arXiv subjects

Lei Han

Publications and source records attributed to Lei Han.

126 records · Page 7Linked to original sources

arXiv2020

GFF: Gated Fully Fusion for Semantic Segmentation

Semantic segmentation generates comprehensive understanding of scenes through densely predicting the category for each pixel. High-level features from Deep Convolutional Neural Networks already demonstrate their effectiveness in semantic segmentation tasks, however the coarse resolution of high-level features often leads to inferior results for small/thin objects where detailed information is important. It is natural to consider importing low level features to compensate for the lost detailed information in high-level features.Unfortunately, simply combining multi-level features suffers from the semantic gap among them. In this paper, we propose a new architecture, named Gated Fully Fusion (GFF), to selectively fuse features from multiple levels using gates in a fully connected way. Specifically, features at each level are enhanced by higher-level features with stronger semantics and lower-level features with more details, and gates are used to control the propagation of useful information which significantly reduces the noises during fusion. We achieve the state of the art results on four challenging scene parsing datasets including Cityscapes, Pascal Context, COCO-stuff and ADE20K.

cs.CV↗

arXiv2020

Convolutional Neural Network for Convective Storm Nowcasting Using 3D Doppler Weather Radar Data

Convective storms are one of the severe weather hazards found during the warm season. Doppler weather radar is the only operational instrument that can frequently sample the detailed structure of convective storm which has a small spatial scale and short lifetime. For the challenging task of short-term convective storm forecasting, 3-D radar images contain information about the processes in convective storm. However, effectively extracting such information from multisource raw data has been problematic due to a lack of methodology and computation limitations. Recent advancements in deep learning techniques and graphics processing units now make it possible. This article investigates the feasibility and performance of an end-to-end deep learning nowcasting method. The nowcasting problem was transformed into a classification problem first, and then, a deep learning method that uses a convolutional neural network was presented to make predictions. On the first layer of CNN, a cross-channel 3D convolution was proposed to fuse 3D raw data. The CNN method eliminates the handcrafted feature engineering, i.e., the process of using domain knowledge of the data to manually design features. Operationally produced historical data of the Beijing-Tianjin-Hebei region in China was used to train the nowcasting system and evaluate its performance; 3737332 samples were collected in the training data set. The experimental results show that the deep learning method improves nowcasting skills compared with traditional machine learning methods.

physics.geo-ph↗

arXiv2019

Application of Multi-channel 3D-cube Successive Convolution Network for Convective Storm Nowcasting

Convective storm nowcasting has attracted substantial attention in various fields. Existing methods under a deep learning framework rely primarily on radar data. Although they perform nowcast storm advection well, it is still challenging to nowcast storm initiation and growth, due to the limitations of the radar observations. This paper describes the first attempt to nowcast storm initiation, growth, and advection simultaneously under a deep learning framework using multi-source meteorological data. To this end, we present a multi-channel 3D-cube successive convolution network (3D-SCN). As real-time re-analysis meteorological data can now provide valuable atmospheric boundary layer thermal dynamic information, which is essential to predict storm initiation and growth, both raw 3D radar and re-analysis data are used directly without any handcraft feature engineering. These data are formulated as multi-channel 3D cubes, to be fed into our network, which are convolved by cross-channel 3D convolutions. By stacking successive convolutional layers without pooling, we build an end-to-end trainable model for nowcasting. Experimental results show that deep learning methods achieve better performance than traditional extrapolation methods. The qualitative analyses of 3D-SCN show encouraging results of nowcasting of storm initiation, growth, and advection.

cs.CV↗

arXiv2019

A Three-dimensional Convolutional-Recurrent Network for Convective Storm Nowcasting

Very short-term convective storm forecasting, termed nowcasting, has long been an important issue and has attracted substantial interest. Existing nowcasting methods rely principally on radar images and are limited in terms of nowcasting storm initiation and growth. Real-time re-analysis of meteorological data supplied by numerical models provides valuable information about three-dimensional (3D), atmospheric, boundary layer thermal dynamics, such as temperature and wind. To mine such data, we here develop a convolution-recurrent, hybrid deep-learning method with the following characteristics: (1) the use of cell-based oversampling to increase the number of training samples; this mitigates the class imbalance issue; (2) the use of both raw 3D radar data and 3D meteorological data re-analyzed via multi-source 3D convolution without any need for handcraft feature engineering; and (3) the stacking of convolutional neural networks on a long short-term memory encoder/decoder that learns the spatiotemporal patterns of convective processes. Experimental results demonstrated that our method performs better than other extrapolation methods. Qualitative analysis yielded encouraging nowcasting results.

cs.CV↗

arXiv2019

Phrase-Level Class based Language Model for Mandarin Smart Speaker Query Recognition

The success of speech assistants requires precise recognition of a number of entities on particular contexts. A common solution is to train a class-based n-gram language model and then expand the classes into specific words or phrases. However, when the class has a huge list, e.g., more than 20 million songs, a fully expansion will cause memory explosion. Worse still, the list items in the class need to be updated frequently, which requires a dynamic model updating technique. In this work, we propose to train pruned language models for the word classes to replace the slots in the root n-gram. We further propose to use a novel technique, named Difference Language Model (DLM), to correct the bias from the pruned language models. Once the decoding graph is built, we only need to recalculate the DLM when the entities in word classes are updated. Results show that the proposed method consistently and significantly outperforms the conventional approaches on all datasets, esp. for large lists, which the conventional approaches cannot handle.

cs.CL↗

arXiv2019

Arena: a toolkit for Multi-Agent Reinforcement Learning

We introduce Arena, a toolkit for multi-agent reinforcement learning (MARL) research. In MARL, it usually requires customizing observations, rewards and actions for each agent, changing cooperative-competitive agent-interaction, and playing with/against a third-party agent, etc. We provide a novel modular design, called Interface, for manipulating such routines in essentially two ways: 1) Different interfaces can be concatenated and combined, which extends the OpenAI Gym Wrappers concept to MARL scenarios. 2) During MARL training or testing, interfaces can be embedded in either wrapped OpenAI Gym compatible Environments or raw environment compatible Agents. We offer off-the-shelf interfaces for several popular MARL platforms, including StarCraft II, Pommerman, ViZDoom, Soccer, etc. The interfaces effectively support self-play RL and cooperative-competitive hybrid MARL. Also, Arena can be conveniently extended to your own favorite MARL platform.

cs.LG↗

arXiv2018

TStarBots: Defeating the Cheating Level Builtin AI in StarCraft II in the Full Game

Starcraft II (SC2) is widely considered as the most challenging Real Time Strategy (RTS) game. The underlying challenges include a large observation space, a huge (continuous and infinite) action space, partial observations, simultaneous move for all players, and long horizon delayed rewards for local decisions. To push the frontier of AI research, Deepmind and Blizzard jointly developed the StarCraft II Learning Environment (SC2LE) as a testbench of complex decision making systems. SC2LE provides a few mini games such as MoveToBeacon, CollectMineralShards, and DefeatRoaches, where some AI agents have achieved the performance level of human professional players. However, for full games, the current AI agents are still far from achieving human professional level performance. To bridge this gap, we present two full game AI agents in this paper - the AI agent TStarBot1 is based on deep reinforcement learning over a flat action structure, and the AI agent TStarBot2 is based on hard-coded rules over a hierarchical action structure. Both TStarBot1 and TStarBot2 are able to defeat the built-in AI agents from level 1 to level 10 in a full game (1v1 Zerg-vs-Zerg game on the AbyssalReef map), noting that level 8, level 9, and level 10 are cheating agents with unfair advantages such as full vision on the whole map and resource harvest boosting. To the best of our knowledge, this is the first public work to investigate AI agents that can defeat the built-in AI in the StarCraft II full game.

cs.AI↗

arXiv2018

RegNet: Learning the Optimization of Direct Image-to-Image Pose Registration

Direct image-to-image alignment that relies on the optimization of photometric error metrics suffers from limited convergence range and sensitivity to lighting conditions. Deep learning approaches has been applied to address this problem by learning better feature representations using convolutional neural networks, yet still require a good initialization. In this paper, we demonstrate that the inaccurate numerical Jacobian limits the convergence range which could be improved greatly using learned approaches. Based on this observation, we propose a novel end-to-end network, RegNet, to learn the optimization of image-to-image pose registration. By jointly learning feature representation for each pixel and partial derivatives that replace handcrafted ones (e.g., numerical differentiation) in the optimization step, the neural network facilitates end-to-end optimization. The energy landscape is constrained on both the feature representation and the learned Jacobian, hence providing more flexibility for the optimization as a consequence leads to more robust and faster convergence. In a series of experiments, including a broad ablation study, we demonstrate that RegNet is able to converge for large-baseline image pairs with fewer iterations.

cs.CV↗

arXiv2018

Parametrized Deep Q-Networks Learning: Reinforcement Learning with Discrete-Continuous Hybrid Action Space

Most existing deep reinforcement learning (DRL) frameworks consider either discrete action space or continuous action space solely. Motivated by applications in computer games, we consider the scenario with discrete-continuous hybrid action space. To handle hybrid action space, previous works either approximate the hybrid space by discretization, or relax it into a continuous set. In this paper, we propose a parametrized deep Q-network (P- DQN) framework for the hybrid action space without approximation or relaxation. Our algorithm combines the spirits of both DQN (dealing with discrete action space) and DDPG (dealing with continuous action space) by seamlessly integrating them. Empirical results on a simulation example, scoring a goal in simulated RoboCup soccer and the solo mode in game King of Glory (KOG) validate the efficiency and effectiveness of our method.

cs.LG↗

arXiv2018

Local Uncertainty Sampling for Large-Scale Multi-Class Logistic Regression

A major challenge for building statistical models in the big data era is that the available data volume far exceeds the computational capability. A common approach for solving this problem is to employ a subsampled dataset that can be handled by available computational resources. In this paper, we propose a general subsampling scheme for large-scale multi-class logistic regression and examine the variance of the resulting estimator. We show that asymptotically, the proposed method always achieves a smaller variance than that of the uniform random sampling. Moreover, when the classes are conditionally imbalanced, significant improvement over uniform sampling can be achieved. Empirical performance of the proposed method is compared to other methods on both simulated and real-world datasets, and these results match and confirm our theoretical analysis.

stat.CO↗

arXiv2018

Candidates vs. Noises Estimation for Large Multi-Class Classification Problem

This paper proposes a method for multi-class classification problems, where the number of classes K is large. The method, referred to as Candidates vs. Noises Estimation (CANE), selects a small subset of candidate classes and samples the remaining classes. We show that CANE is always consistent and computationally efficient. Moreover, the resulting estimator has low statistical variance approaching that of the maximum likelihood estimator, when the observed label belongs to the selected candidates with high probability. In practice, we use a tree structure with leaves as classes to promote fast beam search for candidate selection. We further apply the CANE method to estimate word probabilities in learning large neural language models. Extensive experimental results show that CANE achieves better prediction accuracy over the Noise-Contrastive Estimation (NCE), its variants and a number of the state-of-the-art tree classifiers, while it gains significant speedup compared to standard O(K) methods.

stat.ML↗

arXiv2018

Jiamusi Pulsar Observations: II. Scintillations of 10 Pulsars

Context. Pulsars scintillate. Dynamic spectra show brightness variation of pulsars in the time and frequency domain. Secondary spectra demonstrate the distribution of fluctuation power in the dynamic spectra. Aims. Dynamic spectra strongly depend on observational frequencies, but were often observed at frequencies lower than 1.5 GHz. Scintillation observations at higher frequencies help to constrain the turbulence feature of the interstellar medium over a wide frequency range and can detect the scintillations of more distant pulsars. Methods. Ten pulsars were observed at 2250 MHz (S-band) with the Jiamusi 66 m telescope to study their scintillations. Their dynamic spectra were first obtained, from which the decorrelation bandwidths and time scales of diffractive scintillation were then derived by autocorrelation. Secondary spectra were calculated by forming the Fourier power spectra of the dynamic spectra. Results. Most of the newly obtained dynamic spectra are at the highest frequency or have the longest time span of any published data for these pulsars. For PSRs B0540+23, B2324+60 and B2351+61, these were the first dynamic spectra ever reported. The frequencydependence of scintillation parameters indicates that the intervening medium can rarely be ideally turbulent with a Kolmogorov spectrum. The thin screen model worked well at S-band for the scintillation of PSR B1933+16. Parabolic arcs were detected in the secondary spectra of three pulsars, PSRs B0355+54, B0540+23 and B2154+40, all of which were asymmetrically distributed. The inverted arclets of PSR B0355+54 were seen to evolve along the main parabola within a continuous observing session of 12 hours, from which the angular velocity of the pulsar was estimated that was consistent with the measurement by very long baseline interferometry (VLBI).

astro-ph.HE↗

arXiv2018

Bayesian Model Averaging with Exponentiated Least Square Loss

The model averaging problem is to average multiple models to achieve a prediction accuracy not much worse than that of the best single model in terms of mean squared error. It is known that if the models are misspecified, model averaging is superior to model selection. Specifically, let $n$ be the sample size, then the worst case regret of the former decays at a rate of $O(1/n)$ while the worst case regret of the latter decays at a rate of $O(1/\sqrt{n})$. The recently proposed $Q$-aggregation algorithm \citep{DaiRigZhang12} solves the model averaging problem with the optimal regret of $O(1/n)$ both in expectation and in deviation; however it suffers from two limitations: (1) for continuous dictionary, the proposed greedy algorithm for solving $Q$-aggregation is not applicable; (2) the formulation of $Q$-aggregation appears ad hoc without clear intuition. This paper examines a different approach to model averaging by considering a Bayes estimator for deviation optimal model averaging by using exponentiated least squares loss. We establish a primal-dual relationship of this estimator and that of $Q$-aggregation and propose new greedy procedures that satisfactorily resolve the above mentioned limitations of $Q$-aggregation.

math.ST↗

arXiv2017

Beyond SIFT using Binary features for Loop Closure Detection

In this paper a binary feature based Loop Closure Detection (LCD) method is proposed, which for the first time achieves higher precision-recall (PR) performance compared with state-of-the-art SIFT feature based approaches. The proposed system originates from our previous work Multi-Index hashing for Loop closure Detection (MILD), which employs Multi-Index Hashing (MIH)~\cite{greene1994multi} for Approximate Nearest Neighbor (ANN) search of binary features. As the accuracy of MILD is limited by repeating textures and inaccurate image similarity measurement, burstiness handling is introduced to solve this problem and achieves considerable accuracy improvement. Additionally, a comprehensive theoretical analysis on MIH used in MILD is conducted to further explore the potentials of hashing methods for ANN search of binary features from probabilistic perspective. This analysis provides more freedom on best parameter choosing in MIH for different application scenarios. Experiments on popular public datasets show that the proposed approach achieved the highest accuracy compared with state-of-the-art while running at 30Hz for databases containing thousands of images.

cs.CV↗

arXiv2017

A Machine Learning Nowcasting Method based on Real-time Reanalysis Data

Despite marked progress over the past several decades, convective storm nowcasting remains a challenge because most nowcasting systems are based on linear extrapolation of radar reflectivity without much consideration for other meteorological fields. The variational Doppler radar analysis system (VDRAS) is an advanced convective-scale analysis system capable of providing analysis of 3-D wind, temperature, and humidity by assimilating Doppler radar observations. Although potentially useful, it is still an open question as to how to use these fields to improve nowcasting. In this study, we present results from our first attempt at developing a Support Vector Machine (SVM) Box-based nOWcasting (SBOW) method under the machine learning framework using VDRAS analysis data. The key design points of SBOW are as follows: 1) The study domain is divided into many position-fixed small boxes and the nowcasting problem is transformed into one question, i.e., will a radar echo > 35 dBZ appear in a box in 30 minutes? 2) Box-based temporal and spatial features, which include time trends and surrounding environmental information, are elaborately constructed, and 3) The box-based constructed features are used to first train the SVM classifier, and then the trained classifier is used to make predictions. Compared with complicated and expensive expert systems, the above design of SBOW allows the system to be small, compact, straightforward, and easy to maintain and expand at low cost. The experimental results show that, although no complicated tracking algorithm is used, SBOW can predict the storm movement trend and storm growth with reasonable skill.

physics.ao-ph↗

arXiv2017

MILD: Multi-Index hashing for Loop closure Detection

Loop Closure Detection (LCD) has been proved to be extremely useful in global consistent visual Simultaneously Localization and Mapping (SLAM) and appearance-based robot relocalization. Methods exploiting binary features in bag of words representation have recently gained a lot of popularity for their efficiency, but suffer from low recall due to the inherent drawback that high dimensional binary feature descriptors lack well-defined centroids. In this paper, we propose a realtime LCD approach called MILD (Multi-Index Hashing for Loop closure Detection), in which image similarity is measured by feature matching directly to achieve high recall without introducing extra computational complexity with the aid of Multi-Index Hashing (MIH). A theoretical analysis of the approximate image similarity measurement using MIH is presented, which reveals the trade-off between efficiency and accuracy from a probabilistic perspective. Extensive comparisons with state-of-the-art LCD methods demonstrate the superiority of MILD in both efficiency and accuracy.

cs.CV↗

arXiv2016

Action2Activity: Recognizing Complex Activities from Sensor Data

As compared to simple actions, activities are much more complex, but semantically consistent with a human's real life. Techniques for action recognition from sensor generated data are mature. However, there has been relatively little work on bridging the gap between actions and activities. To this end, this paper presents a novel approach for complex activity recognition comprising of two components. The first component is temporal pattern mining, which provides a mid-level feature representation for activities, encodes temporal relatedness among actions, and captures the intrinsic properties of activities. The second component is adaptive Multi-Task Learning, which captures relatedness among activities and selects discriminant features. Extensive experiments on a real-world dataset demonstrate the effectiveness of our work.

cs.CV↗

arXiv2015

Structure Group and Fermion-Mass-Term in General Nonlocality

In our previous work [J. Math. Phys. 49, 033513 (2008)] two problems remain to be resolved. One is that we lack a minimal group to replace GL(4,C), the other is that the Equation of Motion (EoM) for fermion has no mass term. After careful investigation we find these two problems are linked by conformal group, a subgroup of GL(4,C) group. The Weyl group, a subgroup of conformal group, can bring about the running of mass, charge etc. while making it responsible for the transformation of interaction vertex. However, once concerning the generation of the mass term in EoM, we have to resort to the whole conformal group, in which the generators $K_μ$ play a crucial role in making vacuum vary from space-like (or light-cone-like)to time-like. Physically the starting points are our previous conclusion, $\vec E^2-\vec B^2\neq 0$ for massive bosons, and the two-photon process yielding $e^+ e^-$ pair. Finally we get to the conclusion that the mass term of strong interaction is linearly relevant to (chromo-)magnetic flux as well as angular momentum.

physics.gen-ph↗