In recent years, Graph Neural Networks (GNNs) have shown superior performance on diverse real-world applications. To improve the model capacity, besides designing aggregation operations, GNN topology design is also very important. In general, there are two mainstream GNN topology design manners. The first one is to stack aggregation operations to obtain the higher-level features but easily got performance drop as the network goes deeper. Secondly, the multiple aggregation operations are utilized in each layer which provides adequate and independent feature extraction stage on local neighbors while are costly to obtain the higher-level information. To enjoy the benefits while alleviating the corresponding deficiencies of these two manners, we learn to design the topology of GNNs in a novel feature fusion perspective which is dubbed F$^2$GNN. To be specific, we provide a feature fusion perspective in designing GNN topology and propose a novel framework to unify the existing topology designs with feature selection and fusion strategies. Then we develop a neural architecture search method on top of the unified framework which contains a set of selection and fusion operations in the search space and an improved differentiable search algorithm. The performance gains on eight real-world datasets demonstrate the effectiveness of F$^2$GNN. We further conduct experiments to show that F$^2$GNN can improve the model capacity while alleviating the deficiencies of existing GNN topology design manners, especially alleviating the over-smoothing problem, by utilizing different levels of features adaptively.
In recent years, Graph Neural Networks (GNNs) have shown superior performance on diverse applications on real-world datasets. To improve the model capacity and alleviate the over-smoothing problem, several methods proposed to incorporate the intermediate layers by layer-wise connections. However, due to the highly diverse graph types, the performance of existing methods vary on diverse graphs, leading to a need for data-specific layer-wise connection methods. To address this problem, we propose a novel framework LLC (Learn Layer-wise Connections) based on neural architecture search (NAS) to learn adaptive connections among intermediate layers in GNNs. LLC contains one novel search space which consists of 3 types of blocks and learnable connections, and one differentiable search algorithm to enable the efficient search process. Extensive experiments on five real-world datasets are conducted, and the results show that the searched layer-wise connections can not only improve the performance but also alleviate the over-smoothing problem.
Transformer, an attention-based encoder-decoder architecture, has revolutionized the field of natural language processing. Inspired by this significant achievement, some pioneering works have recently been done on adapting Transformerliked architectures to Computer Vision (CV) fields, which have demonstrated their effectiveness on various CV tasks. Relying on competitive modeling capability, visual Transformers have achieved impressive performance on multiple benchmarks such as ImageNet, COCO, and ADE20k as compared with modern Convolution Neural Networks (CNN). In this paper, we have provided a comprehensive review of over one hundred different visual Transformers for three fundamental CV tasks (classification, detection, and segmentation), where a taxonomy is proposed to organize these methods according to their motivations, structures, and usage scenarios. Because of the differences in training settings and oriented tasks, we have also evaluated these methods on different configurations for easy and intuitive comparison instead of only various benchmarks. Furthermore, we have revealed a series of essential but unexploited aspects that may empower Transformer to stand out from numerous architectures, e.g., slack high-level semantic embeddings to bridge the gap between visual and sequential Transformers. Finally, three promising future research directions are suggested for further investment.
Gait recognition captures gait patterns from the walking sequence of an individual for identification. Most existing gait recognition methods learn features from silhouettes or skeletons for the robustness to clothing, carrying, and other exterior factors. The combination of the two data modalities, however, is not fully exploited. This paper proposes a simple yet effective bimodal fusion (BiFusion) network, which mines the complementary clues of skeletons and silhouettes, to learn rich features for gait identification. Particularly, the inherent hierarchical semantics of body joints in a skeleton is leveraged to design a novel Multi-scale Gait Graph (MSGG) network for the feature extraction of skeletons. Extensive experiments on CASIA-B and OUMVLP demonstrate both the superiority of the proposed MSGG network in modeling skeletons and the effectiveness of the bimodal fusion for gait recognition. Under the most challenging condition of walking in different clothes on CASIA-B, our method achieves the rank-1 accuracy of 92.1%.
Graph classification is an important problem with applications across many domains, like chemistry and bioinformatics, for which graph neural networks (GNNs) have been state-of-the-art (SOTA) methods. GNNs are designed to learn node-level representation based on neighborhood aggregation schemes, and to obtain graph-level representation, pooling methods are applied after the aggregation operation in existing GNN models to generate coarse-grained graphs. However,due to highly diverse applications of graph classification, and the performance of existing pooling methods vary on different graphs. In other words, it is a challenging problem to design a universal pooling architecture to perform well in most cases, leading to a demand for data-specific pooling methods in real-world applications. To address this problem, we propose to use neural architecture search (NAS) to search for adaptive pooling architectures for graph classification. Firstly we designed a unified framework consisting of four modules: Aggregation, Pooling, Readout, and Merge, which can cover existing human-designed pooling methods for graph classification. Based on this framework, a novel search space is designed by incorporating popular operations in human-designed architectures. Then to enable efficient search, a coarsening strategy is proposed to continuously relax the search space, thus a differentiable search method can be adopted. Extensive experiments on six real-world datasets from three domains are conducted, and the results demonstrate the effectiveness and efficiency of the proposed framework.
Deep learning models are notoriously data-hungry. Thus, there is an urging need for data-efficient techniques in medical image analysis, where well-annotated data are costly and time consuming to collect. Motivated by the recently revived "Copy-Paste" augmentation, we propose TumorCP, a simple but effective object-level data augmentation method tailored for tumor segmentation. TumorCP is online and stochastic, providing unlimited augmentation possibilities for tumors' subjects, locations, appearances, as well as morphologies. Experiments on kidney tumor segmentation task demonstrate that TumorCP surpasses the strong baseline by a remarkable margin of 7.12% on tumor Dice. Moreover, together with image-level data augmentation, it beats the current state-of-the-art by 2.32% on tumor Dice. Comprehensive ablation studies are performed to validate the effectiveness of TumorCP. Meanwhile, we show that TumorCP can lead to striking improvements in extremely low-data regimes. Evaluated with only 10% labeled data, TumorCP significantly boosts tumor Dice by 21.87%. To the best of our knowledge, this is the very first work exploring and extending the "Copy-Paste" design in medical imaging domain. Code is available at: https://github.com/YaoZhang93/TumorCP.
Liver cancer is one of the most common cancers worldwide. Due to inconspicuous texture changes of liver tumor, contrast-enhanced computed tomography (CT) imaging is effective for the diagnosis of liver cancer. In this paper, we focus on improving automated liver tumor segmentation by integrating multi-modal CT images. To this end, we propose a novel mutual learning (ML) strategy for effective and robust multi-modal liver tumor segmentation. Different from existing multi-modal methods that fuse information from different modalities by a single model, with ML, an ensemble of modality-specific models learn collaboratively and teach each other to distill both the characteristics and the commonality between high-level representations of different modalities. The proposed ML not only enables the superiority for multi-modal learning but can also handle missing modalities by transferring knowledge from existing modalities to missing ones. Additionally, we present a modality-aware (MA) module, where the modality-specific models are interconnected and calibrated with attention weights for adaptive information exchange. The proposed modality-aware mutual learning (MAML) method achieves promising results for liver tumor segmentation on a large-scale clinical dataset. Moreover, we show the efficacy and robustness of MAML for handling missing modalities on both the liver tumor and public brain tumor (BRATS 2018) datasets. Our code is available at https://github.com/YaoZhang93/MAML.
Accurate segmentation of brain tumors from magnetic resonance imaging (MRI) is clinically relevant in diagnoses, prognoses and surgery treatment, which requires multiple modalities to provide complementary morphological and physiopathologic information. However, missing modality commonly occurs due to image corruption, artifacts, different acquisition protocols or allergies to certain contrast agents in clinical practice. Though existing efforts demonstrate the possibility of a unified model for all missing situations, most of them perform poorly when more than one modality is missing. In this paper, we propose a novel Adversarial Co-training Network (ACN) to solve this issue, in which a series of independent yet related models are trained dedicated to each missing situation with significantly better results. Specifically, ACN adopts a novel co-training network, which enables a coupled learning process for both full modality and missing modality to supplement each other's domain and feature representations, and more importantly, to recover the `missing' information of absent modalities. Then, two unsupervised modules, i.e., entropy and knowledge adversarial learning modules are proposed to minimize the domain gap while enhancing prediction reliability and encouraging the alignment of latent representations, respectively. We also adapt modality-mutual information knowledge transfer learning to ACN to retain the rich mutual information among modalities. Extensive experiments on BraTS2018 dataset show that our proposed method significantly outperforms all state-of-the-art methods under any missing situation.
Medical imaging plays a pivotal role in diagnosis and treatment in clinical practice. Inspired by the significant progress in automatic image captioning, various deep learning (DL)-based architectures have been proposed for generating radiology reports for medical images. However, model uncertainty (i.e., model reliability/confidence on report generation) is still an under-explored problem. In this paper, we propose a novel method to explicitly quantify both the visual uncertainty and the textual uncertainty for the task of radiology report generation. Such multi-modal uncertainties can sufficiently capture the model confidence scores at both the report-level and the sentence-level, and thus they are further leveraged to weight the losses for achieving more comprehensive model optimization. Our experimental results have demonstrated that our proposed method for model uncertainty characterization and estimation can provide more reliable confidence scores for radiology report generation, and our proposed uncertainty-weighted losses can achieve more comprehensive model optimization and result in state-of-the-art performance on a public radiology report dataset.