The problem of missing modalities is both critical and non-trivial to be handled in multi-modal models. It is common for multi-modal tasks that certain modalities contribute more compared to other modalities, and if those important modalities are missing, the model performance drops significantly. Such fact remains unexplored by current multi-modal approaches that recover the representation from missing modalities by feature reconstruction or blind feature aggregation from other modalities, instead of extracting useful information from the best performing modalities. In this paper, we propose a Learnable Cross-modal Knowledge Distillation (LCKD) model to adaptively identify important modalities and distil knowledge from them to help other modalities from the cross-modal perspective for solving the missing modality issue. Our approach introduces a teacher election procedure to select the most ``qualified'' teachers based on their single modality performance on certain tasks. Then, cross-modal knowledge distillation is performed between teacher and student modalities for each task to push the model parameters to a point that is beneficial for all tasks. Hence, even if the teacher modalities for certain tasks are missing during testing, the available student modalities can accomplish the task well enough based on the learned knowledge from their automatically elected teacher modalities. Experiments on the Brain Tumour Segmentation Dataset 2018 (BraTS2018) shows that LCKD outperforms other methods by a considerable margin, improving the state-of-the-art performance by 3.61% for enhancing tumour, 5.99% for tumour core, and 3.76% for whole tumour in terms of segmentation Dice score.
The missing modality issue is critical but non-trivial to be solved by multi-modal models. Current methods aiming to handle the missing modality problem in multi-modal tasks, either deal with missing modalities only during evaluation or train separate models to handle specific missing modality settings. In addition, these models are designed for specific tasks, so for example, classification models are not easily adapted to segmentation tasks and vice versa. In this paper, we propose the Shared-Specific Feature Modelling (ShaSpec) method that is considerably simpler and more effective than competing approaches that address the issues above. ShaSpec is designed to take advantage of all available input modalities during training and evaluation by learning shared and specific features to better represent the input data. This is achieved from a strategy that relies on auxiliary tasks based on distribution alignment and domain classification, in addition to a residual feature fusion procedure. Also, the design simplicity of ShaSpec enables its easy adaptation to multiple tasks, such as classification and segmentation. Experiments are conducted on both medical image segmentation and computer vision classification, with results indicating that ShaSpec outperforms competing methods by a large margin. For instance, on BraTS2018, ShaSpec improves the SOTA by more than 3% for enhancing tumour, 5% for tumour core and 3% for whole tumour.
One key challenge in multi-document summarization is to capture the relations among input documents that distinguish between single document summarization (SDS) and multi-document summarization (MDS). Few existing MDS works address this issue. One effective way is to encode document positional information to assist models in capturing cross-document relations. However, existing MDS models, such as Transformer-based models, only consider token-level positional information. Moreover, these models fail to capture sentences' linguistic structure, which inevitably causes confusions in the generated summaries. Therefore, in this paper, we propose document-aware positional encoding and linguistic-guided encoding that can be fused with Transformer architecture for MDS. For document-aware positional encoding, we introduce a general protocol to guide the selection of document encoding functions. For linguistic-guided encoding, we propose to embed syntactic dependency relations into the dependency relation mask with a simple but effective non-linear encoding learner for feature learning. Extensive experiments show the proposed model can generate summaries with high quality.
Multi-modal learning focuses on training models by equally combining multiple input data modalities during the prediction process. However, this equal combination can be detrimental to the prediction accuracy because different modalities are usually accompanied by varying levels of uncertainty. Using such uncertainty to combine modalities has been studied by a couple of approaches, but with limited success because these approaches are either designed to deal with specific classification or segmentation problems and cannot be easily translated into other tasks, or suffer from numerical instabilities. In this paper, we propose a new Uncertainty-aware Multi-modal Learner that estimates uncertainty by measuring feature density via Cross-modal Random Network Prediction (CRNP). CRNP is designed to require little adaptation to translate between different prediction tasks, while having a stable training process. From a technical point of view, CRNP is the first approach to explore random network prediction to estimate uncertainty and to combine multi-modal data. Experiments on two 3D multi-modal medical image segmentation tasks and three 2D multi-modal computer vision classification tasks show the effectiveness, adaptability and robustness of CRNP. Also, we provide an extensive discussion on different fusion functions and visualization to validate the proposed model.
Within natural language processing tasks, linguistic knowledge can always serve an important role in assisting the model to learn excel representations and better guide the natural language generation. In this work, we develop a neural network based abstractive multi-document summarization (MDS) model which leverages dependency parsing to capture cross-positional dependencies and grammatical structures. More concretely, we process the dependency information into the linguistic-guided attention mechanism and further fuse it with the multi-head attention for better feature representation. With the help of linguistic signals, sentence-level relations can be correctly captured, thus improving MDS performance. Our model has two versions based on Flat-Transformer and Hierarchical Transformer respectively. Empirical studies on both versions demonstrate that this simple but effective method outperforms existing works on the benchmark dataset. Extensive analyses examine different settings and configurations of the proposed model which provide a good reference to the community.
Data hiding is the process of embedding information into a noise-tolerant signal such as a piece of audio, video, or image. Digital watermarking is a form of data hiding where identifying data is robustly embedded so that it can resist tampering and be used to identify the original owners of the media. Steganography, another form of data hiding, embeds data for the purpose of secure and secret communication. This survey summarises recent developments in deep learning techniques for data hiding for the purposes of watermarking and steganography, categorising them based on model architectures and noise injection methods. The objective functions, evaluation metrics, and datasets used for training these data hiding models are comprehensively summarised. Finally, we propose and discuss possible future directions for research into deep data hiding techniques.
Recently, deep neural network models have achieved impressive results in various research fields. Come with it, an increasing number of attentions have been attracted by deep super-resolution (SR) approaches. Many existing methods attempt to restore high-resolution images from directly down-sampled low-resolution images or with the assumption of Gaussian degradation kernels with additive noises for their simplicities. However, in real-world scenarios, highly complex kernels and non-additive noises may be involved, even though the distorted images are visually similar to the clear ones. Existing SR models are facing difficulties to deal with real-world images under such circumstances. In this paper, we introduce a new kernel agnostic SR framework to deal with real-world image SR problem. The framework can be hanged seamlessly to multiple mainstream models. In the proposed framework, the degradation kernels and noises are adaptively modeled rather than explicitly specified. Moreover, we also propose an iterative supervision process and frequency-attended objective from orthogonal perspectives to further boost the performance. The experiments validate the effectiveness of the proposed framework on multiple real-world datasets.
The control of traffic signals is fundamental and critical to alleviate traffic congestion in urban areas. However, it is challenging since traffic dynamics are complicated in real situations. Because of the high complexity of modelling the optimisation problem, experimental settings of current works are often inconsistent. Moreover, it is not trivial to control multiple intersections properly in real complex traffic scenarios due to its vast state and action space. Failing to take intersection topology relations into account also results in inferior traffic condition. To address these issues, in this work we carefully design our settings and propose new data including both synthetic and real traffic data in more complex scenarios. Additionally, we propose a novel and strong baseline model based on deep reinforcement learning with the encoder-decoder structure: an edge-weighted graph convolutional encoder to excavate multi-intersection relations; and a unified structure decoder to jointly model multiple junctions in a comprehensive manner, which significantly reduces the number of the model parameters. By doing so, the proposed model is able to effectively deal with multi-intersection traffic optimisation problems. Models have been trained and tested on both synthetic and real maps and traffic data with the Simulation of Urban Mobility (SUMO) simulator. Experimental results show that the proposed model surpasses existing methods in the literature.
Multi-document summarization (MDS) is an effective tool for information aggregation which generates an informative and concise summary from a cluster of topic-related documents. Our survey structurally overviews the recent deep learning based multi-document summarization models via a proposed taxonomy and it is the first of its kind. Particularly, we propose a novel mechanism to summarize the design strategies of neural networks and conduct a comprehensive summary of the state-of-the-art. We highlight the differences among various objective functions which are rarely discussed in the existing literature. Finally, we propose several future directions pertaining to this new and exciting development of the field.
Automated disease classification of radiology images has been emerging as a promising technique to support clinical diagnosis and treatment planning. Unlike generic image classification tasks, a real-world radiology image classification task is significantly more challenging as it is far more expensive to collect the training data where the labeled data is in nature multi-label; and more seriously samples from easy classes often dominate; training data is highly class-imbalanced problem exists in practice as well. To overcome these challenges, in this paper, we propose a novel scheme of Cross-Attention Networks (CAN) for automated thoracic disease classification from chest x-ray images, which can effectively excavate more meaningful representation from data to boost the performance through cross-attention by only image-level annotations. We also design a new loss function that beyond cross-entropy loss to help cross-attention process and is able to overcome the imbalance between classes and easy-dominated samples within each class. The proposed method achieves state-of-the-art results.