arXiv subjects

Tianyi Li

Publications and source records attributed to Tianyi Li.

137 records · Page 8Linked to original sources

arXiv2019

Self-falsifiable Hierarchical Detection of Overlapping Communities On Social Networks

No community detection algorithm can be optimal for all possible networks, thus it is important to identify whether the algorithm is suitable for a given network. We propose a multi-step algorithmic solution scheme for overlapping community detection based on an advanced label propagation process, which imitates the community formation process on social networks. Our algorithm is parameter-free and is able to reveal the hierarchical order of communities in the graph. The unique property of our solution scheme is self-falsifiability; an automatic quality check of the results is conducted after the detection, and the fitness of the algorithm for the specific network is reported. Extensive experiments show that our algorithm is self-consistent, reliable on networks of a wide range of size and different sorts, and is more robust than existing algorithms on both sparse and large-scale social networks. Results further suggest that our solution scheme may uncover features of networks' intrinsic community structures.

cs.SI↗

arXiv2019

Incorporating Textual Evidence in Visual Storytelling

Previous work on visual storytelling mainly focused on exploring image sequence as evidence for storytelling and neglected textual evidence for guiding story generation. Motivated by human storytelling process which recalls stories for familiar images, we exploit textual evidence from similar images to help generate coherent and meaningful stories. To pick the images which may provide textual experience, we propose a two-step ranking method based on image object recognition techniques. To utilize textual information, we design an extended Seq2Seq model with two-channel encoder and attention. Experiments on the VIST dataset show that our method outperforms state-of-the-art baseline models without heavy engineering.

cs.CL↗

arXiv2019

RFBNet: Deep Multimodal Networks with Residual Fusion Blocks for RGB-D Semantic Segmentation

RGB-D semantic segmentation methods conventionally use two independent encoders to extract features from the RGB and depth data. However, there lacks an effective fusion mechanism to bridge the encoders, for the purpose of fully exploiting the complementary information from multiple modalities. This paper proposes a novel bottom-up interactive fusion structure to model the interdependencies between the encoders. The structure introduces an interaction stream to interconnect the encoders. The interaction stream not only progressively aggregates modality-specific features from the encoders but also computes complementary features for them. To instantiate this structure, the paper proposes a residual fusion block (RFB) to formulate the interdependences of the encoders. The RFB consists of two residual units and one fusion unit with gate mechanism. It learns complementary features for the modality-specific encoders and extracts modality-specific features as well as cross-modal features. Based on the RFB, the paper presents the deep multimodal networks for RGB-D semantic segmentation called RFBNet. The experiments on two datasets demonstrate the effectiveness of modeling the interdependencies and that the RFBNet achieved state-of-the-art performance.

cs.CV↗

arXiv2019

Restricted Deformable Convolution based Road Scene Semantic Segmentation Using Surround View Cameras

Understanding the surrounding environment of the vehicle is still one of the challenges for autonomous driving. This paper addresses 360-degree road scene semantic segmentation using surround view cameras, which are widely equipped in existing production cars. First, in order to address large distortion problem in the fisheye images, Restricted Deformable Convolution (RDC) is proposed for semantic segmentation, which can effectively model geometric transformations by learning the shapes of convolutional filters conditioned on the input feature map. Second, in order to obtain a large-scale training set of surround view images, a novel method called zoom augmentation is proposed to transform conventional images to fisheye images. Finally, an RDC based semantic segmentation model is built; the model is trained for real-world surround view images through a multi-task learning architecture by combining real-world images with transformed images. Experiments demonstrate the effectiveness of the RDC to handle images with large distortions, and that the proposed approach shows a good performance using surround view cameras with the help of the transformed images.

cs.CV↗

arXiv2019

Josephson penetration depth in coplanar junctions based on 2D materials

Josephson junctions and SQUIDs with graphene or other 2D materials as the weak link between superconductors have become a hot topic of research in recent years, with respect to both fundamental physics and potential applications. We have previously reported ultra-wide Josephson junctions (up to 80 μm wide) based on CVD graphene where the critical current was found to be uniformly distributed in the direction perpendicular to the current. In this paper, we demonstrate that the unusually large Josephson penetration depth λ_J that this corresponds to is enabled by the unique geometric structure of Josephson junctions based on 2D materials. We derive a new expression for the Josephson penetration depth of such junctions and verify our assumptions by numerical simulations.

cond-mat.supr-con↗

arXiv2019

A DenseNet Based Approach for Multi-Frame In-Loop Filter in HEVC

High efficiency video coding (HEVC) has brought outperforming efficiency for video compression. To reduce the compression artifacts of HEVC, we propose a DenseNet based approach as the in-loop filter of HEVC, which leverages multiple adjacent frames to enhance the quality of each encoded frame. Specifically, the higher-quality frames are found by a reference frame selector (RFS). Then, a deep neural network for multi-frame in-loop filter (named MIF-Net) is developed to enhance the quality of each encoded frame by utilizing the spatial information of this frame and the temporal information of its neighboring higher-quality frames. The MIF-Net is built on the recently developed DenseNet, benefiting from the improved generalization capacity and computational efficiency. Finally, experimental results verify the effectiveness of our multi-frame in-loop filter, outperforming the HM baseline and other state-of-the-art approaches.

cs.CV↗

arXiv2018

Residential Energy Storage Management with Bidirectional Energy Control

We consider the residential energy storage management system with integrated renewable generation, with the availability of bidirectional energy flow from and to the grid thorough buying and selling. We propose a real-time bidirectional energy control algorithm, aiming to minimize the net system cost, due to energy buying and selling and battery deterioration and inefficiency from storage activities, within a given time period, subject to the battery operational constraints and energy buying and selling constraints. We formulate the problem as a stochastic control optimization problem. We then modify and transform this difficult problem into one that enables us to develop the real-time energy control algorithm through Lyapunov optimization. Our developed algorithm is applicable to arbitrary and unknown statistics of renewable generation, load, and electricity prices. It provides a simple closed-form control solution only based on current system states with minimum complexity for real-time implementation. Furthermore, the solution structure reveals how the battery energy level and energy prices affect the decision on energy flow and storage. The proposed algorithm possesses a bounded performance guarantee to that of the optimal non-causal T-slot look-ahead control policy. Simulation shows the effectiveness of our proposed algorithm as compared with alternative real-time and non-causal algorithms, as well as the effect of selling-to-buying price ratio and battery inefficiency on the storage behavior and system cost.

eess.SY↗

arXiv2018

Reducing Complexity of HEVC: A Deep Learning Approach

High Efficiency Video Coding (HEVC) significantly reduces bit-rates over the proceeding H.264 standard but at the expense of extremely high encoding complexity. In HEVC, the quad-tree partition of coding unit (CU) consumes a large proportion of the HEVC encoding complexity, due to the bruteforce search for rate-distortion optimization (RDO). Therefore, this paper proposes a deep learning approach to predict the CU partition for reducing the HEVC complexity at both intra- and inter-modes, which is based on convolutional neural network (CNN) and long- and short-term memory (LSTM) network. First, we establish a large-scale database including substantial CU partition data for HEVC intra- and inter-modes. This enables deep learning on the CU partition. Second, we represent the CU partition of an entire coding tree unit (CTU) in the form of a hierarchical CU partition map (HCPM). Then, we propose an early-terminated hierarchical CNN (ETH-CNN) for learning to predict the HCPM. Consequently, the encoding complexity of intra-mode HEVC can be drastically reduced by replacing the brute-force search with ETH-CNN to decide the CU partition. Third, an early-terminated hierarchical LSTM (ETH-LSTM) is proposed to learn the temporal correlation of the CU partition. Then, we combine ETH-LSTM and ETH-CNN to predict the CU partition for reducing the HEVC complexity for inter-mode. Finally, experimental results show that our approach outperforms other state-of-the-art approaches in reducing the HEVC complexity at both intra- and inter-modes.

cs.CV↗

arXiv2018

Multi-Frame Quality Enhancement for Compressed Video

The past few years have witnessed great success in applying deep learning to enhance the quality of compressed image/video. The existing approaches mainly focus on enhancing the quality of a single frame, ignoring the similarity between consecutive frames. In this paper, we investigate that heavy quality fluctuation exists across compressed video frames, and thus low quality frames can be enhanced using the neighboring high quality frames, seen as Multi-Frame Quality Enhancement (MFQE). Accordingly, this paper proposes an MFQE approach for compressed video, as a first attempt in this direction. In our approach, we firstly develop a Support Vector Machine (SVM) based detector to locate Peak Quality Frames (PQFs) in compressed video. Then, a novel Multi-Frame Convolutional Neural Network (MF-CNN) is designed to enhance the quality of compressed video, in which the non-PQF and its nearest two PQFs are as the input. The MF-CNN compensates motion between the non-PQF and PQFs through the Motion Compensation subnet (MC-subnet). Subsequently, the Quality Enhancement subnet (QE-subnet) reduces compression artifacts of the non-PQF with the help of its nearest PQFs. Finally, the experiments validate the effectiveness and generality of our MFQE approach in advancing the state-of-the-art quality enhancement of compressed video. The code of our MFQE approach is available at https://github.com/ryangBUAA/MFQE.git

cs.CV↗

arXiv2017

Ballistic Josephson junctions based on CVD graphene

Josephson junctions with graphene as the weak link between superconductors have been intensely studied in recent years, with respect to both fundamental physics and potential applications. However, most of the previous work was based on mechanically exfoliated graphene, which is not compatible with mass production. Here we present our research using graphene grown by chemical vapour deposition (CVD) as the weak link of Josephson junctions. We demonstrate that CVD-graphene-based Josephson junctions with Nb electrodes can work effectively without any thermal hysteresis from 1.5 K down to a base temperature of 320 mK, and they show an ideal Fraunhofer-like interference pattern in a perpendicular magnetic field. We also show that the critical current of the junction can be tuned by a gate voltage. Furthermore, for our shortest junctions (50 nm in length), we find that the normal state resistance oscillates with the gate voltage, indicating that the junctions are in the ballistic regime, a feature not previously observed in CVD-graphene-based Josephson junctions.

cond-mat.supr-con↗

arXiv2016

Real-Time Residential-Side Joint Energy Storage Management and Load Scheduling with Renewable Integration

We consider joint energy storage management and load scheduling at a residential site with integrated renewable generation. Assuming unknown arbitrary dynamics of renewable source, loads, and electricity price, we aim at optimizing the load scheduling and energy storage control simultaneously in order to minimize the overall system cost within a finite time period. Besides incorporating battery operational constraints and costs, we model each individual load task by its requested power intensity and service durations, as well as the maximum and average delay requirements. To tackle this finite time horizon stochastic problem, we propose a real-time scheduling and storage control solution by applying a sequence of modification and transformation to employ Lyapunov optimization that otherwise is not directly applicable. With our proposed algorithm, we show that the joint load scheduling and energy storage control can in fact be separated and sequentially determined. Furthermore, both scheduling and energy control decisions have closed-form solutions for simple implementation. Through analysis, we show that our proposed real-time algorithm has a bounded performance guarantee from the optimal T-slot look-ahead solution and is asymptotically equivalent to it as the battery capacity and time period goes to infinity. The effectiveness of joint load scheduling and energy storage control by our proposed algorithm is demonstrated through simulation as compared with alternative algorithms.

eess.SY↗