arXiv subjects

Kun Qian

Publications and source records attributed to Kun Qian.

At least 109 records · Page 6Linked to original sources

arXiv2020

Multi-Scale Progressive Fusion Learning for Depth Map Super-Resolution

Limited by the cost and technology, the resolution of depth map collected by depth camera is often lower than that of its associated RGB camera. Although there have been many researches on RGB image super-resolution (SR), a major problem with depth map super-resolution is that there will be obvious jagged edges and excessive loss of details. To tackle these difficulties, in this work, we propose a multi-scale progressive fusion network for depth map SR, which possess an asymptotic structure to integrate hierarchical features in different domains. Given a low-resolution (LR) depth map and its associated high-resolution (HR) color image, We utilize two different branches to achieve multi-scale feature learning. Next, we propose a step-wise fusion strategy to restore the HR depth map. Finally, a multi-dimensional loss is introduced to constrain clear boundaries and details. Extensive experiments show that our proposed method produces improved results against state-of-the-art methods both qualitatively and quantitatively.

cs.CV↗

arXiv2020

Towards SAR Tomographic Inversion via Sparse Bayesian Learning

Existing SAR tomography (TomoSAR) algorithms are mostly based on an inversion of the SAR imaging model, which are often computationally expensive. Previous study showed perspective of using data-driven methods like KPCA to decompose the signal and reduce the computational complexity. This paper gives a preliminary demonstration of a new data-driven method based on sparse Bayesian learning. Experiments on simulated data show that the proposed method significantly outperforms KPCA methods in estimating the steering vectors of the scatterers. This gives a perspective of data-drive approach or combining it with model-driven approach for high precision tomographic inversion of large areas.

eess.SP↗

arXiv2020

Learning Structured Representations of Entity Names using Active Learning and Weak Supervision

Structured representations of entity names are useful for many entity-related tasks such as entity normalization and variant generation. Learning the implicit structured representations of entity names without context and external knowledge is particularly challenging. In this paper, we present a novel learning framework that combines active learning and weak supervision to solve this problem. Our experimental evaluation show that this framework enables the learning of high-quality models from merely a dozen or so labeled examples.

cs.CL↗

arXiv2020

Jacobi-Style Iteration for Distributed Submodular Maximization

This paper presents a novel Jacobi-style iteration algorithm for solving the problem of distributed submodular maximization, in which each agent determines its own strategy from a finite set so that the global submodular objective function is jointly maximized. Building on the multi-linear extension of the global submodular function, we expect to achieve the solution from a probabilistic, rather than deterministic, perspective, and thus transfer the considered problem from a discrete domain into a continuous domain. Since it is observed that an unbiased estimation of the gradient of multi-linear extension function~can be obtained by sampling the agents' local decisions, a projected stochastic gradient algorithm is proposed to solve the problem. Our algorithm enables the distributed updates among all individual agents and is proved to asymptotically converge to a desirable equilibrium solution. Such an equilibrium solution is guaranteed to achieve at least 1/2-suboptimal bound, which is comparable to the state-of-art in the literature. Moreover, we further enhance the proposed algorithm by handling the scenario in which agents' communication delays are present. The enhanced algorithmic framework admits a more realistic distributed implementation of our approach. Finally, a movie recommendation task is conducted on a real-world movie rating data set, to validate the numerical performance of the proposed algorithms.

eess.SY↗

arXiv2020

A Survey of the State of Explainable AI for Natural Language Processing

Recent years have seen important advances in the quality of state-of-the-art models, but this has come at the expense of models becoming less interpretable. This survey presents an overview of the current state of Explainable AI (XAI), considered within the domain of Natural Language Processing (NLP). We discuss the main categorization of explanations, as well as the various ways explanations can be arrived at and visualized. We detail the operations and explainability techniques currently available for generating explanations for NLP model predictions, to serve as a resource for model developers in the community. Finally, we point out the current gaps and encourage directions for future work in this important research area.

cs.CL↗

arXiv2020

MeDaS: An open-source platform as service to help break the walls between medicine and informatics

In the past decade, deep learning (DL) has achieved unprecedented success in numerous fields including computer vision, natural language processing, and healthcare. In particular, DL is experiencing an increasing development in applications for advanced medical image analysis in terms of analysis, segmentation, classification, and furthermore. On the one hand, tremendous needs that leverage the power of DL for medical image analysis are arising from the research community of a medical, clinical, and informatics background to jointly share their expertise, knowledge, skills, and experience. On the other hand, barriers between disciplines are on the road for them often hampering a full and efficient collaboration. To this end, we propose our novel open-source platform, i.e., MeDaS -- the MeDical open-source platform as Service. To the best of our knowledge, MeDaS is the first open-source platform proving a collaborative and interactive service for researchers from a medical background easily using DL related toolkits, and at the same time for scientists or engineers from information sciences to understand the medical knowledge side. Based on a series of toolkits and utilities from the idea of RINV (Rapid Implementation aNd Verification), our proposed MeDaS platform can implement pre-processing, post-processing, augmentation, visualization, and other phases needed in medical image analysis. Five tasks including the subjects of lung, liver, brain, chest, and pathology, are validated and demonstrated to be efficiently realisable by using MeDaS.

cs.CV↗

arXiv2020

Real-time Mobile Sensor Management Framework for city-scale environmental monitoring

Environmental disasters such as flash floods are becoming more and more prevalent and carry an increasing burden on human civilization. They are usually unpredictable, fast in development, and extend across large geographical areas. The consequences of such disasters can be reduced through better monitoring, for example using mobile sensing platforms that can give timely and accurate information to first responders and the public. Given the extended scale of the areas to monitor, and the time-varying nature of the phenomenon, we need fast algorithms to quickly determine the best sequence of locations to be monitored. This problem is very challenging: the present informative mobile sensor routing algorithms are either short-sighted or computationally demanding when applied to large scale systems. In this paper, a real-time sensor task scheduling algorithm that suits the features and needs of city-scale environmental monitoring tasks is proposed. The algorithm is run in forward search and makes use of the predictions of an associated distributed parameter system, modeling flash flood propagation. It partly inherits the causal relation expressed by a search tree, which describes all possible sequential decisions. The computationally heavy data assimilation steps in the forward search tree are replaced by functions dependent on the covariance matrix between observation sets. Taking flood tracking in an urban area as a concrete example, numerical experiments in this paper indicate that this scheduling algorithm can achieve better results than myopic planning algorithms and other heuristics based sensor placement algorithms. Furthermore, this paper relies on a deep learning-based data-driven model to track the system states, and experiments suggest that popular estimation techniques have very good performance when applied to precise data-driven models.

eess.SY↗

arXiv2020

An Early Study on Intelligent Analysis of Speech under COVID-19: Severity, Sleep Quality, Fatigue, and Anxiety

The COVID-19 outbreak was announced as a global pandemic by the World Health Organisation in March 2020 and has affected a growing number of people in the past few weeks. In this context, advanced artificial intelligence techniques are brought to the fore in responding to fight against and reduce the impact of this global health crisis. In this study, we focus on developing some potential use-cases of intelligent speech analysis for COVID-19 diagnosed patients. In particular, by analysing speech recordings from these patients, we construct audio-only-based models to automatically categorise the health state of patients from four aspects, including the severity of illness, sleep quality, fatigue, and anxiety. For this purpose, two established acoustic feature sets and support vector machines are utilised. Our experiments show that an average accuracy of .69 obtained estimating the severity of illness, which is derived from the number of days in hospitalisation. We hope that this study can foster an extremely fast, low-cost, and convenient way to automatically detect the COVID-19 disease.

eess.AS↗

arXiv2020

deepSELF: An Open Source Deep Self End-to-End Learning Framework

We introduce an open-source toolkit, i.e., the deep Self End-to-end Learning Framework (deepSELF), as a toolkit of deep self end-to-end learning framework for multi-modal signals. To the best of our knowledge, it is the first public toolkit assembling a series of state-of-the-art deep learning technologies. Highlights of the proposed deepSELF toolkit include: First, it can be used to analyse a variety of multi-modal signals, including images, audio, and single or multi-channel sensor data. Second, we provide multiple options for pre-processing, e.g., filtering, or spectrum image generation by Fourier or wavelet transformation. Third, plenty of topologies in terms of NN, 1D/2D/3D CNN, and RNN/LSTM/GRU can be customised and a series of pretrained 2D CNN models, e.g., AlexNet, VGGNet, ResNet can be used easily. Last but not least, above these features, deepSELF can be flexibly used not only as a single model but also as a fusion of such.

cs.LG↗

arXiv2020

Prediction of mechanical properties of non-equiatomic high-entropy alloy by atomistic simulation and machine learning

High-entropy alloys (HEAs) with multiple constituent elements have been extensively studied in the past 20 years due to their promising engineering application. Previous experimental and computational studies of HEAs focused mainly on equiatomic or near equiatomic HEAs. However, there is probably far more treasure in those non-equiatomic HEAs with carefully designed composition. In this study, molecular dynamics (MD) simulation combined with machine learning (ML) methods were used to predict the mechanical properties of non-equiatomic CuFeNiCrCo HEAs. A database was established based on a tensile test of 900 HEA single-crystal samples by MD simulation. We investigated and compared eight ML models for the learning tasks, ranging from shallow models to deep models. It was found that the kernel-based extreme learning machine (KELM) model outperformed others for the prediction of yield stress and Young's modulus. The accuracy of the KELM model was further verified by the large-sized polycrystal HEA samples.

cond-mat.mtrl-sci↗

arXiv2020

Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction

Better machine understanding of pedestrian behaviors enables faster progress in modeling interactions between agents such as autonomous vehicles and humans. Pedestrian trajectories are not only influenced by the pedestrian itself but also by interaction with surrounding objects. Previous methods modeled these interactions by using a variety of aggregation methods that integrate different learned pedestrians states. We propose the Social Spatio-Temporal Graph Convolutional Neural Network (Social-STGCNN), which substitutes the need of aggregation methods by modeling the interactions as a graph. Our results show an improvement over the state of art by 20% on the Final Displacement Error (FDE) and an improvement on the Average Displacement Error (ADE) with 8.5 times less parameters and up to 48 times faster inference speed than previously reported methods. In addition, our model is data efficient, and exceeds previous state of the art on the ADE metric with only 20% of the training data. We propose a kernel function to embed the social interactions between pedestrians within the adjacency matrix. Through qualitative analysis, we show that our model inherited social behaviors that can be expected between pedestrians trajectories. Code is available at https://github.com/abduallahmohamed/Social-STGCNN.

cs.CV↗

arXiv2020

COVID-19 and Computer Audition: An Overview on What Speech & Sound Analysis Could Contribute in the SARS-CoV-2 Corona Crisis

At the time of writing, the world population is suffering from more than 10,000 registered COVID-19 disease epidemic induced deaths since the outbreak of the Corona virus more than three months ago now officially known as SARS-CoV-2. Since, tremendous efforts have been made worldwide to counter-steer and control the epidemic by now labelled as pandemic. In this contribution, we provide an overview on the potential for computer audition (CA), i.e., the usage of speech and sound analysis by artificial intelligence to help in this scenario. We first survey which types of related or contextually significant phenomena can be automatically assessed from speech or sound. These include the automatic recognition and monitoring of breathing, dry and wet coughing or sneezing sounds, speech under cold, eating behaviour, sleepiness, or pain to name but a few. Then, we consider potential use-cases for exploitation. These include risk assessment and diagnosis based on symptom histograms and their development over time, as well as monitoring of spread, social distancing and its effects, treatment and recovery, and patient wellbeing. We quickly guide further through challenges that need to be faced for real-life usage. We come to the conclusion that CA appears ready for implementation of (pre-)diagnosis and monitoring tools, and more generally provides rich and significant, yet so far untapped potential in the fight against COVID-19 spread.

cs.SD↗

arXiv2020

Broadband Free Space Impedance in $\mathrm{Co_2Z}$ Hexaferrites by Substitution of Quadrivalent Heavy Transition Metal Ions for Miniaturized RF Devices

Polycrystalline samples of Z-type hexaferrites, having nominal compositions $\mathrm{Ba_3Co_{2+x}Fe_{24-2x}M_xO_{41}}$ where M = $\mathrm{Ir^{4+}, Hf^{4+}, Mo^{4+}}$ and x=0 and 0.05, were processed via ceramic processing protocols in pursuit of low magnetic and dielectric losses as well as equivalent permittivity and permeability. Fine process control was conducted to ensure optimal magnetic properties. Organic dispersants (i.e., isobutylene and maleic anhydride) were employed to achieve maximum densities. Crystallographic structure, characterized by X-ray diffraction, revealed that doping with $\mathrm{Ir^{4+}, Hf^{4+}, Mo^{4+}}$ did not adversely affect the crystal structure and phase purity of the Z-type hexaferrite. The measured microwave and magnetic properties show that the resonant frequency shifts depending on the specific dopant allowing for tunability of the operational frequency and bandwidth. The frequency bandwidth in which permittivity and permeability are very near equal (i.e., ~400 MHz for $\mathrm{Mo^{4+}}$ (x), where x=0.05 doping) is shown to occur at frequencies between 0.2 and 1.0 GHz depending on dopant type. These results give rise to low loss at 650 MHz, with considerable size reduction of an order of magnitude, while maintaining the characteristic impedance of free space (i.e., 377 $\mathrmΩ$). These results allow for miniaturization and optimized band-pass performance of magnetodielectric materials for communication devices such as antenna and radomes that can be engineered to operate over desired frequency ranges using cost effective and volumetric processing methodologies.

physics.app-ph↗

arXiv2019

End-to-End Trainable Non-Collaborative Dialog System

End-to-end task-oriented dialog models have achieved promising performance on collaborative tasks where users willingly coordinate with the system to complete a given task. While in non-collaborative settings, for example, negotiation and persuasion, users and systems do not share a common goal. As a result, compared to collaborate tasks, people use social content to build rapport and trust in these non-collaborative settings in order to advance their goals. To handle social content, we introduce a hierarchical intent annotation scheme, which can be generalized to different non-collaborative dialog tasks. Building upon TransferTransfo (Wolf et al. 2019), we propose an end-to-end neural network model to generate diverse coherent responses. Our model utilizes intent and semantic slots as the intermediate sentence representation to guide the generation process. In addition, we design a filter to select appropriate responses based on whether these intermediate representations fit the designed task and conversation constraints. Our non-collaborative dialog model guides users to complete the task while simultaneously keeps them engaged. We test our approach on our newly proposed ANTISCAM dataset and an existing PERSUASIONFORGOOD dataset. Both automatic and human evaluations suggest that our model outperforms multiple baselines in these two non-collaborative tasks.

cs.CL↗

arXiv2019

Adabot: Fault-Tolerant Java Decompiler

Reverse Engineering(RE) has been a fundamental task in software engineering. However, most of the traditional Java reverse engineering tools are strictly rule defined, thus are not fault-tolerant, which pose serious problem when noise and interference were introduced into the system. In this paper, we view reverse engineering as a statistical machine translation task instead of rule-based task, and propose a fault-tolerant Java decompiler based on machine translation models. Our model is based on attention-based Neural Machine Translation (NMT) and Transformer architectures. First, we measure the translation quality on both the redundant and purified datasets. Next, we evaluate the fault-tolerance(anti-noise ability) of our framework on test sets with different unit error probability (UEP). In addition, we compare the suitability of different word segmentation algorithms for decompilation task. Experimental results demonstrate that our model is more robust and fault-tolerant compared to traditional Abstract Syntax Tree (AST) based decompilers. Specifically, in terms of BLEU-4 and Word Error Rate (WER), our performance has reached 94.50% and 2.65% on the redundant test set; 92.30% and 3.48% on the purified test set.

cs.SE↗

arXiv2019

High Performance Metallic Amorphous Magnetic Flake-based Magnetodielectric Inductors

Flake-shaped FeSi-based metallic amorphous alloy particles, having an aspect ratio as high as 175 to 1, were prepared by ball milling gas atomized amorphous powders of an effective diameter of 20 micrometers. The starting powder had a saturation magnetic flux density, Bs, of 1.5 T and a coercivity, Hc, of 94 A/m. The aspect ratio of the flakes, as well as their magnetic properties, were controlled by milling process parameters, such as duration, speed, and the type, mass and diameter of the milling balls. To minimize the oxidation of the charge, the powders were handled in an argon gas-purged glove box, milled in toluene, and subsequently dried in vacuo. Subsequently, soft magnetodielectric composites were prepared by suspending and aligning the FeSi-based powders in paraffin wax or epoxy resin. The composites were then pressed into toroids for measurements of their high frequency complex permeability by a vector network analyzer. The influence of the flake aspect ratio and volume loading fraction on the permeability of the composites were investigated. Results indicate that the composite permeability increases with the flake aspect ratio. For example, for a given loading factor of 30 vol.%, the composite permeability at 0.1 GHz nearly tripled and approached the value of 10 by increasing the aspect ratio of the FeSi-based inclusions from 1 (spheres) to greater than 175 to 1 (flakes).

physics.app-ph↗

arXiv2019

How to Build User Simulators to Train RL-based Dialog Systems

User simulators are essential for training reinforcement learning (RL) based dialog models. The performance of the simulator directly impacts the RL policy. However, building a good user simulator that models real user behaviors is challenging. We propose a method of standardizing user simulator building that can be used by the community to compare dialog system quality using the same set of user simulators fairly. We present implementations of six user simulators trained with different dialog planning and generation methods. We then calculate a set of automatic metrics to evaluate the quality of these simulators both directly and indirectly. We also ask human users to assess the simulators directly and indirectly by rating the simulated dialogs and interacting with the trained systems. This paper presents a comprehensive evaluation framework for user simulator study and provides a better understanding of the pros and cons of different user simulators, as well as their impacts on the trained systems.

cs.CL↗

arXiv2019

Physics Informed Data Driven model for Flood Prediction: Application of Deep Learning in prediction of urban flood development

Flash floods in urban areas occur with increasing frequency. Detecting these floods would greatlyhelp alleviate human and economic losses. However, current flood prediction methods are eithertoo slow or too simplified to capture the flood development in details. Using Deep Neural Networks,this work aims at boosting the computational speed of a physics-based 2-D urban flood predictionmethod, governed by the Shallow Water Equation (SWE). Convolutional Neural Networks(CNN)and conditional Generative Adversarial Neural Networks(cGANs) are applied to extract the dy-namics of flood from the data simulated by a Partial Differential Equation(PDE) solver. Theperformance of the data-driven model is evaluated in terms of Mean Squared Error(MSE) andPeak Signal to Noise Ratio(PSNR). The deep learning-based, data-driven flood prediction modelis shown to be able to provide precise real-time predictions of flood development

eess.SP↗