arXiv subjects

Qiang Fu

Publications and source records attributed to Qiang Fu.

At least 163 records · Page 9Linked to original sources

arXiv2021

Joint Online Multichannel Acoustic Echo Cancellation, Speech Dereverberation and Source Separation

This paper presents a joint source separation algorithm that simultaneously reduces acoustic echo, reverberation and interfering sources. Target speeches are separated from the mixture by maximizing independence with respect to the other sources. It is shown that the separation process can be decomposed into cascading sub-processes that separately relate to acoustic echo cancellation, speech dereverberation and source separation, all of which are solved using the auxiliary function based independent component/vector analysis techniques, and their solving orders are exchangeable. The cascaded solution not only leads to lower computational complexity but also better separation performance than the vanilla joint algorithm.

cs.SD↗

arXiv2021

Mask-ToF: Learning Microlens Masks for Flying Pixel Correction in Time-of-Flight Imaging

We introduce Mask-ToF, a method to reduce flying pixels (FP) in time-of-flight (ToF) depth captures. FPs are pervasive artifacts which occur around depth edges, where light paths from both an object and its background are integrated over the aperture. This light mixes at a sensor pixel to produce erroneous depth estimates, which can adversely affect downstream 3D vision tasks. Mask-ToF starts at the source of these FPs, learning a microlens-level occlusion mask which effectively creates a custom-shaped sub-aperture for each sensor pixel. This modulates the selection of foreground and background light mixtures on a per-pixel basis and thereby encodes scene geometric information directly into the ToF measurements. We develop a differentiable ToF simulator to jointly train a convolutional neural network to decode this information and produce high-fidelity, low-FP depth reconstructions. We test the effectiveness of Mask-ToF on a simulated light field dataset and validate the method with an experimental prototype. To this end, we manufacture the learned amplitude mask and design an optical relay system to virtually place it on a high-resolution ToF sensor. We find that Mask-ToF generalizes well to real data without retraining, cutting FP counts in half.

eess.IV↗

arXiv2021

Weighted Recursive Least Square Filter and Neural Network based Residual Echo Suppression for the AEC-Challenge

This paper presents a real-time Acoustic Echo Cancellation (AEC) algorithm submitted to the AEC-Challenge. The algorithm consists of three modules: Generalized Cross-Correlation with PHAse Transform (GCC-PHAT) based time delay compensation, weighted Recursive Least Square (wRLS) based linear adaptive filtering and neural network based residual echo suppression. The wRLS filter is derived from a novel semi-blind source separation perspective. The neural network model predicts a Phase-Sensitive Mask (PSM) based on the aligned reference and the linear filter output. The algorithm achieved a mean subjective score of 4.00 and ranked 2nd in the AEC-Challenge.

cs.SD↗

arXiv2021

Multi-Agent Deep Reinforcement Learning for Request Dispatching in Distributed-Controller Software-Defined Networking

Recently, distributed controller architectures have been quickly gaining popularity in Software-Defined Networking (SDN). However, the use of distributed controllers introduces a new and important Request Dispatching (RD) problem with the goal for every SDN switch to properly dispatch their requests among all controllers so as to optimize network performance. This goal can be fulfilled by designing an RD policy to guide distribution of requests at each switch. In this paper, we propose a Multi-Agent Deep Reinforcement Learning (MA-DRL) approach to automatically design RD policies with high adaptability and performance. This is achieved through a new problem formulation in the form of a Multi-Agent Markov Decision Process (MA-MDP), a new adaptive RD policy design and a new MA-DRL algorithm called MA-PPO. Extensive simulation studies show that our MA-DRL technique can effectively train RD policies to significantly outperform man-made policies, model-based policies, as well as RD policies learned via single-agent DRL algorithms.

cs.NI↗

arXiv2020

Towards Playing Full MOBA Games with Deep Reinforcement Learning

MOBA games, e.g., Honor of Kings, League of Legends, and Dota 2, pose grand challenges to AI systems such as multi-agent, enormous state-action space, complex action control, etc. Developing AI for playing MOBA games has raised much attention accordingly. However, existing work falls short in handling the raw game complexity caused by the explosion of agent combinations, i.e., lineups, when expanding the hero pool in case that OpenAI's Dota AI limits the play to a pool of only 17 heroes. As a result, full MOBA games without restrictions are far from being mastered by any existing AI system. In this paper, we propose a MOBA AI learning paradigm that methodologically enables playing full MOBA games with deep reinforcement learning. Specifically, we develop a combination of novel and existing learning techniques, including curriculum self-play learning, policy distillation, off-policy adaption, multi-head value estimation, and Monte-Carlo tree-search, in training and playing a large pool of heroes, meanwhile addressing the scalability issue skillfully. Tested on Honor of Kings, a popular MOBA game, we show how to build superhuman AI agents that can defeat top esports players. The superiority of our AI is demonstrated by the first large-scale performance test of MOBA AI agent in the literature.

cs.AI↗

arXiv2020

Optimal Ternary Linear Complementary Dual Codes

Linear complementary dual (LCD) codes introduced by Massey are the codes whose intersections with their dual codes are trivial. It can help to improve the security of the information processed by sensitive devices, especially against side-channel attacks (SCA) and fault invasive attacks. In this paper, By construction of puncturing, extending, shortening and combination codes, many good ternary LCD codes are presented. We give a Table 1 with the values of $d_{LCD}(n,k)$ for length $ n \leq 20$. In addition, Many of these ternary LCD codes given in this paper are optimal which are saturating the lower or upper bound of Grassl's codetable in \cite{Grassl} and some of them are nearly optimal.

cs.IT↗

arXiv2020

Mastering Complex Control in MOBA Games with Deep Reinforcement Learning

We study the reinforcement learning problem of complex action control in the Multi-player Online Battle Arena (MOBA) 1v1 games. This problem involves far more complicated state and action spaces than those of traditional 1v1 games, such as Go and Atari series, which makes it very difficult to search any policies with human-level performance. In this paper, we present a deep reinforcement learning framework to tackle this problem from the perspectives of both system and algorithm. Our system is of low coupling and high scalability, which enables efficient explorations at large scale. Our algorithm includes several novel strategies, including control dependency decoupling, action mask, target attention, and dual-clip PPO, with which our proposed actor-critic network can be effectively trained in our system. Tested on the MOBA game Honor of Kings, our AI agent, called Tencent Solo, can defeat top professional human players in full 1v1 games.

cs.AI↗

arXiv2020

Supervised Learning Achieves Human-Level Performance in MOBA Games: A Case Study of Honor of Kings

We present JueWu-SL, the first supervised-learning-based artificial intelligence (AI) program that achieves human-level performance in playing multiplayer online battle arena (MOBA) games. Unlike prior attempts, we integrate the macro-strategy and the micromanagement of MOBA-game-playing into neural networks in a supervised and end-to-end manner. Tested on Honor of Kings, the most popular MOBA at present, our AI performs competitively at the level of High King players in standard 5v5 games.

cs.AI↗

arXiv2020

Revisiting the Quasi Biennial Oscillation as Seen in ERA5. Part I: Description and Momentum Budget

The dynamics and momentum budget of the quasi-biennial oscillation (QBO) are examined in the ERA5 reanalysis and compared with those in ERA-I. Because of ERA5's higher spatial resolution it is capable of resolving a broader spectrum of atmospheric waves and allows for a better representation of the wave-mean flow interactions, both of which are of crucial importance for QBO studies. It is shown that the QBO-induced mean meridional circulation, which is mainly confined to the winter hemisphere, is strong enough to interrupt the tropical upwelling during the descent of the westerly shear zones. Since the momentum advection tends to damp the QBO, the wave forcing is responsible for both the downward propagation and for the maintenance of the QBO. It is shown that half the required wave forcing is provided by resolved waves during the descent of both westerly and easterly regimes. Planetary-scale waves account for most of the resolved wave forcing of the descent of westerly shear zones and small-scale gravity (SSG) waves with wavelengths shorter than 2000 km account for the remainder. SSG waves account for most of the resolved forcing of the descent of the easterly shear zones. The representation of the mean fields in the QBO is very similar in ERA5 and ERA-I but the resolved wave forcing is shown to be substantially stronger in ERA5. The contributions of the various equatorially-trapped wave modes to the QBO forcing are documented in Part II.

physics.ao-ph↗

arXiv2020

Revisiting the Quasi Biennial Oscillation as Seen in ERA5.Part II: Evaluation of Waves and Wave Forcing

This paper describes stratospheric waves in ERA5 reanalysis and evaluates the contributions of different types of waves to the driving of the quasi-biennial oscillation (QBO). Because of its higher spatial resolution compared to its predecessors, ERA5 is capable of resolving a broader spectrum of waves. It is shown that the resolved waves contribute to both eastward and westward accelerations near the equator, mainly by the way of the vertical flux of zonal momentum. The eastward accelerations by the resolved waves are mainly due to Kelvin waves and small-scale gravity (SSG) waves with zonal wavelengths smaller than 2000 km, whereas the westward accelerations are forced mainly by SSG waves, with smaller contributions from inertio-gravity and mixed-Rossby-gravity waves. Extratropical Rossby waves propagate into the tropical region and impart a westward acceleration to the zonal flow. They appear to be responsible for at least some of the irregularities in the QBO cycle.

physics.ao-ph↗

arXiv2020

The Lunar Lander Neutron and Dosimetry (LND) Experiment on Chang'E 4

Chang'E 4 is the first mission to the far side of the Moon and consists of a lander, a rover, and a relay spacecraft. Lander and rover were launched at 18:23 UTC on December 7, 2018 and landed in the von Kármán crater at 02:26 UTC on January 3, 2019. Here we describe the Lunar Lander Neutron \& Dosimetry experiment (LND) which is part of the Chang'E 4 Lander scientific payload. Its chief scientific goal is to obtain first active dosimetric measurements on the surface of the Moon. LND also provides observations of fast neutrons which are a result of the interaction of high-energy particle radiation with the lunar regolith and of their thermalized counterpart, thermal neutrons, which are a sensitive indicator of subsurface water content.

astro-ph.IM↗

arXiv2020

Interlayer Decoupling in 30° Twisted Bilayer Graphene Quasicrystal

Stacking order has strong influence on the coupling between the two layers of twisted bilayer graphene (BLG), which in turn determines its physical properties. Here, we report the investigation of the interlayer coupling of the epitaxially grown single-crystal 30° twisted BLG on Cu(111) at the atomic scale. The stacking order and morphology of BLG is controlled by a rationally designed two-step growth process, that is, the thermodynamically controlled nucleation and kinetically controlled growth. The crystal structure of the 30°-twisted bilayer graphene (30°-tBLG) is determined to have the quasicrystal like symmetry. The electronic properties and interlayer coupling of the 30°-tBLG is investigated using scanning tunneling microscopy (STM) and spectroscopy (STS). The energy-dependent local density of states (DOS) with in-situ electrostatic doping shows that the electronic states in two graphene layers are decoupled near the Dirac point. A linear dispersion originated from the constituent graphene monolayers is discovered with doubled degeneracy. This study contributes to controlled growth of twist-angle-defined BLG, and provides insights of the electronic properties and interlayer coupling in this intriguing system.

cond-mat.mes-hall↗

arXiv2019

Pion Valence Quark Distributions from Maximum Entropy Method

Valence quark distributions of pion at very low resolution scale $Q^{2}_0 \sim 0.1~GeV^2$ are deduced from a maximum entropy method, under the assumption that pion consists of only a valence quark and a valence anti-quark at such a low scale. Taking the obtained initial quark distributions as the nonperturbative input in the modified Dokshitzer-Gribov-Lipatov-Altarelli-Parisi (with the GLR-MQ-ZRS corrections) evolution, the generated valence quark distribution functions at high $Q^2$ are consistent with the measured ones from a Drell-Yan experiment. The maximum entropy method is also applied to estimate the valence quark distributions at relatively higher $Q^2$ = 0.26 GeV$^{2}$. At this higher scale, other components (sea quarks and gluons) should be considered in order to match the experimental data. The first three moments of pion quark distributions at high $Q^2$ are calculated and compared with the other theoretical predictions.

hep-ph↗

arXiv2019

Real time backbone for semantic segmentation

The rapid development of autonomous driving in recent years presents lots of challenges for scene understanding. As an essential step towards scene understanding, semantic segmentation thus received lots of attention in past few years. Although deep learning based state-of-the-arts have achieved great success in improving the segmentation accuracy, most of them suffer from an inefficiency problem and can hardly applied to practical applications. In this paper, we systematically analyze the computation cost of Convolutional Neural Network(CNN) and found that the inefficiency of CNN is mainly caused by its wide structure rather than the deep structure. In addition, the success of pruning based model compression methods proved that there are many redundant channels in CNN. Thus, we designed a very narrow while deep backbone network to improve the efficiency of semantic segmentation. By casting our network to FCN32 segmentation architecture, the basic structure of most segmentation methods, we achieved 60.6\% mIoU on Cityscape val dataset with 54 frame per seconds(FPS) on $1024\times2048$ inputs, which already outperforms one of the earliest real time deep learning based segmentation methods: ENet.

cs.CV↗

arXiv2019

Optimizing Controller Placement for Software-Defined Networks

Controller placement problem (CPP) is a key issue for Software-Defined Networking (SDN) with distributed controller architectures. This problem aims to determine a suitable number of controllers deployed in important locations so as to optimize the overall network performance. In comparison to communication delay, existing literature on the CPP assumes that the influence of controller workload distribution on network performance is negligible. In this paper, we tackle the CPP that simultaneously considers the communication delay, the control plane utilization, and the controller workload distribution. Due to this reason, our CPP is intrinsically different from and clearly more difficult than any previously studied CPPs that are NP-hard. To tackle this challenging issue, we develop a new algorithm that seamlessly integrates the genetic algorithm (GA) and the gradient descent (GD) optimization method. Particularly, GA is used to search for suitable CPP solutions. The quality of each solution is further evaluated through GD. Simulation results on two representative network scenarios (small-scale and large-scale) show that our algorithm can effectively strike the trade-off between the control plane utilization and the network response time.

cs.NI↗

arXiv2019

Federated Collaborative Filtering for Privacy-Preserving Personalized Recommendation System

The increasing interest in user privacy is leading to new privacy preserving machine learning paradigms. In the Federated Learning paradigm, a master machine learning model is distributed to user clients, the clients use their locally stored data and model for both inference and calculating model updates. The model updates are sent back and aggregated on the server to update the master model then redistributed to the clients. In this paradigm, the user data never leaves the client, greatly enhancing the user' privacy, in contrast to the traditional paradigm of collecting, storing and processing user data on a backend server beyond the user's control. In this paper we introduce, as far as we are aware, the first federated implementation of a Collaborative Filter. The federated updates to the model are based on a stochastic gradient approach. As a classical case study in machine learning, we explore a personalized recommendation system based on users' implicit feedback and demonstrate the method's applicability to both the MovieLens and an in-house dataset. Empirical validation confirms a collaborative filter can be federated without a loss of accuracy compared to a standard implementation, hence enhancing the user's privacy in a widely used recommender application while maintaining recommender performance.

cs.IR↗

arXiv2018

Hierarchical Macro Strategy Model for MOBA Game AI

The next challenge of game AI lies in Real Time Strategy (RTS) games. RTS games provide partially observable gaming environments, where agents interact with one another in an action space much larger than that of GO. Mastering RTS games requires both strong macro strategies and delicate micro level execution. Recently, great progress has been made in micro level execution, while complete solutions for macro strategies are still lacking. In this paper, we propose a novel learning-based Hierarchical Macro Strategy model for mastering MOBA games, a sub-genre of RTS games. Trained by the Hierarchical Macro Strategy model, agents explicitly make macro strategy decisions and further guide their micro level execution. Moreover, each of the agents makes independent strategy decisions, while simultaneously communicating with the allies through leveraging a novel imitated cross-agent communication mechanism. We perform comprehensive evaluations on a popular 5v5 Multiplayer Online Battle Arena (MOBA) game. Our 5-AI team achieves a 48% winning rate against human player teams which are ranked top 1% in the player ranking system.

cs.MA↗

arXiv2018

Polaron-induced band renormalization due to linear and quadratic electron-phonon coupling

We present a novel approach to electron-lattice interaction beyond the linear-coupling regime. Based on the solution of a Holstein-Peierls-type model, we derive explicit analytical expressions for the eigenvalue spectrum of the Hamiltonian, resulting in a narrowing of bands as a function of temperature. Our approach enables the intuitive interpretation in terms of quasiparticles, i.e. polaron bands and dressed-phonon frequencies. Being nonperturbative, the formalism also applies in the strong-coupling case. We apply it to the organic crystal naphthalene, with the coupling strengths obtained by \textit{ab initio} calculations.

cond-mat.mtrl-sci↗