Interaction network

About: Interaction network is a research topic. Over the lifetime, 2700 publications have been published within this topic receiving 113372 citations.

The Correlation Fractal Dimension of Complex Networks

Abstract: The fractality of complex networks is studied by estimating the correlation dimensions of the networks. Comparing with the previous algorithms of estimating the box dimension, our algorithm achieves a significant reduction in time complexity. For four benchmark cases tested, that is, the Escherichia coli (E. Coli) metabolic network, the Homo sapiens protein interaction network (H. Sapiens PIN), the Saccharomyces cerevisiae protein interaction network (S. Cerevisiae PIN) and the World Wide Web (WWW), experiments are provided to demonstrate the validity of our algorithm.

Fraud Detection in Dynamic Interaction Network

Abstract: Fraud detection from massive user behaviors is often regarded as trying to find a needle in a haystack. In this paper, we suggest abnormal behavioral patterns can be better revealed if both sequential and interaction behaviors of users can be modeled simultaneously, which however has rarely been addressed in prior work. Along this line, we propose a COllective Sequence and INteraction (COSIN) model, in which the behavioral sequences and interactions between source and target users in a dynamic interaction network are modeled uniformly in a probabilistic graphical model. More specifically, the sequential schema is modeled with a hierarchical Hidden Markov Model, and meanwhile it is shifted to the interaction schema to generate the interaction counts through Poisson factorization. A hybrid Gibbs-Variational algorithm is then proposed for efficient parameter estimation of the COSIN model. We conduct extensive experiments on both synthetic and real-world telecom datasets in different scales, and the results show that the proposed model outperforms some competitive baseline methods and is scalable. A case is further presented to show the precious explainability of the model.

Automatic annotation of BIND molecular interactions from three‐dimensional structures

Abstract: Software to automate the process of extracting molecular interactions from three- dimensional (3D) structures has been developed that records these as Biomolecular Interaction Network Database (BIND) pairwise interaction records. Full annotation of BIND records is provided through a database processing tool called MMDBind, including detailed atom-atom and residue-residue level interaction information. BIND three-dimensional interaction annotation is synthesized by combining information from the Molecular Modeling Database (MMDB), and the HET (heterogen) group dictionary of small molecules in the macromolecular Crystallographic Information Format (mmCIF). Interactions are validated using the Protein Quaternary Structure (PQS) system. A total of 18,166 interactions were removed as being redundant or biologically irrelevant after PQS validation. This first pass MMDBind annotation creates two new divisions of BIND, 3D Biopolymers (BIND-3DBP) comprising 16,737 initial interaction records, and 3D Small Molecules (BIND-3DSM) comprising 48,219 records. Visualization of interacting residues and nucleotides within a macromolecular structure is possible directly from the BIND database owing to added 3D feature annotation within the BIND records that can be conveniently seen using Cn3D ("see-in-3D") after query from the BIND Data Manager. These interaction records provide a further demonstration of the completeness of the BIND data specification and its capabilities as storage and exchange format for all kinds of molecular interactions, including RNA, DNA, protein, and small molecules. Data from the 3DBP and 3DSM sets are available for downloading in Abstract Syntax Notation.1 (ASN.1) or Extensible Markup Language (XML) formats at ftp://ftp.bind.ca/DB/ MMDBBind. Data from the 3DBP set is available for interactive query from the BIND Data Manager at www.bind.ca. © 2002 Wiley Periodicals, Inc. Biopoly (Nucleic Acid Sci) 61: 111-120, 2002; DOI 10.1002/bip.10143

Embryonic stem cells: protein interaction networks.

Abstract: Embryonic stem cells have the ability to differentiate into nearly all cell types. However, the molecular mechanism of its pluripotency is still unclear. Oct3/4, Sox2 and Nanog are important factors of pluripotency. Oct3/4 (hereafter referred to as Oct4), in particular, has been an irreplaceable factor in the induction of pluripotency in adult cells. Proteins interacting with Oct4 and Nanog have been identified via affinity purification and mass spectrometry. These data, together with iterative purifications of interacting proteins allowed a protein interaction network to be constructed. The network currently includes 77 transcription factors, all of which are interconnected in one network. In-depth studies of some of these transcription factors show that they all recruit the NuRD complex. Hence, transcription factor clustering and chromosomal remodeling are key mechanism used by embryonic stem cells. Studies using RNA interference suggest that more pluripotency genes are yet to be discovered via protein-protein interactions. More work is required to complete and curate the embryonic stem cell protein interaction network. Analysis of a saturated protein interaction network by system biology tools can greatly aid in the understanding of the embryonic stem cell pluripotency network.

HDINet: Hierarchical Dual-Sensor Interaction Network for RGBT Tracking

Abstract: RGBT tracking is a practical solution that combines RGB and thermal infrared modes to solve tracking failures in complex environments to achieve all-day and all-weather work, which makes it gradually applied in multifarious fields. The fundamental reason is that it could avoid the damage of tracking performance caused by the limitation of the imaging characteristics of a single sensor. The existing work aggregates features in different ways, without considering hierarchical complementary interactions and the value of the initial input that may affect subsequent aggregation. In this paper, a novel hierarchical dual-sensor interaction network is proposed, which is mainly composed of dual-sensor interaction, sensor-specific and instance learning. Specifically, our network mainly benefits from the design of two modules, called feature interaction module and data encoding module. The dominant information of the dual sensor is extracted and supplemented by the former based on attention. The latter encodes the raw data into the initial input of the first feature interaction module, whose quality has a key influence on the follow-up. We investigate the performance through extensive experiments compared with the recent state-of-the-art RGB and RGBT trackers on the GTOT and RGBT234 datasets, which verify that our network is effective in both quantitative and qualitative evaluation.