Accurate and unbiased examinations of skin lesions are critical for the early diagnosis and treatment of skin conditions and disorders. Visual features of skin lesions vary significantly because the skin images are collected from patients with different skin colours and morphologies by using dissimilar imaging equipment. Recent studies have reported ensembled convolutional neural networks (CNNs) to classify the images for early diagnosis of skin disorders. However, the practical use of these ensembled CNNs is limited because they are heavyweight and inadequate for using contextual information. Although lightweight networks (e.g., MobileNetV3 and EfficientNet) were developed to achieve parameters reduction for implementing deep neural networks on mobile devices, insufficient depth of feature representation restricts the performance. To address the existing limitations, we introduce a new lite and effective neural network, namely HierAttn. The HierAttn applies a novel strategy to learn the local and global features by using multi-stage and multi-branch attention mechanisms. The efficacy of HierAttn was evaluated by using the dermoscopy images dataset ISIC2019 and smartphone photos dataset PAD-UFES-20. The experimental results show that HierAttn achieves the best top-1 accuracy and AUC among the state-of-the-art lightweight networks. The code is available at https://github.com/anthonyweidai/HierAttn.
Gliomas are one of the most prevalent types of primary brain tumours, accounting for more than 30\% of all cases and they develop from the glial stem or progenitor cells. In theory, the majority of brain tumours could well be identified exclusively by the use of Magnetic Resonance Imaging (MRI). Each MRI modality delivers distinct information on the soft tissue of the human brain and integrating all of them would provide comprehensive data for the accurate segmentation of the glioma, which is crucial for the patient's prognosis, diagnosis, and determining the best follow-up treatment. Unfortunately, MRI is prone to artifacts for a variety of reasons, which might result in missing one or more MRI modalities. Various strategies have been proposed over the years to synthesize the missing modality or compensate for the influence it has on automated segmentation models. However, these methods usually fail to model the underlying missing information. In this paper, we propose a style matching U-Net (SMU-Net) for brain tumour segmentation on MRI images. Our co-training approach utilizes a content and style-matching mechanism to distill the informative features from the full-modality network into a missing modality network. To do so, we encode both full-modality and missing-modality data into a latent space, then we decompose the representation space into a style and content representation. Our style matching module adaptively recalibrates the representation space by learning a matching function to transfer the informative and textural features from a full-modality path into a missing-modality path. Moreover, by modelling the mutual information, our content module surpasses the less informative features and re-calibrates the representation space based on discriminative semantic features. The evaluation process on the BraTS 2018 dataset shows a significant results.
Geometrical structures and the internal local region relationship, such as symmetry, regular array, junction, etc., are essential for understanding a 3D shape. This paper proposes a point cloud feature extraction network named PointSCNet, to capture the geometrical structure information and local region correlation information of a point cloud. The PointSCNet consists of three main modules: the space-filling curve-guided sampling module, the information fusion module, and the channel-spatial attention module. The space-filling curve-guided sampling module uses Z-order curve coding to sample points that contain geometrical correlation. The information fusion module uses a correlation tensor and a set of skip connections to fuse the structure and correlation information. The channel-spatial attention module enhances the representation of key points and crucial feature channels to refine the network. The proposed PointSCNet is evaluated on shape classification and part segmentation tasks. The experimental results demonstrate that the PointSCNet outperforms or is on par with state-of-the-art methods by learning the structure and correlation of point clouds effectively.
Short-term future population prediction is a crucial problem in urban computing. Accurate future population prediction can provide rich insights for urban planners or developers. However, predicting the future population is a challenging task due to its complex spatiotemporal dependencies. Many existing works have attempted to capture spatial correlations by partitioning a city into grids and using Convolutional Neural Networks (CNN). However, CNN merely captures spatial correlations by using a rectangle filter; it ignores urban environmental information such as distribution of railroads and location of POI. Moreover, the importance of those kinds of information for population prediction differs in each region and is affected by contextual situations such as weather conditions and day of the week. To tackle this problem, we propose a novel deep learning model called Attention-based Contextual Multi-View Graph Convolutional Networks (ACMV-GCNs). We first construct multiple graphs based on urban environmental information, and then ACMV-GCNs captures spatial correlations from various views with graph convolutional networks. Further, we add an attention module to consider the contextual situations when leveraging urban environmental information for future population prediction. Using statistics population count data collected through mobile phones, we demonstrate that our proposed model outperforms baseline methods. In addition, by visualizing weights calculated by an attention module, we show that our model learns an efficient way to utilize urban environment information without any prior knowledge.
There has been rapidly growing interests in Automatic Diagnosis (AD) and Automatic Symptom Detection (ASD) systems in the machine learning research literature, aiming to assist doctors in telemedicine services. These systems are designed to interact with patients, collect evidence relevant to their concerns, and make predictions about the underlying diseases. Doctors would review the interaction, including the evidence and the predictions, before making their final decisions. Despite the recent progress, an important piece of doctors' interactions with patients is missing in the design of AD and ASD systems, namely the differential diagnosis. Its absence is largely due to the lack of datasets that include such information for models to train on. In this work, we present a large-scale synthetic dataset that includes a differential diagnosis, along with the ground truth pathology, for each patient. In addition, this dataset includes more pathologies, as well as types of symtoms and antecedents. As a proof-of-concept, we extend several existing AD and ASD systems to incorporate differential diagnosis, and provide empirical evidence that using differentials in training signals is essential for such systems to learn to predict differentials. Dataset available at https://github.com/bruzwen/ddxplus
Dense correspondence between humans carries powerful semantic information that can be utilized to solve fundamental problems for full-body understanding such as in-the-wild surface matching, tracking and reconstruction. In this paper we present BodyMap, a new framework for obtaining high-definition full-body and continuous dense correspondence between in-the-wild images of clothed humans and the surface of a 3D template model. The correspondences cover fine details such as hands and hair, while capturing regions far from the body surface, such as loose clothing. Prior methods for estimating such dense surface correspondence i) cut a 3D body into parts which are unwrapped to a 2D UV space, producing discontinuities along part seams, or ii) use a single surface for representing the whole body, but none handled body details. Here, we introduce a novel network architecture with Vision Transformers that learn fine-level features on a continuous body surface. BodyMap outperforms prior work on various metrics and datasets, including DensePose-COCO by a large margin. Furthermore, we show various applications ranging from multi-layer dense cloth correspondence, neural rendering with novel-view synthesis and appearance swapping.
In recent years, Knowledge Graph (KG) development has attracted significant researches considering the applications in web search, relation prediction, natural language processing, information retrieval, question answering to name a few. However, often KGs are incomplete due to which Knowledge Graph Completion (KGC) has emerged as a sub-domain of research to automatically track down the missing connections in a KG. Numerous strategies have been suggested to work out the KGC dependent on different representation procedures intended to embed triples into a low-dimensional vector space. Given the difficulties related to KGC, researchers around the world are attempting to comprehend the attributes of the problem statement. This study intends to provide an overview of knowledge bases combined with different challenges and their impacts. We discuss existing KGC approaches, including the state-of-the-art Knowledge Graph Embeddings (KGE), not only on static graphs but also for the latest trends such as multimodal, temporal, and uncertain knowledge graphs. In addition, reinforcement learning techniques are reviewed to model complex queries as a link prediction problem. Subsequently, we explored popular software packages for model training and examine open research challenges that can guide future research.
In delay-sensitive industrial internet of things (IIoT) applications, the age of information (AoI) is employed to characterize the freshness of information. Meanwhile, the emerging network function virtualization provides flexibility and agility for service providers to deliver a given network service using a sequence of virtual network functions (VNFs). However, suitable VNF placement and scheduling in these schemes is NP-hard and finding a globally optimal solution by traditional approaches is complex. Recently, deep reinforcement learning (DRL) has appeared as a viable way to solve such problems. In this paper, we first utilize single agent low-complex compound action actor-critic RL to cover both discrete and continuous actions and jointly minimize VNF cost and AoI in terms of network resources under end-to end Quality of Service constraints. To surmount the single-agent capacity limitation for learning, we then extend our solution to a multi-agent DRL scheme in which agents collaborate with each other. Simulation results demonstrate that single-agent schemes significantly outperform the greedy algorithm in terms of average network cost and AoI. Moreover, multi-agent solution decreases the average cost by dividing the tasks between the agents. However, it needs more iterations to be learned due to the requirement on the agents collaboration.
Annotated data have traditionally been used to provide the input for training a supervised machine learning (ML) model. However, current pre-trained ML models for natural language processing (NLP) contain embedded linguistic information that can be used to inform the annotation process. We use the BERT neural language model to feed information back into an annotation task that involves semantic labelling of dialog behavior in a question-asking game called Emotion Twenty Questions (EMO20Q). First we describe the background of BERT, the EMO20Q data, and assisted annotation tasks. Then we describe the methods for fine-tuning BERT for the purpose of checking the annotated labels. To do this, we use the paraphrase task as a way to check that all utterances with the same annotation label are classified as paraphrases of each other. We show this method to be an effective way to assess and revise annotations of textual user data with complex, utterance-level semantic labels.
Despite of the promising results on shape and color recovery using self-supervision, the multi-layer perceptrons-based methods usually costs hours to train the deep neural network due to the implicit surface representation. Moreover, it is quite computational intensive to render a single image, since a forward network inference is required for each pixel. To tackle these challenges, in this paper, we propose an efficient coarse-to-fine approach to recover the textured mesh from multi-view images. Specifically, we take advantage of a differentiable Poisson Solver to represent the shape, which is able to produce topology-agnostic and watertight surfaces. To account for the depth information, we optimize the shape geometry by minimizing the difference between the rendered mesh with the depth predicted by the learning-based multi-view stereo algorithm. In contrast to the implicit neural representation on shape and color, we introduce a physically based inverse rendering scheme to jointly estimate the lighting and reflectance of the objects, which is able to render the high resolution image at real-time. Additionally, we fine-tune the extracted mesh by inverse rendering to obtain the mesh with fine details and high fidelity image. We have conducted the extensive experiments on several multi-view stereo datasets, whose promising results demonstrate the efficacy of our proposed approach. We will make our full implementation publicly available.