arXiv subjects

Huaping Liu

Publications and source records attributed to Huaping Liu.

125 records · Page 7Linked to original sources

arXiv2020

Joint User Identification and Channel Estimation Over Rician Fading Channels

This paper considers crowded massive multiple input multiple output (MIMO) communications over a Rician fading channel, where the number of users is much greater than the number of available pilot sequences. A joint user identification and line-of-sight (LOS) component derivation algorithm is proposed without requiring a threshold. Based on the derived LOS component, we design a LOS-only channel estimator and an updated channel estimator.

cs.IT↗

arXiv2020

TrajectoryNet: a new spatio-temporal feature learning network for human motion prediction

Human motion prediction is an increasingly interesting topic in computer vision and robotics. In this paper, we propose a new 2D CNN based network, TrajectoryNet, to predict future poses in the trajectory space. Compared with most existing methods, our model focuses on modeling the motion dynamics with coupled spatio-temporal features, local-global spatial features and global temporal co-occurrence features of the previous pose sequence. Specifically, the coupled spatio-temporal features describe the spatial and temporal structure information hidden in the natural human motion sequence, which can be mined by covering the space and time dimensions of the input pose sequence with the convolutional filters. The local-global spatial features that encode different correlations of different joints of the human body (e.g. strong correlations between joints of one limb, weak correlations between joints of different limbs) are captured hierarchically by enlarging the receptive field layer by layer and residual connections from the lower layers to the deeper layers in our proposed convolutional network. And the global temporal co-occurrence features represent the co-occurrence relationship that different subsequences in a complex motion sequence are appeared simultaneously, which can be obtained automatically with our proposed TrajectoryNet by reorganizing the temporal information as the depth dimension of the input tensor. Finally, future poses are approximated based on the captured motion dynamics features. Extensive experiments show that our method achieves state-of-the-art performance on three challenging benchmarks (e.g. Human3.6M, G3D, and FNTU), which demonstrates the effectiveness of our proposed method. The code will be available if the paper is accepted.

cs.CV↗

arXiv2020

Energy-based Periodicity Mining with Deep Features for Action Repetition Counting in Unconstrained Videos

Action repetition counting is to estimate the occurrence times of the repetitive motion in one action, which is a relatively new, important but challenging measurement problem. To solve this problem, we propose a new method superior to the traditional ways in two aspects, without preprocessing and applicable for arbitrary periodicity actions. Without preprocessing, the proposed model makes our method convenient for real applications; processing the arbitrary periodicity action makes our model more suitable for the actual circumstance. In terms of methodology, firstly, we analyze the movement patterns of the repetitive actions based on the spatial and temporal features of actions extracted by deep ConvNets; Secondly, the Principal Component Analysis algorithm is used to generate the intuitive periodic information from the chaotic high-dimensional deep features; Thirdly, the periodicity is mined based on the high-energy rule using Fourier transform; Finally, the inverse Fourier transform with a multi-stage threshold filter is proposed to improve the quality of the mined periodicity, and peak detection is introduced to finish the repetition counting. Our work features two-fold: 1) An important insight that deep features extracted for action recognition can well model the self-similarity periodicity of the repetitive action is presented. 2) A high-energy based periodicity mining rule using deep features is presented, which can process arbitrary actions without preprocessing. Experimental results show that our method achieves comparable results on the public datasets YT Segments and QUVA.

cs.CV↗

arXiv2019

Reinforcement Learning from Imperfect Demonstrations under Soft Expert Guidance

In this paper, we study Reinforcement Learning from Demonstrations (RLfD) that improves the exploration efficiency of Reinforcement Learning (RL) by providing expert demonstrations. Most of existing RLfD methods require demonstrations to be perfect and sufficient, which yet is unrealistic to meet in practice. To work on imperfect demonstrations, we first define an imperfect expert setting for RLfD in a formal way, and then point out that previous methods suffer from two issues in terms of optimality and convergence, respectively. Upon the theoretical findings we have derived, we tackle these two issues by regarding the expert guidance as a soft constraint on regulating the policy exploration of the agent, which eventually leads to a constrained optimization problem. We further demonstrate that such problem is able to be addressed efficiently by performing a local linear search on its dual form. Considerable empirical evaluations on a comprehensive collection of benchmarks indicate our method attains consistent improvement over other RLfD counterparts.

cs.LG↗

arXiv2019

Imitation Learning from Observations by Minimizing Inverse Dynamics Disagreement

This paper studies Learning from Observations (LfO) for imitation learning with access to state-only demonstrations. In contrast to Learning from Demonstration (LfD) that involves both action and state supervision, LfO is more practical in leveraging previously inapplicable resources (e.g. videos), yet more challenging due to the incomplete expert guidance. In this paper, we investigate LfO and its difference with LfD in both theoretical and practical perspectives. We first prove that the gap between LfD and LfO actually lies in the disagreement of inverse dynamics models between the imitator and the expert, if following the modeling approach of GAIL. More importantly, the upper bound of this gap is revealed by a negative causal entropy which can be minimized in a model-free way. We term our method as Inverse-Dynamics-Disagreement-Minimization (IDDM) which enhances the conventional LfO method through further bridging the gap to LfD. Considerable empirical results on challenging benchmarks indicate that our method attains consistent improvements over other LfO counterparts.

cs.LG↗

arXiv2019

Deep Point-wise Prediction for Action Temporal Proposal

Detecting actions in videos is an important yet challenging task. Previous works usually utilize (a) sliding window paradigms, or (b) per-frame action scoring and grouping to enumerate the possible temporal locations. Their performances are also limited to the designs of sliding windows or grouping strategies. In this paper, we present a simple and effective method for temporal action proposal generation, named Deep Point-wise Prediction (DPP). DPP simultaneously predicts the action existing possibility and the corresponding temporal locations, without the utilization of any handcrafted sliding window or grouping. The whole system is end-to-end trained with joint loss of temporal action proposal classification and location prediction. We conduct extensive experiments to verify its effectiveness, generality and robustness on standard THUMOS14 dataset. DPP runs more than 1000 frames per second, which largely satisfies the real-time requirement. The code is available at https://github.com/liluxuan1997/DPP.

cs.CV↗

arXiv2019

PISEP^2: Pseudo Image Sequence Evolution based 3D Pose Prediction

Pose prediction is to predict future poses given a window of previous poses. In this paper, we propose a new problem that predicts poses using 3D joint coordinate sequences. Different from the traditional pose prediction based on Mocap frames, this problem is convenient to use in real applications due to its simple sensors to capture data. We also present a new framework, PISEP^2 (Pseudo Image Sequence Evolution based 3D Pose Prediction), to address this new problem. Specifically, a skeletal representation is proposed by transforming the joint coordinate sequence into an image sequence, which can model the different correlations of different joints. With this image based skeletal representation, we model the pose prediction as the evolution of image sequence. Moreover, a novel inference network is proposed to predict all future poses in one step by decoupling the decoders in a non-recursive manner. Compared with the recursive sequence to sequence model, we can improve the computational efficiency and avoid error accumulation significantly. Extensive experiments are carried out on two benchmark datasets (e.g. G3D and FNTU). The proposed method achieves the state-of-the-art performance on both datasets, which demonstrates the effectiveness of our proposed method.

cs.CV↗

arXiv2019

Task Transfer by Preference-Based Cost Learning

The goal of task transfer in reinforcement learning is migrating the action policy of an agent to the target task from the source task. Given their successes on robotic action planning, current methods mostly rely on two requirements: exactly-relevant expert demonstrations or the explicitly-coded cost function on target task, both of which, however, are inconvenient to obtain in practice. In this paper, we relax these two strong conditions by developing a novel task transfer framework where the expert preference is applied as a guidance. In particular, we alternate the following two steps: Firstly, letting experts apply pre-defined preference rules to select related expert demonstrates for the target task. Secondly, based on the selection result, we learn the target cost function and trajectory distribution simultaneously via enhanced Adversarial MaxEnt IRL and generate more trajectories by the learned target distribution for the next preference selection. The theoretical analysis on the distribution learning and convergence of the proposed algorithm are provided. Extensive simulations on several benchmarks have been conducted for further verifying the effectiveness of the proposed method.

cs.LG↗

arXiv2019

Consistent Optimization for Single-Shot Object Detection

We present consistent optimization for single stage object detection. Previous works of single stage object detectors usually rely on the regular, dense sampled anchors to generate hypothesis for the optimization of the model. Through an examination of the behavior of the detector, we observe that the misalignment between the optimization target and inference configurations has hindered the performance improvement. We propose to bride this gap by consistent optimization, which is an extension of the traditional single stage detector's optimization strategy. Consistent optimization focuses on matching the training hypotheses and the inference quality by utilizing of the refined anchors during training. To evaluate its effectiveness, we conduct various design choices based on the state-of-the-art RetinaNet detector. We demonstrate it is the consistent optimization, not the architecture design, that yields the performance boosts. Consistent optimization is nearly cost-free, and achieves stable performance gains independent of the model capacities or input scales. Specifically, utilizing consistent optimization improves RetinaNet from 39.1 AP to 40.1 AP on COCO dataset without any bells or whistles, which surpasses the accuracy of all existing state-of-the-art one-stage detectors when adopting ResNet-101 as backbone. The code will be made available.

cs.CV↗

arXiv2018

Learning and Inferring Movement with Deep Generative Model

Learning and inference movement is a very challenging problem due to its high dimensionality and dependency to varied environments or tasks. In this paper, we propose an effective probabilistic method for learning and inference of basic movements. The motion planning problem is formulated as learning on a directed graphic model and deep generative model is used to perform learning and inference from demonstrations. An important characteristic of this method is that it flexibly incorporates the task descriptors and context information for long-term planning and it can be combined with dynamic systems for robot control. The experimental validations on robotic approaching path planning tasks show the advantages over the base methods with limited training data.

cs.LG↗

arXiv2018

Deep Feature Pyramid Reconfiguration for Object Detection

State-of-the-art object detectors usually learn multi-scale representations to get better results by employing feature pyramids. However, the current designs for feature pyramids are still inefficient to integrate the semantic information over different scales. In this paper, we begin by investigating current feature pyramids solutions, and then reformulate the feature pyramid construction as the feature reconfiguration process. Finally, we propose a novel reconfiguration architecture to combine low-level representations with high-level semantic features in a highly-nonlinear yet efficient way. In particular, our architecture which consists of global attention and local reconfigurations, is able to gather task-oriented features across different spatial locations and scales, globally and locally. Both the global attention and local reconfiguration are lightweight, in-place, and end-to-end trainable. Using this method in the basic SSD system, our models achieve consistent and significant boosts compared with the original model and its other variations, without losing real-time processing speed.

cs.CV↗

arXiv2017

RON: Reverse Connection with Objectness Prior Networks for Object Detection

We present RON, an efficient and effective framework for generic object detection. Our motivation is to smartly associate the best of the region-based (e.g., Faster R-CNN) and region-free (e.g., SSD) methodologies. Under fully convolutional architecture, RON mainly focuses on two fundamental problems: (a) multi-scale object localization and (b) negative sample mining. To address (a), we design the reverse connection, which enables the network to detect objects on multi-levels of CNNs. To deal with (b), we propose the objectness prior to significantly reduce the searching space of objects. We optimize the reverse connection, objectness prior and object detector jointly by a multi-task loss function, thus RON can directly predict final detection results from all locations of various feature maps. Extensive experiments on the challenging PASCAL VOC 2007, PASCAL VOC 2012 and MS COCO benchmarks demonstrate the competitive performance of RON. Specifically, with VGG-16 and low resolution 384X384 input size, the network gets 81.3% mAP on PASCAL VOC 2007, 80.7% mAP on PASCAL VOC 2012 datasets. Its superiority increases when datasets become larger and more difficult, as demonstrated by the results on the MS COCO dataset. With 1.5G GPU memory at test phase, the speed of the network is 15 FPS, 3X faster than the Faster R-CNN counterpart.

cs.CV↗

arXiv2013

Exciton-phonon Bound Complex in Single-walled Carbon Nanotubes Revealed by High-field Magneto-optical Spectroscopy

High-field magneto-optical spectroscopy was conducted on highly-selected chiral (6,5) specific single-walled carbon nanotubes. Spectra of phonon sidebands in both 1st and 2nd sub-bands were observed to be unchanged by the application of an external magnetic field up to 52 T. Our analyses led to the conclusion that both phonon sidebands in respective sub-band originate from the dark K-momentum singlet (D-K-S) excitons. Moreover, while the relative ordering between the bandedge bright exciton and its zero-momentum anti-bonding counterpart was found to be opposite for the 1st and 2nd sub-bands, the relative ordering between the D-K-S exciton and the band-edge bright exciton was clarified to be the same for both sub-bands. Energy of these D-K-S excitons was estimated to be ~ 21.5 and ~ 37.3 meV above the band-edge bright exciton for the 1st and 2nd sub-bands, respectively.

cond-mat.mes-hall↗

arXiv2013

Band-edge Exciton States in a Single-walled Carbon Nanotube Revealed by Magneto-optical Spectroscopy in Ultra-high Magnetic Field

We report high field magneto-optical study on the first and second sub-band transitions of single-chirality single-walled carbon nanotubes. The ordering and relative energy splitting between bright and dark excitonic states were found to be inverse between the first and second subbands. We verified that the zero-momentum dark singlet exciton lies below the bright exciton for the first subband transitions, while for the second sub-band transitions, it was found to have higher energy than the bright excitonic state. Effect of this peculiar excitonic structure was found to manifest itself in distinctive Aharonov-Bohm splitting in ultra-high magnetic fields up to 190 T.

cond-mat.mes-hall↗

arXiv2013

Survey of Exciton-Phonon Sidebands by Magneto-optical Spectroscopy Using Highly Purified (6,5) Single-walled Carbon Nanotubes

We report the first high-field magneto-optical study on the exciton-phonon sideband of single-walled carbon nanotubes (SWNTs) consisting only of (6,5) species. Both energy and intensity of the observed phonon sideband were found to be independent of the external magnetic field. Comparing with theoretical calculations, we confirmed that these sidebands originate from the optically forbidden K-momentum singlet excitons. Energy of these K-momentum dark excitons was extracted to be ~ 21.5 meV above the bright zero-momentum singlet excitons, in close agreement with recent theoretical predictions and experimentally determined values.

cond-mat.mes-hall↗

arXiv2012

Transmission Schemes for Four-Way Relaying in Wireless Cellular Systems

Two-way relaying in wireless systems has initiated a large research effort during the past few years. While one-way relay with a single data flow introduces loss in spectral efficiency due to its half-duplex operation, two-way relaying based on wireless network coding regains part of this loss by simultaneously processing the two data flows. In a broader perspective, the two-way traffic pattern is rather limited and it is of interest to investigate other traffic patterns where such a simultaneous processing of information flows can bring performance advantage. In this paper we consider a scenario beyond the usual two-way relaying: a four-way relaying, where each of the two Mobile Stations (MSs) has a two-way connection to the same Base Station (BS), while each connection is through a dedicated Relay Station (RS). While both RSs are in the range of the same BS, they are assumed to have antipodal positions within the cell, such that they do not interfere with each other. We introduce and analyze a two-phase transmission scheme to serve the four-way traffic pattern defined in this scenario. Each phase consists of combined broadcast and multiple access. We analyze the achievable rate region of the new schemes for two different operational models for the RS, Decode-and-Forward (DF) and Amplify-and-Forward (AF), respectively. We compare the performance with a state-of-the-art reference scheme, time sharing is used between the two MSs, while each MS is served through a two-way relaying scheme. The results indicate that, when the RS operates in a DF mode, the achievable rate regions are significantly enlarged. On the other hand, for AF relaying, the gains are rather modest. The practical implication of the presented work is a novel insight on how to improve the spatial reuse in wireless cellular networks by coordinating the transmissions of the antipodal relays.

cs.IT↗

arXiv2012

Sum-Rate Optimization in a Two-Way Relay Network with Buffering

A Relay Station (RS) uses a buffer to store and process the received data packets before forwarding them. Recently, the buffer has been exploited in one-way relaying to opportunistically schedule the two different links according to their channel quality. The intuition is that, if the channel to the destination is poor, then RS stores more data from the source, in order to use it when the channel to the destination is good. We apply this intuition to the case of half-duplex two-way relaying, where the interactions among the buffers and the links become more complex. We investigate the sum-rate maximization problem in the Time Division Broadcast (TDBC): the users send signals to the RS in different time slots, the RS decodes and stores messages in the buffers. For downlink transmission, the RS re-encodes and sends using the optimal broadcast strategy. The operation in each time slot is not determined in advance, but depends on the channel state information (CSI). We derive the decision function for adaptive link selection with respect to CSI using the Karush-Kuhn-Tucker (KKT) conditions. The thresholds of the decision function are obtained under Rayleigh fading channel conditions. The numerical results show that the sum-rate of the adaptive link selection protocol with buffering is significantly larger compared to the reference protocol with fixed transmission schedule.

cs.IT↗