Towards Efficient Diffusion-Based Image Editing with Instant Attention Masks

Diffusion-based Image Editing (DIE) is an emerging research hot-spot, which often applies a semantic mask to control the target area for diffusion-based editing. However, most existing solutions obtain these masks via manual operations or off-line processing, greatly reducing their efficiency. In this paper, we propose a novel and efficient image editing method for Text-to-Image (T2I) diffusion models, termed Instant Diffusion Editing(InstDiffEdit). In particular, InstDiffEdit aims to employ the cross-modal attention ability of existing diffusion models to achieve instant mask guidance during the diffusion steps. To reduce the noise of attention maps and realize the full automatics, we equip InstDiffEdit with a training-free refinement scheme to adaptively aggregate the attention distributions for the automatic yet accurate mask generation. Meanwhile, to supplement the existing evaluations of DIE, we propose a new benchmark called Editing-Mask to examine the mask accuracy and local editing ability of existing methods. To validate InstDiffEdit, we also conduct extensive experiments on ImageNet and Imagen, and compare it with a bunch of the SOTA methods. The experimental results show that InstDiffEdit not only outperforms the SOTA methods in both image quality and editing results, but also has a much faster inference speed, i.e., +5 to +6 times.

Traffic Video Object Detection using Motion Prior

Traffic videos inherently differ from generic videos in their stationary camera setup, thus providing a strong motion prior where objects often move in a specific direction over a short time interval. Existing works predominantly employ generic video object detection framework for traffic video object detection, which yield certain advantages such as broad applicability and robustness to diverse scenarios. However, they fail to harness the strength of motion prior to enhance detection accuracy. In this work, we propose two innovative methods to exploit the motion prior and boost the performance of both fully-supervised and semi-supervised traffic video object detection. Firstly, we introduce a new self-attention module that leverages the motion prior to guide temporal information integration in the fully-supervised setting. Secondly, we utilise the motion prior to develop a pseudo-labelling mechanism to eliminate noisy pseudo labels for the semi-supervised setting. Both of our motion-prior-centred methods consistently demonstrates superior performance, outperforming existing state-of-the-art approaches by a margin of 2% in terms of mAP.

Can LSH (Locality-Sensitive Hashing) Be Replaced by Neural Network?

With the rapid development of GPU (Graphics Processing Unit) technologies and neural networks, we can explore more appropriate data structures and algorithms. Recent progress shows that neural networks can partly replace traditional data structures. In this paper, we proposed a novel DNN (Deep Neural Network)-based learned locality-sensitive hashing, called LLSH, to efficiently and flexibly map high-dimensional data to low-dimensional space. LLSH replaces the traditional LSH (Locality-sensitive Hashing) function families with parallel multi-layer neural networks, which reduces the time and memory consumption and guarantees query accuracy simultaneously. The proposed LLSH demonstrate the feasibility of replacing the hash index with learning-based neural networks and open a new door for developers to design and configure data organization more accurately to improve information-searching performance. Extensive experiments on different types of datasets show the superiority of the proposed method in query accuracy, time consumption, and memory usage.

Deep Learning and Medical Imaging for COVID-19 Diagnosis: A Comprehensive Survey

COVID-19 (Coronavirus disease 2019) has been quickly spreading since its outbreak, impacting financial markets and healthcare systems globally. Countries all around the world have adopted a number of extraordinary steps to restrict the spreading virus, where early COVID-19 diagnosis is essential. Medical images such as X-ray images and Computed Tomography scans are becoming one of the main diagnostic tools to combat COVID-19 with the aid of deep learning-based systems. In this survey, we investigate the main contributions of deep learning applications using medical images in fighting against COVID-19 from the aspects of image classification, lesion localization, and severity quantification, and review different deep learning architectures and some image preprocessing techniques for achieving a preciser diagnosis. We also provide a summary of the X-ray and CT image datasets used in various studies for COVID-19 detection. The key difficulties and potential applications of deep learning in fighting against COVID-19 are finally discussed. This work summarizes the latest methods of deep learning using medical images to diagnose COVID-19, highlighting the challenges and inspiring more studies to keep utilizing the advantages of deep learning to combat COVID-19.

Exploiting Expert Knowledge for Assigning Firms to Industries: A Novel Deep Learning Method

Industry assignment, which assigns firms to industries according to a predefined Industry Classification System (ICS), is fundamental to a large number of critical business practices, ranging from operations and strategic decision making by firms to economic analyses by government agencies. Three types of expert knowledge are essential to effective industry assignment: definition-based knowledge (i.e., expert definitions of each industry), structure-based knowledge (i.e., structural relationships among industries as specified in an ICS), and assignment-based knowledge (i.e., prior firm-industry assignments performed by domain experts). Existing industry assignment methods utilize only assignment-based knowledge to learn a model that classifies unassigned firms to industries, and overlook definition-based and structure-based knowledge. Moreover, these methods only consider which industry a firm has been assigned to, but ignore the time-specificity of assignment-based knowledge, i.e., when the assignment occurs. To address the limitations of existing methods, we propose a novel deep learning-based method that not only seamlessly integrates the three types of knowledge for industry assignment but also takes the time-specificity of assignment-based knowledge into account. Methodologically, our method features two innovations: dynamic industry representation and hierarchical assignment. The former represents an industry as a sequence of time-specific vectors by integrating the three types of knowledge through our proposed temporal and spatial aggregation mechanisms. The latter takes industry and firm representations as inputs, computes the probability of assigning a firm to different industries, and assigns the firm to the industry with the highest probability.

PixelFolder: An Efficient Progressive Pixel Synthesis Network for Image Generation

Pixel synthesis is a promising research paradigm for image generation, which can well exploit pixel-wise prior knowledge for generation. However, existing methods still suffer from excessive memory footprint and computation overhead. In this paper, we propose a progressive pixel synthesis network towards efficient image generation, coined as PixelFolder. Specifically, PixelFolder formulates image generation as a progressive pixel regression problem and synthesizes images by a multi-stage paradigm, which can greatly reduce the overhead caused by large tensor transformations. In addition, we introduce novel pixel folding operations to further improve model efficiency while maintaining pixel-wise prior knowledge for end-to-end regression. With these innovative designs, we greatly reduce the expenditure of pixel synthesis, e.g., reducing 90% computation and 57% parameters compared to the latest pixel synthesis method called CIPS. To validate our approach, we conduct extensive experiments on two benchmark datasets, namely FFHQ and LSUN Church. The experimental results show that with much less expenditure, PixelFolder obtains new state-of-the-art (SOTA) performance on two benchmark datasets, i.e., 3.77 FID and 2.45 FID on FFHQ and LSUN Church, respectively. Meanwhile, PixelFolder is also more efficient than the SOTA methods like StyleGAN2, reducing about 74% computation and 36% parameters, respectively. These results greatly validate the effectiveness of the proposed PixelFolder.

Reinforcement learning for multi-item retrieval in the puzzle-based storage system

Nowadays, fast delivery services have created the need for high-density warehouses. The puzzle-based storage system is a practical way to enhance the storage density, however, facing difficulties in the retrieval process. In this work, a deep reinforcement learning algorithm, specifically the Double&Dueling Deep Q Network, is developed to solve the multi-item retrieval problem in the system with general settings, where multiple desired items, escorts, and I/O points are placed randomly. Additionally, we propose a general compact integer programming model to evaluate the solution quality. Extensive numerical experiments demonstrate that the reinforcement learning approach can yield high-quality solutions and outperforms three related state-of-the-art heuristic algorithms. Furthermore, a conversion algorithm and a decomposition framework are proposed to handle simultaneous movement and large-scale instances respectively, thus improving the applicability of the PBS system.

Modelling matrix time series via a tensor CP-decomposition

We propose to model matrix time series based on a tensor CP-decomposition. Instead of using an iterative algorithm which is the standard practice for estimating CP-decompositions, we propose a new and one-pass estimation procedure based on a generalized eigenanalysis constructed from the serial dependence structure of the underlying process. A key idea of the new procedure is to project a generalized eigenequation defined in terms of rank-reduced matrices to a lower-dimensional one with full-ranked matrices, to avoid the intricacy of the former of which the number of eigenvalues can be zero, finite and infinity. The asymptotic theory has been established under a general setting without the stationarity. It shows, for example, that all the component coefficient vectors in the CP-decomposition are estimated consistently with the different error rates, depending on the relative sizes between the dimensions of time series and the sample size. The proposed model and the estimation method are further illustrated with both simulated and real data; showing effective dimension-reduction in modelling and forecasting matrix time series.

How to eliminate detour behaviors in E-hailing: On-line detection and Pricing regulation

With the fast development of information and communication technology (ICT), taxi business becomes a typical electronic commerce mode. However, one traditional problem still exists in taxi service, that greedy taxi drivers may deliberately take unnecessary detours to overcharge passengers. The detection of these fraudulent behaviors is essential to ensure high-quality taxi service. In this paper, we propose a novel framework for detecting and analyzing the detour misbehaviors both in off-line database and among on-line trips. Applying our framework to real-world taxi data, a remarkable performance (AUC > 0.98, 100% of detour trajectories can be detected at less than 10% false alarm rate) has been achieved in off-line phases, meanwhile, an excellent precision (AUC > 0.9) also has arrived in on-line detection. In additional, some constructive suggestions upon pricing regulation are also provided to control the happening of detours. Finally, some commercial value-added applications in DiDi benefited from our method have yielded good results to improve the map service.