Search

Electric field control of magnetism in solids, i.e. the converse magnetoelectricity, is highly desired for applications of scalable energy-efficient logic devices. However, it is not only a technical challenge but also a scientific paradox, since in principle the electric and magnetic degrees of freedom obey distinct rules of symmetries. Despite the great progresses obtained in the community of multiferroics during the past decades, the success of magnetoelectricity remains on its way and more alternative approaches with conceptual revolution are urgently needed. Here, by introducing the concept of topology into multiferroics, an exotic magnetoelectric double-leaf Riemann-surface is unveiled based on the mechanism of spin-dependent $d-p$ hybridization in a two-dimensional magnet: GdI$_2$ monolayer. Protected by the topology, a $180^\circ$ spin reversal can be precisely achieved by an electric cycle, leading to a robust and dissipationless converse magnetoelectric function. Such a topological magnetoelectricity allows the nontrivial manipulation of magnetization by AC electric field. In this category, more candidate materials with better performance are designed targetedly, which pave the road to the potential applications with topological magnetoelectrics., Comment: 7 pages. 5 figures

The Jahn-Teller distortion and the resulting orbital order usually cause some fascinating correlated electronic behaviors, and generally lead to antiferromagnetism in perovskite bulks. Here we demonstrate that the Jahn-Teller distortion present in the perovskite fluoride KCrF$_3$ bulk can be retained to the two-dimensional limit, resulting in a staggered orbital order and ferromagnetism in the perovskite monolayer. Octahedral tilt and rotation distortion also appear in the ground-state structure of the perovskite monolayer, which have minor effects on the electronic and magnetic properties with respect to the Jahn-Teller distortion. In addition, in the prototype phase without structural distortion, the partial occupation of the $e_g$ orbitals leads to a ferromagnetic metallic state. This work facilitates the design of two-dimensional ferromagnets and functional properties based on Jahn-Teller distortion and orbital order, Comment: 8 pages, 5 figures, 1 table

Two-dimensional (2D) ferroelectricity has attracted extensive attention since its discovery in the monolayers of van der Waals materials. Here we show that 2D ferroelectricity induced by octahedral rotation distortion is widely present in the perovskite bilayer system through first-principles calculations. The perovskite tolerance factor plays a crucial role in the lattice dynamics and ground-state structure of the perovskite monolayers and bilayers, thus providing an important indicator for screening this hybrid improper ferroelectricity. Generally, the ferroelectric switching via an orthorhombic twin state has the lowest energy barrier. Epitaxial strain can effectively tune the ferroelectric polarization and ferroelectric switching by changing the amplitude of octahedral rotation and tilt distortion. The increasing compressive strain causes a polar to nonpolar phase transition by suppressing the tilt distortion. The cooperative effect of octahedral distortion at the interface with the substrate can reduce the energy barrier of the reversing rotation mode and can even change the lowest-energy ferroelectric switching path., Comment: 9 pages, 6 figures, 2 tables

Object detection models shipped with camera-equipped edge devices cannot cover the objects of interest for every user. Therefore, the incremental learning capability is a critical feature for a robust and personalized object detection system that many applications would rely on. In this paper, we present an efficient yet practical system, RILOD, to incrementally train an existing object detection model such that it can detect new object classes without losing its capability to detect old classes. The key component of RILOD is a novel incremental learning algorithm that trains end-to-end for one-stage deep object detection models only using training data of new object classes. Specifically to avoid catastrophic forgetting, the algorithm distills three types of knowledge from the old model to mimic the old model's behavior on object classification, bounding box regression and feature extraction. In addition, since the training data for the new classes may not be available, a real-time dataset construction pipeline is designed to collect training images on-the-fly and automatically label the images with both category and bounding box annotations. We have implemented RILOD under both edge-cloud and edge-only setups. Experiment results show that the proposed system can learn to detect a new object class in just a few minutes, including both dataset construction and model training. In comparison, traditional fine-tuning based method may take a few hours for training, and in most cases would also need a tedious and costly manual dataset labeling step., Comment: Camera-ready for ACM/IEEE SEC 2019

Lifelong learning, the problem of continual learning where tasks arrive in sequence, has been lately attracting more attention in the computer vision community. The aim of lifelong learning is to develop a system that can learn new tasks while maintaining the performance on the previously learned tasks. However, there are two obstacles for lifelong learning of deep neural networks: catastrophic forgetting and capacity limitation. To solve the above issues, inspired by the recent breakthroughs in automatically learning good neural network architectures, we develop a Multi-task based lifelong learning via nonexpansive AutoML framework termed Regularize, Expand and Compress (REC). REC is composed of three stages: 1) continually learns the sequential tasks without the learned tasks' data via a newly proposed multi-task weight consolidation (MWC) algorithm; 2) expands the network to help the lifelong learning with potentially improved model capability and performance by network-transformation based AutoML; 3) compresses the expanded model after learning every new task to maintain model efficiency and performance. The proposed MWC and REC algorithms achieve superior performance over other lifelong learning algorithms on four different datasets., Comment: 9 pages, 6 figures

Deep neural networks (DNNs) often suffer from "catastrophic forgetting" during incremental learning (IL) --- an abrupt degradation of performance on the original set of classes when the training objective is adapted to a newly added set of classes. Existing IL approaches tend to produce a model that is biased towards either the old classes or new classes, unless with the help of exemplars of the old data. To address this issue, we propose a class-incremental learning paradigm called Deep Model Consolidation (DMC), which works well even when the original training data is not available. The idea is to first train a separate model only for the new classes, and then combine the two individual models trained on data of two distinct set of classes (old classes and new classes) via a novel double distillation training objective. The two existing models are consolidated by exploiting publicly available unlabeled auxiliary data. This overcomes the potential difficulties due to the unavailability of original training data. Compared to the state-of-the-art techniques, DMC demonstrates significantly better performance in image classification (CIFAR-100 and CUB-200) and object detection (PASCAL VOC 2007) in the single-headed IL setting., Comment: WACV 2020 camera-ready

The key challenge of generative Visual Dialogue (VD) systems is to respond to human queries with informative answers in natural and contiguous conversation flow. Traditional Maximum Likelihood Estimation (MLE)-based methods only learn from positive responses but ignore the negative responses, and consequently tend to yield safe or generic responses. To address this issue, we propose a novel training scheme in conjunction with weighted likelihood estimation (WLE) method. Furthermore, an adaptive multi-modal reasoning module is designed, to accommodate various dialogue scenarios automatically and select relevant information accordingly. The experimental results on the VisDial benchmark demonstrate the superiority of our proposed algorithm over other state-of-the-art approaches, with an improvement of 5.81% on recall@10., Comment: IJCAI 2019

A domain adaptation method for urban scene segmentation is proposed in this work. We develop a fully convolutional tri-branch network, where two branches assign pseudo labels to images in the unlabeled target domain while the third branch is trained with supervision based on images in the pseudo-labeled target domain. The re-labeling and re-training processes alternate. With this design, the tri-branch network learns target-specific discriminative representations progressively and, as a result, the cross-domain capability of the segmenter improves. We evaluate the proposed network on large-scale domain adaptation experiments using both synthetic (GTA) and real (Cityscapes) images. It is shown that our solution achieves the state-of-the-art performance and it outperforms previous methods by a significant margin., Comment: Accepted by ICASSP 2018