Humans are able to identify and categorize novel compositions of known concepts. The task in Compositional Zero-Shot learning (CZSL) is to learn composition of primitive concepts, i.e. objects and states, in such a wa...
详细信息
ISBN:
(纸本)9781450392037
Humans are able to identify and categorize novel compositions of known concepts. The task in Compositional Zero-Shot learning (CZSL) is to learn composition of primitive concepts, i.e. objects and states, in such a way that even their novel compositions can be zero-shot classified. In this work, we do not assume any prior knowledge on the feasibility of novel compositions ***-world setting, where infeasible compositions dominate the search space. We propose a Compositional variational graph autoencoder (CVGAE) approach for learning the variational embeddings of the primitive concepts (nodes) as well as feasibility of their compositions (via edges). Such modelling makes CVGAE scalable to real-world application scenarios. This is in contrast to SOTA method, CGE [33], which is computationally very expensive. *** benchmark C-GQA dataset, CGE requires 3.94x10(5) nodes, whereas CVGAE requires only 1323 nodes. We learn a mapping of the graph and image embeddings onto a common embedding space. CVGAE adopts a deep metric learning approach and learns a similarity metric in this space via bi-directional contrastive loss between projected graph and image embeddings. We validate the effectiveness of our approach on three benchmark *** also demonstrate via an image retrieval task that the representations learnt by CVGAE are better suited for compositional generalization.
Bitcoin provides pseudo-anonymity to its users, leading to many transactions related to illicit activities. The advent of mixing services like OnionBC, Bitcoin Fog, and *** has allowed users to increase their anonymit...
详细信息
ISBN:
(纸本)9781728191270
Bitcoin provides pseudo-anonymity to its users, leading to many transactions related to illicit activities. The advent of mixing services like OnionBC, Bitcoin Fog, and *** has allowed users to increase their anonymity further. This paper tackles the pseudo-anonymity of the Bitcoin blockchain by developing a scalable spark based framework to find patterns in the transaction data. The efficacy of the framework is demonstrated by performing exploratory analysis. Furthermore, the paper shows the capabilities of bitcoin-based graph representations and addresses the issue of user profiling based on unsupervised learning approaches for analysing Bitcoin transactions and users. The authors convert the transaction graph of the Bitcoin data to contain only Wallet-IDs and generate graph embeddings using variational graph autoencoder [1]. Additionally, the authors use explainable-AI techniques and Kohonen self organizing maps to visualize and understand the results obtained from the unsupervised learning methods.
Diverse types of cells interact and communicate with each other to maintain tissue homeostasis and perform biological functions. Perturbations to these interactions can break the homeostasis of the tissue microenviron...
详细信息
Diverse types of cells interact and communicate with each other to maintain tissue homeostasis and perform biological functions. Perturbations to these interactions can break the homeostasis of the tissue microenvironment, leading to disease. Understanding intercellular communication changes in disease is critical for therapeutic development. Cell-cell communication networks (CCCNs) inferred from single-cell RNA sequencing data are highly variable and only capture a snapshot of the dynamic intercellular communication system. We develop a graphical generative model to compare CCCNs between disease and control samples to identify disease associated perturbations to intercellular communications. The distribution of CCCNs is learned using variational graph autoencoder (VGAE) in disease and control groups separately. Then a large number of graphs is generated to assess the significance of the difference between the two distributions using different graph distance measures. We demonstrate the advantage of this approach in improving the power of identifying disease associated perturbations to intercellular communications through both simulation studies and real scRNA-seq datasets.
Social recommendation has been an effective approach to solve the new user recommendation problem based on user -item interactions and user -user social relations. Although lots of research has been done, it is still ...
详细信息
Social recommendation has been an effective approach to solve the new user recommendation problem based on user -item interactions and user -user social relations. Although lots of research has been done, it is still an emergent and challenging issue to predict the behaviors of new users without any historical interaction. Firstly, the previous methods fail to consider social structures and social semantics when looking for potential social neighbors for new users, resulting in inconsistent preferences of these neighbors. Secondly, existing methods employ deterministic modeling way to represent and aggregate neighbors, limiting the diversity and robustness of new user representations. Therefore, we present a novel new user preference uncertainty modeling framework, named Disentangled -feature and Composite -prior VAE(DC-VAE), to predict the behaviors of new users without any interaction. Concretely, a length -adaptive similarity metric considering the length of user behaviors and social relationships is designed for all users to choose more analogous neighbors, especially more effective for new users due to the metric incorporating the social structures and social semantics. Then the Neighbor -based Disentangled Features module is proposed to disentangle different types of neighbor characteristics and model more diversified new user representations. Next, unlike traditional Gaussian prior constraint, the Neighbor -based Composite prior module is proposed to fuse the priors of neighbors and obtain more expressive and robust new user representations. Finally, we theoretically prove the advantages of composite prior and disentangled features. Extensive experiments on three datasets demonstrate that our model DC-VAE is remarkably superior to other baselines.
With the rapid development of graphvariational Bayes theory, some representative community detection methods have been proposed. Although these methods are well designed, they are limited by the inherent constraints ...
详细信息
With the rapid development of graphvariational Bayes theory, some representative community detection methods have been proposed. Although these methods are well designed, they are limited by the inherent constraints of static networks. Therefore, how to re-examine the information in static networks from a new perspective becomes a challenge. To this end, in this paper, we propose EG-VGAE and its variant EGC-VGAE, going beyond the constraints of static networks from a new perspective. Specifically, we first demonstrate that the evolution relations on static networks can be simulated reasonably. And then, our EG-VGAE method combines evolution information with static network information to realize a fine-grained propagation of local low -frequency signals to global low -frequency signals, thereby improving the accuracy of community assignment for each node. Building on this progress, our method EGC-VGAE imposes the smoothness constraint on adjacent slices, significantly enhancing the sensitivity of the method to evolution information and mitigating the impact of network noise. The comprehensive experimental results on real static networks well validate that our methods outperform state-of-the-art methods in most cases. The code is available at https://***/GDM-SCNU/EG-VGAE.
Long noncoding RNAs (lncRNAs) participate in various biological processes and have close linkages with diseases. In vivo and in vitro experiments have validated many associations between lncRNAs and diseases. However,...
详细信息
Long noncoding RNAs (lncRNAs) participate in various biological processes and have close linkages with diseases. In vivo and in vitro experiments have validated many associations between lncRNAs and diseases. However, biological experiments are time-consuming and expensive. Here, we introduce LDA-VGHB, an lncRNA-disease association (LDA) identification framework, by incorporating feature extraction based on singular value decomposition and variational graph autoencoder and LDA classification based on heterogeneous Newton boosting machine. LDA-VGHB was compared with four classical LDA prediction methods (i.e. SDLDA, LDNFSGB, IPCARF and LDASR) and four popular boosting models (XGBoost, AdaBoost, CatBoost and LightGBM) under 5-fold cross-validations on lncRNAs, diseases, lncRNA-disease pairs and independent lncRNAs and independent diseases, respectively. It greatly outperformed the other methods with its prominent performance under four different cross-validations on the lncRNADisease and MNDR databases. We further investigated potential lncRNAs for lung cancer, breast cancer, colorectal cancer and kidney neoplasms and inferred the top 20 lncRNAs associated with them among all their unobserved lncRNAs. The results showed that most of the predicted top 20 lncRNAs have been verified by biomedical experiments provided by the Lnc2Cancer 3.0, lncRNADisease v2.0 and RNADisease databases as well as publications. We found that HAR1A, KCNQ1DN, ZFAT-AS1 and HAR1B could associate with lung cancer, breast cancer, colorectal cancer and kidney neoplasms, respectively. The results need further biological experimental validation. We foresee that LDA-VGHB was capable of identifying possible lncRNAs for complex diseases. LDA-VGHB is publicly available at https://***/plhhnu/LDA-VGHB.
暂无评论