Deep learning's performance has been extensively recognized recently. Graph neural networks (GNNs) are designed to deal with graph-structural data that classical deep learning does not easily manage. Since most GNNs were created using distinct theories, direct comparisons are impossible. Prior research has primarily concentrated on categorizing existing models, with little attention paid to their intrinsic connections. The purpose of this study is to establish a unified framework that integrates GNNs based on spectral graph and approximation theory. The framework incorporates a strong integration between spatial- and spectral-based GNNs while tightly associating approaches that exist within each respective domain.
Federated Learning (FL) is known to perform Machine Learning tasks in a distributed manner. Over the years, this has become an emerging technology especially with various data protection and privacy policies being imposed FL allows performing machine learning tasks whilst adhering to these challenges. As with the emerging of any new technology, there are going to be challenges and benefits. A challenge that exists in FL is the communication costs, as FL takes place in a distributed environment where devices connected over the network have to constantly share their updates this can create a communication bottleneck. In this paper, we present a survey of the research that is performed to overcome the communication constraints in an FL setting.
During the past decade, deep learning's performance has been widely recognized in a variety of machine learning tasks, ranging from image classification, speech recognition to natural language understanding. Graph neural networks (GNN) are a type of deep learning that is designed to handle non-Euclidean issues using graph-structured data that are difficult to solve with traditional deep learning techniques. The majority of GNNs were created using a variety of processes, including random walk, PageRank, graph convolution, and heat diffusion, making direct comparisons impossible. Previous studies have primarily focused on classifying current models into distinct categories, with little investigation of their internal relationships. This research proposes a unified theoretical framework and a novel perspective that can methodologically integrate existing GNN into our framework. We survey and categorize existing GNN models into spatial and spectral domains, as well as show linkages between subcategories within each domain. Further investigation reveals a strong relationship between the spatial, spectral, and subgroups of these domains.
Recently a novel framework has been proposed for designing the molecular structure of chemical compounds using both artificial neural networks (ANNs) and mixed integer linear programming (MILP). In the framework, we first define a feature vector $f(C)$ of a chemical graph $C$ and construct an ANN that maps $x=f(C)$ to a predicted value $\eta(x)$ of a chemical property $\pi$ to $C$. After this, we formulate an MILP that simulates the computation process of $f(C)$ from $C$ and that of $\eta(x)$ from $x$. Given a target value $y^*$ of the chemical property $\pi$, we infer a chemical graph $C^\dagger$ such that $\eta(f(C^\dagger))=y^*$ by solving the MILP. In this paper, we use linear regression to construct a prediction function $\eta$ instead of ANNs. For this, we derive an MILP formulation that simulates the computation process of a prediction function by linear regression. The results of computational experiments suggest our method can infer chemical graphs with around up to 50 non-hydrogen atoms.
Recently developed large pre-trained language models, e.g., BERT, have achieved remarkable performance in many downstream natural language processing applications. These pre-trained language models often contain hundreds of millions of parameters and suffer from high computation and latency in real-world applications. It is desirable to reduce the computation overhead of the models for fast training and inference while keeping the model performance in downstream applications. Several lines of work utilize knowledge distillation to compress the teacher model to a smaller student model. However, they usually discard the teacher's knowledge when in inference. Differently, in this paper, we propose RefBERT to leverage the knowledge learned from the teacher, i.e., facilitating the pre-computed BERT representation on the reference sample and compressing BERT into a smaller student model. To guarantee our proposal, we provide theoretical justification on the loss function and the usage of reference samples. Significantly, the theoretical result shows that including the pre-computed teacher's representations on the reference samples indeed increases the mutual information in learning the student model. Finally, we conduct the empirical evaluation and show that our RefBERT can beat the vanilla TinyBERT over 8.1\% and achieves more than 94\% of the performance of $\BERTBASE$ on the GLUE benchmark. Meanwhile, RefBERT is 7.4x smaller and 9.5x faster on inference than BERT$_{\rm BASE}$.
This paper presents a novel algorithm that registers a collection of mono-modal 3D images in a simultaneous fashion, named as Direct Simultaneous Registration (DSR). The algorithm optimizes global poses of local frames directly based on the intensities of images (without extracting features from the images). To obtain the optimal result, we start with formulating a Direct Bundle Adjustment (DBA) problem which jointly optimizes pose parameters of local frames and intensities of panoramic image. By proving the independence of the pose from panoramic image in the iterative process, DSR is proposed and proved to be able to generate the same optimal poses as DBA, but without optimizing the intensities of the panoramic image. The proposed DSR method is particularly suitable in mono-modal registration and in the scenarios where distinct features are not available, such as Transesophageal Echocardiography (TEE) images. The proposed method is validated via simulated and in-vivo 3D TEE images. It is shown that the proposed method outperforms conventional sequential registration method in terms of accuracy and the obtained results can produce good alignment in in-vivo images.
The Graph Augmented Multi-layer Perceptron (GA-MLP) model is an attractive alternative to Graph Neural Networks (GNNs). This is because it is resistant to the over-smoothing problem, and deeper GA-MLP models yield better performance. GA-MLP models are traditionally optimized by the Stochastic Gradient Descent (SGD). However, SGD suffers from the layer dependency problem, which prevents the gradients of different layers of GA-MLP models from being calculated in parallel. In this paper, we propose a parallel deep learning Alternating Direction Method of Multipliers (pdADMM) framework to achieve model parallelism: parameters in each layer of GA-MLP models can be updated in parallel. The extended pdADMM-Q algorithm reduces communication cost by utilizing the quantization technique. Theoretical convergence to a critical point of the pdADMM algorithm and the pdADMM-Q algorithm is provided with a sublinear convergence rate $o(1/k)$. Extensive experiments in six benchmark datasets demonstrate that the pdADMM can lead to high speedup, and outperforms all the existing state-of-the-art comparison methods.
Continual learning is considered a promising step towards next-generation Artificial Intelligence (AI), where deep neural networks (DNNs) make decisions by continuously learning a sequence of different tasks akin to human learning processes. It is still quite primitive, with existing works focusing primarily on avoiding (catastrophic) forgetting. However, since forgetting is inevitable given bounded memory and unbounded task loads, 'how to reasonably forget' is a problem continual learning must address in order to reduce the performance gap between AIs and humans, in terms of 1) memory efficiency, 2) generalizability, and 3) robustness when dealing with noisy data. To address this, we propose a novel ScheMAtic memory peRsistence and Transience (SMART) framework for continual learning with external memory that builds on recent advances in neuroscience. The efficiency and generalizability are enhanced by a novel long-term forgetting mechanism and schematic memory, using sparsity and 'backward positive transfer' constraints with theoretical guarantees on the error bound. Robust enhancement is achieved using a novel short-term forgetting mechanism inspired by background information-gated learning. Finally, an extensive experimental analysis on both benchmark and real-world datasets demonstrates the effectiveness and efficiency of our model.
Handwriting recognition is of crucial importance to both Human Computer Interaction (HCI) and paperwork digitization. In the general field of Optical Character Recognition (OCR), handwritten Chinese character recognition faces tremendous challenges due to the enormously large character sets and the amazing diversity of writing styles. Learning an appropriate distance metric to measure the difference between data inputs is the foundation of accurate handwritten character recognition. Existing distance metric learning approaches either produce unacceptable error rates, or provide little interpretability in the results. In this paper, we propose an interpretable distance metric learning approach for handwritten Chinese character recognition. The learned metric is a linear combination of intelligible base metrics, and thus provides meaningful insights to ordinary users. Our experimental results on a benchmark dataset demonstrate the superior efficiency, accuracy and interpretability of our proposed approach.