arXiv subjects

Hao Wu

Publications and source records attributed to Hao Wu.

At least 829 records · Page 46Linked to original sources

arXiv2018

A Multi-Axis Annotation Scheme for Event Temporal Relations

Existing temporal relation (TempRel) annotation schemes often have low inter-annotator agreements (IAA) even between experts, suggesting that the current annotation task needs a better definition. This paper proposes a new multi-axis modeling to better capture the temporal structure of events. In addition, we identify that event end-points are a major source of confusion in annotation, so we also propose to annotate TempRels based on start-points only. A pilot expert annotation using the proposed scheme shows significant improvement in IAA from the conventional 60's to 80's (Cohen's Kappa). This better-defined annotation scheme further enables the use of crowdsourcing to alleviate the labor intensity for each annotator. We hope that this work can foster more interesting studies towards event understanding.

cs.CL↗

arXiv2018

A deep convolutional encoder-decoder neural network in assisting seismic horizon tracking

Seismic horizons are geologically significant surfaces that can be used for building geology structure and stratigraphy models. However, horizon tracking in 3D seismic data is a time-consuming and challenging problem. Relief human from the tedious seismic interpretation is one of the hot research topics. We proposed a novel automatically seismic horizon tracking method by using a deep convolutional neural network. We employ a state-of-art end-to-end semantic segmentation method to track the seismic horizons automatically. Experiment result shows that our proposed neural network can automatically track multiple horizons simultaneously. We validate the effectiveness and robustness of our proposed method by comparing automatically tracked horizons with manually picked horizons.

physics.geo-ph↗

arXiv2018

MorphNet: Fast & Simple Resource-Constrained Structure Learning of Deep Networks

We present MorphNet, an approach to automate the design of neural network structures. MorphNet iteratively shrinks and expands a network, shrinking via a resource-weighted sparsifying regularizer on activations and expanding via a uniform multiplicative factor on all layers. In contrast to previous approaches, our method is scalable to large networks, adaptable to specific resource constraints (e.g. the number of floating-point operations per inference), and capable of increasing the network's performance. When applied to standard network architectures on a wide variety of datasets, our approach discovers novel structures in each domain, obtaining higher performance while respecting the resource constraint.

cs.LG↗

arXiv2018

Improving Temporal Relation Extraction with a Globally Acquired Statistical Resource

Extracting temporal relations (before, after, overlapping, etc.) is a key aspect of understanding events described in natural language. We argue that this task would gain from the availability of a resource that provides prior knowledge in the form of the temporal order that events usually follow. This paper develops such a resource -- a probabilistic knowledge base acquired in the news domain -- by extracting temporal relations between events from the New York Times (NYT) articles over a 20-year span (1987--2007). We show that existing temporal extraction systems can be improved via this resource. As a byproduct, we also show that interesting statistics can be retrieved from this resource, which can potentially benefit other time-aware tasks. The proposed system and resource are both publicly available.

cs.AI↗

arXiv2018

Loss Rank Mining: A General Hard Example Mining Method for Real-time Detectors

Modern object detectors usually suffer from low accuracy issues, as foregrounds always drown in tons of backgrounds and become hard examples during training. Compared with those proposal-based ones, real-time detectors are in far more serious trouble since they renounce the use of region-proposing stage which is used to filter a majority of backgrounds for achieving real-time rates. Though foregrounds as hard examples are in urgent need of being mined from tons of backgrounds, a considerable number of state-of-the-art real-time detectors, like YOLO series, have yet to profit from existing hard example mining methods, as using these methods need detectors fit series of prerequisites. In this paper, we propose a general hard example mining method named Loss Rank Mining (LRM) to fill the gap. LRM is a general method for real-time detectors, as it utilizes the final feature map which exists in all real-time detectors to mine hard examples. By using LRM, some elements representing easy examples in final feature map are filtered and detectors are forced to concentrate on hard examples during training. Extensive experiments validate the effectiveness of our method. With our method, the improvements of YOLOv2 detector on auto-driving related dataset KITTI and more general dataset PASCAL VOC are over 5% and 2% mAP, respectively. In addition, LRM is the first hard example mining strategy which could fit YOLOv2 perfectly and make it better applied in series of real scenarios where both real-time rates and accurate detection are strongly demanded.

cs.CV↗

arXiv2018

The freezing Rènyi quantum discord

As a universal quantum character of quantum correlation, the freezing phenomenon is researched by geometry and quantum discord methods, respectively. In this paper, the properties of R`enyi discord is studied for two independent Dimer System coupled to two correlated Fermi-spin environments under the non-Markovian condition. We further demonstrate that the freezing behaviors still exist for R`enyi discord and study the effects of different parameters on this behaviors.

quant-ph↗

arXiv2018

Betti Numbers of the HOMFLYPT Homology

In arXiv:math/0508510, Rasmussen observed that the Khovanov-Rozansky homology of a link is a finitely generated module over the polynomial ring generated by the components of this link. In the current paper, we study the module structure of the middle HOMFLYPT homology, especially the Betti numbers of this module. For each link, these Betti numbers are supported on a finite subset of $\mathbb{Z}^4$. One can easily recover from these Betti numbers the Poincaré polynomial of the middle HOMFLYPT homology. We explain why the Betti numbers can be viewed as a generalization of the reduced HOMFLYPT homology of knots. As an application, we prove that the projective dimension of the middle HOMFLYPT homology is additive under split union of links and provides a new obstruction to split links.

math.GT↗

arXiv2018

Energy and Delay Optimization for Cache-Enabled Dense Small Cell Networks

Caching popular files in small base stations (SBSs) has been proved to be an effective way to reduce bandwidth pressure on the backhaul links of dense small cell networks (DSCNs). Many existing studies on cache-enabled DSCNs attempt to improve user experience by optimizing end-to-end file delivery delay. However, under practical scenarios where files (e.g., video files) have diverse quality of service requirements, energy consumption at SBSs should also be concerned from the network perspective. In this paper,we attempt to optimize these two critical metrics in cache-enabled DSCNs. Firstly, we formulate the energy-delay optimization problem as a Mixed Integer Programming (MIP) problem, where file placement, user association and power control are jointly considered. To model the tradeoff relationship between energy consumption and end-to-end file delivery delay, a utility function linearly combining these two metrics is used as an objective function of the optimization problem. Then, we solve the problem in two stages, i.e. caching stage and delivery stage, based on the observation that caching is performed during off-peak time. At the caching stage, a local popular file placement policy is proposed by estimating user preference at each SBS. At the delivery stage, with given caching status at SBSs, the MIP problem is further decomposed by Benders' decomposition method. An efficient algorithm is proposed to approach the optimal association and power solution by iteratively shrinking the gap of the upper and lower bounds. Finally, extension simulations are performed to validate our analytical and algorithmic work. The results demonstrate that the proposed algorithms can achieve the optimal tradeoff between energy consumption and end-to-end file delivery delay.

cs.NI↗

arXiv2018

Generation of multiphoton entangled quantum states with a single silicon nanowire

Multiphoton entanglement plays a critical role in quantum information processing, and greatly improves our fundamental understanding of the quantum world. Despite tremendous efforts in either bulk media or fiber-based devices, nonlinear interactions in integrated circuits show great promise as an excellent platform for photon pair generation with its high brightness, stability and scalability \cite{Caspani2017}. Here, we demonstrate the generation of bi- and multiphoton polarization entangled qubits in a single silicon nanowire waveguide, and these qubits directly compatible with the dense wavelength division multiplexing in telecommunication system. Multiphoton interference and quantum state tomography were used to characterize the quality of the entangled states. Four-photon entanglement states among two frequency channels were ascertained with a fidelity of $0.78\pm0.02$. Our work realizes the integrated multiphoton source in a relatively simple pattern and paves a way for the revolution of multiphoton quantum science.

quant-ph↗

arXiv2018

Enhancing radical molecular beams by skimmer cooling

A high-intensity supersonic beam source has been a key component in studies of molecular collisions, molecule-surface interaction, chemical reactions, and precision spectroscopy. However, the molecular density available for experiments in a downstream science chamber is limited by skimmer clogging, which constrains the separation between a valve and a skimmer to at least several hundred nozzle diameters. A recent experiment (Science Advances, 2017, 3, e1602258) has introduced a new strategy to address this challenge: when a skimmer is cooled to a temperature below the freezing point of the carrier gas, skimmer clogging can be effectively suppressed. We go beyond this proof-of-principle work in several key ways. Firstly, we apply the skimmer cooling approach to discharge-produced radical and metastable beams entrained in a carrier gas. We also identify two different processes for skimmer clogging mitigation-shockwave suppression at temperatures around the carrier gas freezing point and diffusive clogging at even lower temperatures. With the carrier clogging removed, we now fully optimize the production of entrained species such as hydroxyl radicals, resulting in a gain of 30 in density over the best commercial devices. The gain arises from both clogging mitigation and favorable geometry with a much shorter valve-skimmer distance.

physics.chem-ph↗

arXiv2018

Hypergeometric SLE: Conformal Markov Characterization and Applications

This article pertains to the classification of pairs of simple random curves with conformal Markov property and symmetry. We give the complete classification of such curves: conformal Markov property and symmetry single out a two-parameter family of random curves---Hypergeometric SLE---denoted by hSLE$_κ(ν)$ for $κ\in (0,4]$ and $ν<κ-6$. The proof relies crucially on Dubédat's commutation relation [Dub07] and a uniqueness result proved in [MS16b]. The classification indicates that hypergeometric SLE is the only possible scaling limit of the interfaces in critical lattice models (conjectured or proved to be conformal invariant) in topological rectangles with alternating boundary conditions. We also prove various properties of hSLE: continuity, reversibility, target-independence, and conditional law characterization. As by-products, we give two applications of these properties. The first one is about the critical Ising interfaces. We prove the convergence of the Ising interface in rectangles with alternating boundary conditions. This result was first proved by Izyurov in [Izy15], but our proof is new which is based on the properties of hSLE. The second application is the existence of the so-called pure partition functions of multiple SLEs. Such existence was proved for $κ\in (0,8)\setminus \mathbb{Q}$ in [KP16], and it was later proved for $κ\in (0,4]$ in [PW17]. We give a new proof of the existence for $κ\in (0,6]$ using the properties of hSLE.

math.PR↗

arXiv2018

Mixed Precision Training

Deep neural networks have enabled progress in a wide variety of applications. Growing the size of the neural network typically results in improved accuracy. As model sizes grow, the memory and compute requirements for training these models also increases. We introduce a technique to train deep neural networks using half precision floating point numbers. In our technique, weights, activations and gradients are stored in IEEE half-precision format. Half-precision floating numbers have limited numerical range compared to single-precision numbers. We propose two techniques to handle this loss of information. Firstly, we recommend maintaining a single-precision copy of the weights that accumulates the gradients after each optimizer step. This single-precision copy is rounded to half-precision format during training. Secondly, we propose scaling the loss appropriately to handle the loss of information with half-precision gradients. We demonstrate that this approach works for a wide variety of models including convolution neural networks, recurrent neural networks and generative adversarial networks. This technique works for large scale models with more than 100 million parameters trained on large datasets. Using this approach, we can reduce the memory consumption of deep learning models by nearly 2x. In future processors, we can also expect a significant computation speedup using half-precision hardware units.

cs.AI↗

arXiv2018

DeepTravel: a Neural Network Based Travel Time Estimation Model with Auxiliary Supervision

Estimating the travel time of a path is of great importance to smart urban mobility. Existing approaches are either based on estimating the time cost of each road segment which are not able to capture many cross-segment complex factors, or designed heuristically in a non-learning-based way which fail to utilize the existing abundant temporal labels of the data, i.e., the time stamp of each trajectory point. In this paper, we leverage on new development of deep neural networks and propose a novel auxiliary supervision model, namely DeepTravel, that can automatically and effectively extract different features, as well as make full use of the temporal labels of the trajectory data. We have conducted comprehensive experiments on real datasets to demonstrate the out-performance of DeepTravel over existing approaches.

cs.LG↗

arXiv2018

Role of dimensional crossover on spin-orbit torque efficiency in magnetic insulator thin films

Magnetic insulators (MIs) attract tremendous interest for spintronic applications due to low Gilbert damping and absence of Ohmic loss. Magnetic order of MIs can be manipulated and even switched by spin-orbit torques (SOTs) generated through spin Hall effect and Rashba-Edelstein effect in heavy metal/MI bilayers. SOTs on MIs are more intriguing than magnetic metals since SOTs cannot be transferred to MIs through direct injection of electron spins. Understanding of SOTs on MIs remains elusive, especially how SOTs scale with the film thickness. Here, we observe the critical role of dimensionality on the SOT efficiency by systematically studying the MI layer thickness dependent SOT efficiency in tungsten/thulium iron garnet (W/TmIG) bilayers. We first show that the TmIG thin film evolves from two-dimensional to three-dimensional magnetic phase transitions as the thickness increases, due to the suppression of long-wavelength thermal fluctuation. Then, we report the significant enhancement of the measured SOT efficiency as the thickness increases. We attribute this effect to the increase of the magnetic moment density in concert with the suppression of thermal fluctuations. At last, we demonstrate the current-induced SOT switching in the W/TmIG bilayers with a TmIG thickness up to 15 nm. The switching current density is comparable with those of heavy metal/ferromagnetic metal cases. Our findings shed light on the understanding of SOTs in MIs, which is important for the future development of ultrathin MI-based low-power spintronics.

cond-mat.mes-hall↗

arXiv2018

Polychromatic Arm Exponents for the Critical Planar FK-Ising model

Schramm Loewner Evolution (SLE) is a one-parameter family of random planar curves introduced by Oded Schramm in 1999 as the candidates for the scaling limits of the interfaces in the planar critical lattice models. This is the only possible process with conformal invariance and a certain "domain Markov property". In 2010, Chelkak and Smirnov proved the conformal invariance of the scaling limits of the critial planar FK-Ising model which gave the convergence of the interface to SLE$_{16/3}$. We derive the arm exponents of SLE$_κ$ for $κ\in (4,8)$. Combining with the convergence of the interface, we derive the arm exponents of the critical FK-Ising model. We obtain six different patterns of boundary arm exponents and three different patterns of interior arm exponents of the critical planar FK-Ising model on the square lattice.

math.PR↗

arXiv2018

Optimal data-driven estimation of generalized Markov state models for non-equilibrium dynamics

There are multiple ways in which a stochastic system can be out of statistical equilibrium. It might be subject to time-varying forcing; or be in a transient phase on its way towards equilibrium; it might even be in equilibrium without us noticing it, due to insufficient observations; and it even might be a system failing to admit an equilibrium distribution at all. We review some of the approaches that model the effective statistical behavior of equilibrium and non-equilibrium dynamical systems, and show that both cases can be considered under the unified framework of optimal low-rank approximation of so-called transfer operators. Particular attention is given to the connection between these methods, Markov state models, and the concept of metastability, further to the estimation of such reduced order models from finite simulation data. We illustrate our considerations by numerical examples.

math.DS↗

arXiv2018

Lyapunov exponent and Wasserstein metric as validation tools for assessing short-time dynamics and quantitative model evaluation of large-eddy simulation

In this work, methods for the evaluation of LES-quality and LES-accuracy are presented, which include the Lyapunov exponent for the analysis of short-time predictability of LES-calculation and the Wasserstein metric for the quantitative assessment of simulation results. Both methods are derived and evaluated in application to the Volvo test case. Both the non-reacting and reacting cases are calculated. For the non- reacting cases, good agreement with the experimental data is achieved by solvers at high numerical resolution. The reacting cases are more challenging due to the small length scale of the flame and the suppression of sinuous mode of absolute instability by the density ratio. The analysis of the turbulent simulation data using the concept of the Lyapunov exponent and the Wasserstein metric provides a more quantitative approach to assess the mesh dependency of the simulation results. The convergence of the Lyapunov exponent is shown to be a more sensitive and stronger indication of mesh-independence. Though grid convergence for the reacting cases cannot be reached with the chosen resolutions, the Lyapunov exponents and the Wasserstein metric are shown to be capable of identifying quantity-specific sensitivities with respect to the numerical resolution, while requiring significantly less computational resources than acquiring profiles of conventional turbulent statistics.

physics.flu-dyn↗

arXiv2017

VAMPnets: Deep learning of molecular kinetics

There is an increasing demand for computing the relevant structures, equilibria and long-timescale kinetics of biomolecular processes, such as protein-drug binding, from high-throughput molecular dynamics simulations. Current methods employ transformation of simulated coordinates into structural features, dimension reduction, clustering the dimension-reduced data, and estimation of a Markov state model or related model of the interconversion rates between molecular structures. This handcrafted approach demands a substantial amount of modeling expertise, as poor decisions at any step will lead to large modeling errors. Here we employ the variational approach for Markov processes (VAMP) to develop a deep learning framework for molecular kinetics using neural networks, dubbed VAMPnets. A VAMPnet encodes the entire mapping from molecular coordinates to Markov states, thus combining the whole data processing pipeline in a single end-to-end framework. Our method performs equally or better than state-of-the art Markov modeling methods and provides easily interpretable few-state kinetic models.

stat.ML↗