Searcharxiv⌕ Search

arXiv subjects

Ning Kang

Publications and source records attributed to Ning Kang.

34 records · Page 2Linked to original sources
arXiv2021

iFlow: Numerically Invertible Flows for Efficient Lossless Compression via a Uniform Coder

It was estimated that the world produced $59 ZB$ ($5.9 \times 10^{13} GB$) of data in 2020, resulting in the enormous costs of both data storage and transmission. Fortunately, recent advances in deep generative models have spearheaded a new class of so-called "neural compression" algorithms, which significantly outperform traditional codecs in terms of compression ratio. Unfortunately, the application of neural compression garners little commercial interest due to its limited bandwidth; therefore, developing highly efficient frameworks is of critical practical importance. In this paper, we discuss lossless compression using normalizing flows which have demonstrated a great capacity for achieving high compression ratios. As such, we introduce iFlow, a new method for achieving efficient lossless compression. We first propose Modular Scale Transform (MST) and a novel family of numerically invertible flow transformations based on MST. Then we introduce the Uniform Base Conversion System (UBCS), a fast uniform-distribution codec incorporated into iFlow, enabling efficient compression. iFlow achieves state-of-the-art compression ratios and is $5\times$ quicker than other high-performance schemes. Furthermore, the techniques presented in this paper can be used to accelerate coding time for a broad class of flow-based algorithms.

cs.LG↗
arXiv2021

NASOA: Towards Faster Task-oriented Online Fine-tuning with a Zoo of Models

Fine-tuning from pre-trained ImageNet models has been a simple, effective, and popular approach for various computer vision tasks. The common practice of fine-tuning is to adopt a default hyperparameter setting with a fixed pre-trained model, while both of them are not optimized for specific tasks and time constraints. Moreover, in cloud computing or GPU clusters where the tasks arrive sequentially in a stream, faster online fine-tuning is a more desired and realistic strategy for saving money, energy consumption, and CO2 emission. In this paper, we propose a joint Neural Architecture Search and Online Adaption framework named NASOA towards a faster task-oriented fine-tuning upon the request of users. Specifically, NASOA first adopts an offline NAS to identify a group of training-efficient networks to form a pretrained model zoo. We propose a novel joint block and macro-level search space to enable a flexible and efficient search. Then, by estimating fine-tuning performance via an adaptive model by accumulating experience from the past tasks, an online schedule generator is proposed to pick up the most suitable model and generate a personalized training regime with respect to each desired task in a one-shot fashion. The resulting model zoo is more training efficient than SOTA models, e.g. 6x faster than RegNetY-16GF, and 1.7x faster than EfficientNetB3. Experiments on multiple datasets also show that NASOA achieves much better fine-tuning results, i.e. improving around 2.1% accuracy than the best performance in RegNet series under various constraints and tasks; 40x faster compared to the BOHB.

cs.CV↗
arXiv2021

First radiative shock experiments on the SG-II laser

We report on the design and first results from experiments looking at the formation of radiative shocks on the Shenguang-II (SG-II) laser at the Shanghai Institute of Optics and Fine Mechanics in China. Laser-heating of a two-layer CH/CH-Br foil drives a $\sim$40 km/s shock inside a gas-cell filled with argon at an initial pressure of 1 bar. The use of gas-cell targets with large (several mm) lateral and axial extent allows the shock to propagate freely without any wall interactions, and permits a large field of view to image single and colliding counter-propagating shocks with time resolved, point-projection X-ray backlighting ($\sim20$ $μ$m source size, 4.3 keV photon energy). Single shocks were imaged up to 100 ns after the onset of the laser drive allowing to probe the growth of spatial non-uniformities in the shock apex. These results are compared with experiments looking at counter-propagating shocks, showing a symmetric drive which leads to a collision and stagnation from $\sim$40 ns onward. We present a preliminary comparison with numerical simulations with the radiation hydrodynamics code ARWEN, which provides expected plasma parameters for the design of future experiments in this facility.

physics.plasm-ph↗
arXiv2020

VEGA: Towards an End-to-End Configurable AutoML Pipeline

Automated Machine Learning (AutoML) is an important industrial solution for automatic discovery and deployment of the machine learning models. However, designing an integrated AutoML system faces four great challenges of configurability, scalability, integrability, and platform diversity. In this work, we present VEGA, an efficient and comprehensive AutoML framework that is compatible and optimized for multiple hardware platforms. a) The VEGA pipeline integrates various modules of AutoML, including Neural Architecture Search (NAS), Hyperparameter Optimization (HPO), Auto Data Augmentation, Model Compression, and Fully Train. b) To support a variety of search algorithms and tasks, we design a novel fine-grained search space and its description language to enable easy adaptation to different search algorithms and tasks. c) We abstract the common components of deep learning frameworks into a unified interface. VEGA can be executed with multiple back-ends and hardwares. Extensive benchmark experiments on multiple tasks demonstrate that VEGA can improve the existing AutoML algorithms and discover new high-performance models against SOTA methods, e.g. the searched DNet model zoo for Ascend 10x faster than EfficientNet-B5 and 9.2x faster than RegNetX-32GF on ImageNet. VEGA is open-sourced at https://github.com/huawei-noah/vega.

cs.CV↗
arXiv2020

Bremsstrahlung emission and plasma characterization driven by moderately relativistic laser-plasma interactions

Relativistic electrons generated by the interaction of petawatt-class short laser pulses with solid targets can be used to generate bright X-rays via bremsstrahlung. The efficiency of laser energy transfer into these electrons depends on multiple parameters including the focused intensity and pre-plasma level. This paper reports experimental results from the interaction of a high intensity petawatt-class glass laser pulses with solid targets at a maximum intensity of $10^{19}$ W/cm$^2$. In-situ measurements of specularly reflected light are used to provide an upper bound of laser absorption and to characterize focused laser intensity, the pre-plasma level and the generation mechanism of second harmonic light. The measured spectrum of electrons and bremsstrahlung radiation provide information about the efficiency of laser energy transfer.

physics.plasm-ph↗
arXiv2019

AIM 2019 Challenge on Constrained Super-Resolution: Methods and Results

This paper reviews the AIM 2019 challenge on constrained example-based single image super-resolution with focus on proposed solutions and results. The challenge had 3 tracks. Taking the three main aspects (i.e., number of parameters, inference/running time, fidelity (PSNR)) of MSRResNet as the baseline, Track 1 aims to reduce the amount of parameters while being constrained to maintain or improve the running time and the PSNR result, Tracks 2 and 3 aim to optimize running time and PSNR result with constrain of the other two aspects, respectively. Each track had an average of 64 registered participants, and 12 teams submitted the final results. They gauge the state-of-the-art in single image super-resolution.

eess.IV↗
arXiv2018

Online Makespan Minimization: The Power of Restart

We consider the online makespan minimization problem on identical machines. Chen and Vestjens (ORL 1997) show that the largest processing time first (LPT) algorithm is 1.5-competitive. For the special case of two machines, Noga and Seiden (TCS 2001) introduce the SLEEPY algorithm that achieves a competitive ratio of $(5 - \sqrt{5})/2 \approx 1.382$, matching the lower bound by Chen and Vestjens (ORL 1997). Furthermore, Noga and Seiden note that in many applications one can kill a job and restart it later, and they leave an open problem whether algorithms with restart can obtain better competitive ratios. We resolve this long-standing open problem on the positive end. Our algorithm has a natural rule for killing a processing job: a newly-arrived job replaces the smallest processing job if 1) the new job is larger than other pending jobs, 2) the new job is much larger than the processing one, and 3) the processed portion is small relative to the size of the new job. With appropriate choice of parameters, we show that our algorithm improves the 1.5 competitive ratio for the general case, and the 1.382 competitive ratio for the two-machine case.

cs.DS↗
arXiv2018

How to Match when All Vertices Arrive Online

We introduce a fully online model of maximum cardinality matching in which all vertices arrive online. On the arrival of a vertex, its incident edges to previously-arrived vertices are revealed. Each vertex has a deadline that is after all its neighbors' arrivals. If a vertex remains unmatched until its deadline, the algorithm must then irrevocably either match it to an unmatched neighbor, or leave it unmatched. The model generalizes the existing one-sided online model and is motivated by applications including ride-sharing platforms, real-estate agency, etc. We show that the Ranking algorithm by Karp et al. (STOC 1990) is $0.5211$-competitive in our fully online model for general graphs. Our analysis brings a novel charging mechanic into the randomized primal dual technique by Devanur et al. (SODA 2013), allowing a vertex other than the two endpoints of a matched edge to share the gain. To our knowledge, this is the first analysis of Ranking that beats $0.5$ on general graphs in an online matching problem, a first step towards solving the open problem by Karp et al. (STOC 1990) about the optimality of Ranking on general graphs. If the graph is bipartite, we show that the competitive ratio of Ranking is between $0.5541$ and $0.5671$. Finally, we prove that the fully online model is strictly harder than the previous model as no online algorithm can be $0.6317 < 1-\frac{1}{e}$-competitive in our model even for bipartite graphs.

cs.DS↗
arXiv2017

Electron-Hole Symmetry Breaking in Charge Transport in Nitrogen-Doped Graphene

Graphitic nitrogen-doped graphene is an excellent platform to study scattering processes of massless Dirac fermions by charged impurities, in which high mobility can be preserved due to the absence of lattice defects through direct substitution of carbon atoms in the graphene lattice by nitrogen atoms. In this work, we report on electrical and magnetotransport measurements of high-quality graphitic nitrogen-doped graphene. We show that the substitutional nitrogen dopants in graphene introduce atomically sharp scatters for electrons but long-range Coulomb scatters for holes and, thus, graphitic nitrogen-doped graphene exhibits clear electron-hole asymmetry in transport properties. Dominant scattering processes of charge carriers in graphitic nitrogen-doped graphene are analyzed. It is shown that the electron-hole asymmetry originates from a distinct difference in intervalley scattering of electrons and holes. We have also carried out the magnetotransport measurements of graphitic nitrogen-doped graphene at different temperatures and the temperature dependences of intervalley scattering, intravalley scattering and phase coherent scattering rates are extracted and discussed. Our results provide an evidence for the electron-hole asymmetry in the intervalley scattering induced by substitutional nitrogen dopants in graphene and shine a light on versatile and potential applications of graphitic nitrogen-doped graphene in electronic and valleytronic devices.

cond-mat.mes-hall↗
arXiv2016

Online Submodular Maximization with Free Disposal: Randomization Beats 0.25 for Partition Matroids

We study the online submodular maximization problem with free disposal under a matroid constraint. Elements from some ground set arrive one by one in rounds, and the algorithm maintains a feasible set that is independent in the underlying matroid. In each round when a new element arrives, the algorithm may accept the new element into its feasible set and possibly remove elements from it, provided that the resulting set is still independent. The goal is to maximize the value of the final feasible set under some monotone submodular function, to which the algorithm has oracle access. For $k$-uniform matroids, we give a deterministic algorithm with competitive ratio at least $0.2959$, and the ratio approaches $\frac{1}{α_\infty} \approx 0.3178$ as $k$ approaches infinity, improving the previous best ratio of $0.25$ by Chakrabarti and Kale (IPCO 2014), Buchbinder et al. (SODA 2015) and Chekuri et al. (ICALP 2015). We also show that our algorithm is optimal among a class of deterministic monotone algorithms that accept a new arriving element only if the objective is strictly increased. Further, we prove that no deterministic monotone algorithm can be strictly better than $0.25$-competitive even for partition matroids, the most modest generalization of $k$-uniform matroids, matching the competitive ratio by Chakrabarti and Kale (IPCO 2014) and Chekuri et al. (ICALP 2015). Interestingly, we show that randomized algorithms are strictly more powerful by giving a (non-monotone) randomized algorithm for partition matroids with ratio $\frac{1}{α_\infty} \approx 0.3178$.

cs.DM↗
arXiv2015

Free-Standing Two-Dimensional Single-Crystalline InSb Nanosheets

Growth of high-quality single-crystalline InSb layers remains challenging in material science. Such layered InSb materials are highly desired for searching for and manipulation of Majorana fermions in solid state, a fundamental research task in physics today, and for development of novel high-speed nanoelectronic and infrared optoelectronic devices. Here we report on a new route towards growth of single-crystalline, layered InSb materials. We demonstrate the successful growth of free-standing, two-dimensional InSb nanosheets on one-dimensional InAs nanowires by molecular-beam epitaxy. The grown InSb nanosheets are pure zinc-blende single crystals. The length and width of the InSb nanosheets are up to several micrometers and the thickness is down to ~10 nm. The InSb nanosheets show a clear ambipolar behavior and a high electron mobility. Our work will open up new technology routes towards the development of InSb-based devices for applications in nanoelectronics, optoelectronics and quantum electronics, and for study of fundamental physical phenomena.

cond-mat.mtrl-sci↗
arXiv2015

Online Convex Covering and Packing Problems

We study the online convex covering problem and online convex packing problem. The (offline) convex covering problem is modeled by the following convex program: $\min_{x \in R_+^n} f(x) \ \text{s.t}\ A x \ge 1$, where $f : R_+^n \mapsto R_+$ is a monotone and convex cost function, and $A$ is an $m \times n$ matrix with non-negative entries. Each row of the constraint matrix $A$ corresponds to a covering constraint. In the online problem, each row of $A$ comes online and the algorithm must maintain a feasible assignment $x$ and may only increase $x$ over time. The (offline) convex packing problem is modeled by the following convex program: $\max_{y\in R_+^m} \sum_{j = 1}^m y_j - g(A^T y)$, where $g : R_+^n \mapsto R_+$ is a monotone and convex cost function. It is the Fenchel dual program of convex covering when $g$ is the convex conjugate of $f$. In the online problem, each variable $y_j$ arrives online and the algorithm must decide the value of $y_j$ on its arrival. We propose simple online algorithms for both problems using the online primal dual technique, and obtain nearly optimal competitive ratios for both problems for the important special case of polynomial cost functions. For any convex polynomial cost functions with non-negative coefficients and maximum degree $τ$, we introduce an $O(τ\log n)^τ$-competitive online convex covering algorithm, and an $O(τ)$-competitive online convex packing algorithm, matching the known $Ω(τ\log n)^τ$ and $Ω(τ)$ lower bounds respectively. There is a large family of online resource allocation problems that can be modeled under this online convex covering and packing framework, including online covering and packing problems (with linear objectives), online mixed covering and packing, and online combinatorial auction. Our framework allows us to study these problems using a unified approach.

cs.DS↗
arXiv2010

A Concept Annotation System for Clinical Records

Unstructured information comprises a valuable source of data in clinical records. For text mining in clinical records, concept extraction is the first step in finding assertions and relationships. This study presents a system developed for the annotation of medical concepts, including medical problems, tests, and treatments, mentioned in clinical records. The system combines six publicly available named entity recognition system into one framework, and uses a simple voting scheme that allows to tune precision and recall of the system to specific needs. The system provides both a web service interface and a UIMA interface which can be easily used by other systems. The system was tested in the fourth i2b2 challenge and achieved an F-score of 82.1% for the concept exact match task, a score which is among the top-ranking systems. To our knowledge, this is the first publicly available clinical record concept annotation system.

cs.IR↗
arXiv2010

Domain walls at the spin density wave endpoint of the organic superconductor (TMTSF)2PF6 under pressure

We report the first comprehensive investigation of the organic superconductor (TMTSF)2PF6 in the vicinity of the endpoint of the spin density wave - metal phase transition where phase coexistence occurs. At low temperature, the transition of metallic domains towards superconductivity is used to reveal the various textures. In particular, we demonstrate experimentally the existence of 1D and 2D metallic domains with a cross-over from a filamentary superconductivity mostly along the c?-axis to a 2D superconductivity in the b?c-plane perpendicular to the most conducting direction. The formation of these domain walls may be related to the proposal of a soliton phase in the vicinity of the critical pressure of the (TMTSF)2PF6 phase diagram.

cond-mat.str-el↗
arXiv2007

Observation of Spin-Orbit Berry's Phase in Magnetoresistance of a Two-Dimensional Hole Anti-dot System

We report observation of spin-orbit Berry's phase in the Aharonov-Bohm (AB) type oscillation of weak field magnetoresistance in an anti-dot lattice (ADL) of a two-dimensional hole system. An AB-type oscillation is superposed on the commensurability peak, and the main peak in the Fourier transform is clearly split up due to variation in Berry's phase originating from the spin-orbit interaction. A simulation considering Berry's phase and the phase arising from the spin-orbit shift in the momentum space shows qualitative agreement with the experiment.

cond-mat.mes-hall↗