely dark conditions is discussed. Source Code: https://github.com/LingLIx/SCL-LLE.

cs.CV↗

arXiv2021

Conditional Local Convolution for Spatio-temporal Meteorological Forecasting

Spatio-temporal forecasting is challenging attributing to the high nonlinearity in temporal dynamics as well as complex location-characterized patterns in spatial domains, especially in fields like weather forecasting. Graph convolutions are usually used for modeling the spatial dependency in meteorology to handle the irregular distribution of sensors' spatial location. In this work, a novel graph-based convolution for imitating the meteorological flows is proposed to capture the local spatial patterns. Based on the assumption of smoothness of location-characterized patterns, we propose conditional local convolution whose shared kernel on nodes' local space is approximated by feedforward networks, with local representations of coordinate obtained by horizon maps into cylindrical-tangent space as its input. The established united standard of local coordinate system preserves the orientation on geography. We further propose the distance and orientation scaling terms to reduce the impacts of irregular spatial distribution. The convolution is embedded in a Recurrent Neural Network architecture to model the temporal dynamics, leading to the Conditional Local Convolution Recurrent Network (CLCRN). Our model is evaluated on real-world weather benchmark datasets, achieving state-of-the-art performance with obvious improvements. We conduct further analysis on local pattern visualization, model's framework choice, advantages of horizon maps and etc.

cs.LG↗

arXiv2021

Griffin: Rethinking Sparse Optimization for Deep Learning Architectures

This paper examines the design space trade-offs of DNNs accelerators aiming to achieve competitive performance and efficiency metrics for all four combinations of dense or sparse activation/weight tensors. To do so, we systematically examine the overheads of supporting sparsity on top of an optimized dense core. These overheads are modeled based on parameters that indicate how a multiplier can borrow a nonzero operation from the neighboring multipliers or future cycles. As a result of this exploration, we identify a few promising designs that perform better than prior work. Our findings suggest that even the best design targeting dual sparsity yields a 20%-30% drop in power efficiency when performing on single sparse models, i.e., those with only sparse weight or sparse activation tensors. We found that one can reuse resources of the same core to maintain high performance and efficiency when running single sparsity or dense models. We call this hybrid architecture Griffin. Griffin is 1.2, 3.0, 3.1, and 1.4X more power-efficient than state-of-the-art sparse architectures, for dense, weight-only sparse, activation-only sparse, and dual sparse models, respectively.

cs.AR↗

arXiv2021

Bayesian data assimilation for estimating epidemic evolution: a COVID-19 study

The evolution of epidemiological parameters, such as instantaneous reproduction number Rt, is important for understanding the transmission dynamics of infectious diseases. Current estimates of time-varying epidemiological parameters often face problems such as lagging observations, averaging inference, and improper quantification of uncertainties. To address these problems, we propose a Bayesian data assimilation framework for time-varying parameter estimation. Specifically, this framework is applied to Rt estimation, resulting in the state-of-the-art DARt system. With DARt, time misalignment caused by lagging observations is tackled by incorporating observation delays into the joint inference of infections and Rt; the drawback of averaging is overcome by instantaneously updating upon new observations and developing a model selection mechanism that captures abrupt changes; the uncertainty is quantified and reduced by employing Bayesian smoothing. We validate the performance of DARt and demonstrate its power in revealing the transmission dynamics of COVID-19. The proposed approach provides a promising solution for accurate and timely estimating transmission dynamics from reported data.

stat.AP↗

arXiv2021

Finite-Size Analysis of the Collapse of Dry Granular Columns

In this letter, we focus on the size effect of granular column collapses, which are potentially connected to the dynamics of complex geophysical flows, even if the link between microscopic structures of granular assemblies and their macroscopic behaviors is still not well understood. Using the sphero-polyhedral discrete element method (DEM), we show that the column radius/grain size ratio has a strong influence on the collapse behavior. A finite-size analysis, which is inspired by a phase transition around an inflection point, is performed to obtain a general scaling equation with critical exponents for run-out distances. We further link the size effect with the strong force network and formalize a correlation length scale that exponentially scales with the effective aspect ratio. Such a scaling solution shows similarities with the percolation problem of two-dimensional random networks and can be extended to other similar natural and engineering systems.

cond-mat.soft↗

arXiv2021

Eden: A Unified Environment Framework for Booming Reinforcement Learning Algorithms

With AlphaGo defeats top human players, reinforcement learning(RL) algorithms have gradually become the code-base of building stronger artificial intelligence(AI). The RL algorithm design firstly needs to adapt to the specific environment, so the designed environment guides the rapid and profound development of RL algorithms. However, the existing environments, which can be divided into real world games and customized toy environments, have obvious shortcomings. For real world games, it is designed for human entertainment, and too much difficult for most of RL researchers. For customized toy environments, there is no widely accepted unified evaluation standard for all RL algorithms. Therefore, we introduce the first virtual user-friendly environment framework for RL. In this framework, the environment can be easily configured to realize all kinds of RL tasks in the mainstream research. Then all the mainstream state-of-the-art(SOTA) RL algorithms can be conveniently evaluated and compared. Therefore, our contributions mainly includes the following aspects: 1.single configured environment for all classification of SOTA RL algorithms; 2.combined environment of more than one classification RL algorithms; 3.the evaluation standard for all kinds of RL algorithms. With all these efforts, a possibility for breeding an AI with capability of general competency in a variety of tasks is provided, and maybe it will open up a new chapter for AI.

cs.LG↗

arXiv2021

Structured Context Enhancement Network for Mouse Pose Estimation

Automated analysis of mouse behaviours is crucial for many applications in neuroscience. However, quantifying mouse behaviours from videos or images remains a challenging problem, where pose estimation plays an important role in describing mouse behaviours. Although deep learning based methods have made promising advances in human pose estimation, they cannot be directly applied to pose estimation of mice due to different physiological natures. Particularly, since mouse body is highly deformable, it is a challenge to accurately locate different keypoints on the mouse body. In this paper, we propose a novel Hourglass network based model, namely Graphical Model based Structured Context Enhancement Network (GM-SCENet) where two effective modules, i.e., Structured Context Mixer (SCM) and Cascaded Multi-Level Supervision (CMLS) are subsequently implemented. SCM can adaptively learn and enhance the proposed structured context information of each mouse part by a novel graphical model that takes into account the motion difference between body parts. Then, the CMLS module is designed to jointly train the proposed SCM and the Hourglass network by generating multi-level information, increasing the robustness of the whole network.Using the multi-level prediction information from SCM and CMLS, we develop an inference method to ensure the accuracy of the localisation results. Finally, we evaluate our proposed approach against several baselines...

cs.CV↗

arXiv2021

Model structures and recollements induced by duality pairs

We give some equivalent characterizations of $\mathcal{GP}$, the class of Gorenstein $(\mathcal{L}, \mathcal{A})$-projective modules, and construct some model structures associated to duality pairs and Frobenius pairs. Moreover, some rings are described by Frobenius pairs. Meanwhile, we investigate strongly Gorenstein $(\mathcal{L}, \mathcal{A})$-projective modules and obtain some equivalent characterizations of them. Also, some model structures and recollements associated to strongly Gorenstein $(\mathcal{L}, \mathcal{A})$-projective modules are constructed.

math.RA↗

arXiv2021

Muti-view Mouse Social Behaviour Recognition with Deep Graphical Model

Home-cage social behaviour analysis of mice is an invaluable tool to assess therapeutic efficacy of neurodegenerative diseases. Despite tremendous efforts made within the research community, single-camera video recordings are mainly used for such analysis. Because of the potential to create rich descriptions of mouse social behaviors, the use of multi-view video recordings for rodent observations is increasingly receiving much attention. However, identifying social behaviours from various views is still challenging due to the lack of correspondence across data sources. To address this problem, we here propose a novel multiview latent-attention and dynamic discriminative model that jointly learns view-specific and view-shared sub-structures, where the former captures unique dynamics of each view whilst the latter encodes the interaction between the views. Furthermore, a novel multi-view latent-attention variational autoencoder model is introduced in learning the acquired features, enabling us to learn discriminative features in each view. Experimental results on the standard CRMI13 and our multi-view Parkinson's Disease Mouse Behaviour (PDMB) datasets demonstrate that our model outperforms the other state of the arts technologies and effectively deals with the imbalanced data problem.

cs.CV↗

arXiv2021

Deep learning based low-dose synchrotron radiation CT reconstruction

Synchrotron radiation sources are widely used in various fields, among which computed tomography (CT) is one of the most important. The amount of effort expended by the operator varies depending on the subject. If the number of angles needed to be used can be greatly reduced under the condition of similar imaging effects, the working time and workload of the experimentalists will be greatly reduced. However, decreasing the sampling angle can produce serious artifacts and blur the details. We try to use a deep learning model which can build high quality reconstruction sparse data sampling from the angle of the image and ResAttUnet are put forward. ResAttUnet is roughly a symmetrical U-shaped network that incorporates similar mechanisms to ResNet and attention. In addition, the mixed precision is adopted to reduce the demand for video memory of the model and training time.

eess.IV↗

arXiv2021

Deposition Morphology of Granular Column Collapses

Granular column collapses result in an array of flow phenomena and deposition morphologies, the understanding of which brings insights into studying granular flows in both natural and engineering systems. Guided by experiments, we carried out computational studies with the discrete element method (DEM) to identify fundamental links between the macro-scale behavior and micro-scale properties of granular columns. A physics-based non-dimensional number combining particle and bulk properties of the column, $α_{\textrm{eff}}$, was found to determinate three collapse regimes (quasi-static, inertial, and liquid-like), revealing universal trends of flow regimes and deposition morphologies under different conditions. This non-dimensional number represents physically the competing inertial and frictional effects that govern the behavior of the granular column collapse, including energy conversion and dissipation. The finding is important for understanding quantitatively the flow of granular materials and their deposits.

cond-mat.soft↗

arXiv2021

Slip flow regimes in nanofluidics: a universal superexponential model

Many experiments have shown large flow enhancement ratios (up to 10^5) in carbon nanotubes (CNT) with diameters larger than 5nm. However, molecular dynamics simulations have never replicated these results maintaining a three-order-of-magnitude gap with measurements. Our study provides a generic model of nanofluidics for continuum slip flow (diameter>3nm) that fills this significant gap and sheds light on its origin. Compared to 140 literature cases, the model explains the entire range of experimental flow enhancements by changes of nanotube diameters and finite variations of interfacial energies. Despite large variations of flow enhancement ratios spanning 5 orders of magnitude in experimental results, the ratio between these data and corresponding model predictions approaches unity for the majority of experiments. The role of viscous entrance effects is discussed. The model provides insight into puzzling observations such as differences of CNTs and boron nitride nanotubes, the slip on low-contact-angle surfaces and massive functionalization effects. This study could advance our understanding of nano-scale transport mechanisms and aid the design of tailored nanomembranes.

cond-mat.soft↗