N-ary relational knowledge bases (KBs) represent knowledge with binary and beyond-binary relational facts. Especially, in an n-ary relational fact, the involved entities play different roles, e.g., the ternary relation PlayCharacterIn consists of three roles, ACTOR, CHARACTER and MOVIE. However, existing approaches are often directly extended from binary relational KBs, i.e., knowledge graphs, while missing the important semantic property of role. Therefore, we start from the role level, and propose a Role-Aware Modeling, RAM for short, for facts in n-ary relational KBs. RAM explores a latent space that contains basis vectors, and represents roles by linear combinations of these vectors. This way encourages semantically related roles to have close representations. RAM further introduces a pattern matrix that captures the compatibility between the role and all involved entities. To this end, it presents a multilinear scoring function to measure the plausibility of a fact composed by certain roles and entities. We show that RAM achieves both theoretical full expressiveness and computation efficiency, which also provides an elegant generalization for approaches in binary relational KBs. Experiments demonstrate that RAM outperforms representative baselines on both n-ary and binary relational datasets.
Recent years have witnessed the popularity and success of graph neural networks (GNN) in various scenarios. To obtain data-specific GNN architectures, researchers turn to neural architecture search (NAS), which has made impressive success in discovering effective architectures in convolutional neural networks. However, it is non-trivial to apply NAS approaches to GNN due to challenges in search space design and the expensive searching cost of existing NAS methods. In this work, to obtain the data-specific GNN architectures and address the computational challenges facing by NAS approaches, we propose a framework, which tries to Search to Aggregate NEighborhood (SANE), to automatically design data-specific GNN architectures. By designing a novel and expressive search space, we propose a differentiable search algorithm, which is more efficient than previous reinforcement learning based methods. Experimental results on four tasks and seven real-world datasets demonstrate the superiority of SANE compared to existing GNN models and NAS approaches in terms of effectiveness and efficiency. (Code is available at: https://github.com/AutoML-4Paradigm/SANE).
Recently, a special kind of graph, i.e., supernet, which allows two nodes connected by multi-choice edges, has exhibited its power in neural architecture search (NAS) by searching for better architectures for computer vision (CV) and natural language processing (NLP) tasks. In this paper, we discover that the design of such discrete architectures also appears in many other important learning tasks, e.g., logical chain inference in knowledge graphs (KGs) and meta-path discovery in heterogeneous information networks (HINs). Thus, we are motivated to generalize the supernet search problem on a broader horizon. However, none of the existing works are effective since the supernet topology is highly task-dependent and diverse. To address this issue, we propose to tensorize the supernet, i.e., unify the subgraph search problems by a tensor formulation and encode the topology inside the supernet by a tensor network. We further propose an efficient algorithm that admits both stochastic and deterministic objectives to solve the search problem. Finally, we perform extensive experiments on diverse learning tasks, i.e., architecture design for CV, logic inference for KG, and meta-path discovery for HIN. Empirical results demonstrate that our method leads to better performance and architectures.
Since convolutional neural networks (ConvNets) can easily memorize noisy labels, which are ubiquitous in visual classification tasks, it has been a great challenge to train ConvNets against them robustly. Various solutions, e.g., sample selection, label correction, and robustifying loss functions, have been proposed for this challenge, and most of them stick to the end-to-end training of the representation (feature extractor) and classifier. In this paper, by a deep rethinking and careful re-examining on learning behaviors of the representation and classifier, we discover that the representation is much more fragile in the presence of noisy labels than the classifier. Thus, we are motivated to design a new method, i.e., REED, to leverage above discoveries to learn from noisy labels robustly. The proposed method contains three stages, i.e., obtaining the representation by self-supervised learning without any labels, transferring the noisy label learning problem into a semisupervised one by the classifier directly and reliably trained with noisy labels, and joint semi-supervised retraining of both the representation and classifier. Extensive experiments are performed on both synthetic and real benchmark datasets. Results demonstrate that the proposed method can beat the state-of-the-art ones by a large margin, especially under high noise level.
Classical machine learning implicitly assumes that labels of the training data are sampled from a clean distribution, which can be too restrictive for real-world scenarios. However, statistical learning-based methods may not train deep learning models robustly with these noisy labels. Therefore, it is urgent to design Label-Noise Representation Learning (LNRL) methods for robustly training deep models with noisy labels. To fully understand LNRL, we conduct a survey study. We first clarify a formal definition for LNRL from the perspective of machine learning. Then, via the lens of learning theory and empirical study, we figure out why noisy labels affect deep models' performance. Based on the theoretical guidance, we categorize different LNRL methods into three directions. Under this unified taxonomy, we provide a thorough discussion of the pros and cons of different categories. More importantly, we summarize the essential components of robust LNRL, which can spark new directions. Lastly, we propose possible research directions within LNRL, such as new datasets, instance-dependent LNRL, and adversarial LNRL. Finally, we envision potential directions beyond LNRL, such as learning with feature-noise, preference-noise, domain-noise, similarity-noise, graph-noise, and demonstration-noise.
Graph neural network (GNN) has recently been established as an effective representation learning framework on graph data. However, the popular message passing models rely on local permutation invariant aggregate functions, which gives rise to the concerns about their representational power. Here, we introduce the concept of automorphic equivalence to theoretically analyze GNN's expressiveness in differentiating node's structural role. We show that the existing message passing GNNs have limitations in learning expressive representations. Moreover, we design a novel GNN class that leverages learnable automorphic equivalence filters to explicitly differentiate the structural roles of each node's neighbors, and uses a squeeze-and-excitation module to fuse various structural information. We theoretically prove that the proposed model is expressive in terms of generating distinct representations for nodes with different structural feature. Besides, we empirically validate our model on eight real-world graph data, including social network, e-commerce co-purchase network and citation network, and show that it consistently outperforms strong baselines.
Non-local low-rank tensor approximation has been developed as a state-of-the-art method for hyperspectral image (HSI) restoration, which includes the tasks of denoising, compressed HSI reconstruction and inpainting. Unfortunately, while its restoration performance benefits from more spectral bands, its runtime also substantially increases. In this paper, we claim that the HSI lies in a global spectral low-rank subspace, and the spectral subspaces of each full band patch group should lie in this global low-rank subspace. This motivates us to propose a unified paradigm combining the spatial and spectral properties for HSI restoration. The proposed paradigm enjoys performance superiority from the non-local spatial denoising and light computation complexity from the low-rank orthogonal basis exploration. An efficient alternating minimization algorithm with rank adaptation is developed. It is done by first solving a fidelity term-related problem for the update of a latent input image, and then learning a low-dimensional orthogonal basis and the related reduced image from the latent input image. Subsequently, non-local low-rank denoising is developed to refine the reduced image and orthogonal basis iteratively. Finally, the experiments on HSI denoising, compressed reconstruction, and inpainting tasks, with both simulated and real datasets, demonstrate its superiority with respect to state-of-the-art HSI restoration methods.
Negative sampling, which samples negative triplets from non-observed ones in knowledge graph (KG), is an essential step in KG embedding. Recently, generative adversarial network (GAN), has been introduced in negative sampling. By sampling negative triplets with large gradients, these methods avoid the problem of vanishing gradient and thus obtain better performance. However, they make the original model more complex and harder to train. In this paper, motivated by the observation that negative triplets with large gradients are important but rare, we propose to directly keep track of them with the cache. In this way, our method acts as a "distilled" version of previous GAN-based methods, which does not waste training time on additional parameters to fit the full distribution of negative triplets. However, how to sample from and update the cache are two critical questions. We propose to solve these issues by automated machine learning techniques. The automated version also covers GAN-based methods as special cases. Theoretical explanation of NSCaching is also provided, justifying the superior over fixed sampling scheme. Besides, we further extend NSCaching with skip-gram model for graph embedding. Finally, extensive experiments show that our method can gain significant improvements on various KG embedding models and the skip-gram model, and outperforms the state-of-the-art negative sampling methods.
This paper presents a novel neural architecture search (NAS) framework for graph neural networks (GNNs). We design an expressive search space that focuses on a common and critical component of GNNs -- propagation model. Specifically, we search for propagation matrices and the connections between propagation steps. Our search space covers various graph types, e.g., homogeneous graphs, heterogeneous graphs, and can be naturally extended to higher-dimensional recommender systems and spatial-temporal data. We propose a sampling-based one-shot NAS algorithm to search for appropriate propagation patterns efficiently. Extensive experiments in three different scenarios are used to evaluate the proposed framework. We show that the performance of the models obtained by our framework is better than state-of-the-art GNN methods. Furthermore, our framework can discover explainable meta-graphs in heterogeneous graphs.