Global Contextual Dependency Network for Object Detection

Current two-stage object detectors extract the local visual features of Regions of Interest (RoIs) for object recognition and bounding-box regression. However, only using local visual features will lose global contextual dependencies, which are helpful to recognize objects with featureless appearances and restrain false detections. To tackle the problem, a simple framework, named Global Contextual Dependency Network (GCDN), is presented to enhance the classification ability of two-stage detectors. Our GCDN mainly consists of two components, Context Representation Module (CRM) and Context Dependency Module (CDM). Specifically, a CRM is proposed to construct multi-scale context representations. With CRM, contextual information can be fully explored at different scales. Moreover, the CDM is designed to capture global contextual dependencies. Our GCDN includes multiple CDMs. Each CDM utilizes local Region of Interest (RoI) features and single-scale context representation to generate single-scale contextual RoI features via the attention mechanism. Finally, the contextual RoI features generated by parallel CDMs independently are combined with the original RoI features to help classification. Experiments on MS-COCO 2017 benchmark dataset show that our approach brings continuous improvements for two-stage detectors.

DDN2.0: R and Python packages for differential dependency network analysis of biological systems

Data-driven differential dependency network analysis identifies in a complex and often unknown overall molecular circuitry a network of differentially connected molecular entities (pairwise selective coupling or uncoupling depending on the specific phenotypes or experimental conditions). Such differential dependency networks are typically used to assist in the inference of potential key pathways. Based on our previously developed Differential Dependency Network (DDN) method, we report here the fully implemented R and Python software tool packages for public use. The DDN algorithm uses a fused Lasso model and block-wise coordinate descent to estimate both the common and differential edges of dependency networks. The identified DDN can help to provide plausible interpretation of data, gain new insight of disease biology, and generate novel hypotheses for further validation and investigations.

Dependency network-based approach to the implicit structure and semantic diffusion modes of semantic prosody

This paper studies the implicit structures and the diffusion modes of semantic prosody on the dependency networks of some English words such as cause and their Chinese equivalents. It is found that the structure of semantic prosody is a bi-stratified network consisting of a few large clusters gathering in the center with most nodes of low dependency capability scattered around. With regard to the diffusion modes, results show that: (i) within one shortest path length, the core words directly attract the nodes with the same or similar semantic characteristics and exclude those with conflicting ones, creating the clearest and the most intense semantic diffusion; (ii) over one shortest path length, semantic diffusion is achieved through content words or function words, and the semantic diffusion modes created with function words as bridges are relatively vaguer and more complicated ones. This conclusion also results in the semantic prosodies of other English words and their Chinese equivalent words, revealing, to some extent, a common cognitive approach to understanding the internal structure and the diffusion modes of semantic prosody.

Network-adjusted Kendall’s Tau Measure for Feature Screening with Application to High-dimensional Survival Genomic Data

Motivation In high-dimensional genetic/genomic data, the identification of genes related to clinical survival trait is a challenging and important issue. In particular, right-censored survival outcomes and contaminated biomarker data make the relevant feature screening difficult. Several independence screening methods have been developed, but they fail to account for gene–gene dependency information, and may be sensitive to outlying feature data. Results We improve the inverse probability-of-censoring weighted (IPCW) Kendall’s tau statistic by using Google’s PageRank Markov matrix to incorporate feature dependency network information. Also, to tackle outlying feature data, the nonparanormal approach transforming the feature data to multivariate normal variates are utilized in the graphical lasso procedure to estimate the network structure in feature data. Simulation studies under various scenarios show that the proposed network-adjusted weighted Kendall’s tau approach leads to more accurate feature selection and survival prediction than the methods without accounting for feature dependency network information and outlying feature data. The applications on the clinical survival outcome data of diffuse large B-cell lymphoma and of The Cancer Genome Atlas lung adenocarcinoma patients demonstrate clearly the advantages of the new proposal over the alternative methods. Supplementary information Supplementary data are available at Bioinformatics online.

A syntactic dependency network approach to the study of translational language

Complex network approach provides language research with quantitative measures that can capture global features of language. Although translational language has been recognized as a ‘third code’ by some researchers, its independence still calls for further and quantitative validation in an overall manner. In this study, we intend to examine this independence and explore comprehensively its features. We investigated macroscopically translational language from English into Chinese and from Chinese into English by comparing with its source language and native language through syntactic dependency networks. The results show that: (1) translational language presents small-world and scale-free properties like most languages do; (2) however, it is independent of and different from both source language and native language in terms of its network parameters; (3) its network parameters show values eclectic between source language and native language, and this eclectic tendency may be regarded as a new candidate for universal features of translational language, which certainly needs further validation in other genres and language pairs. This study also corroborates that quantitative linguistic method of complex network approach can be well utilized in the study of translational language.