2018.05.22 Machine learning and complex networks for complex systems big data analysis

2019-07-07 00:55:36 5






题目:Machine learning and complex networks for complex systems big data analysis


报告人:Carlo Vittorio Cannistraci


Biomedical Cybernetics Group, Technical University Dresden (Germany)


时  间:2018年05月22日(周二)13:00-14:00


地  点:北京大学老化学楼东配楼101报告厅


主持人:汤超 教授



  The talk will present our research at the Biomedical Cybernetics Group that I established about four years ago in Technical University Dresden. We adopt a transdisciplinary approach integrating information theory, machine learning and network science to investigate the physics of networked adaptive complex systems at different scales, from molecules to ecological and social systems, with a particular attention to biology and medicine, and a new emerging interest for the analysis of complex big data in social and economic science. Our theoretical effort is to translate advanced mathematical paradigms typically adopted in theoretical physics (such as topology, network and manifold theory) to characterize many-body interactions in complex systems. We apply the theoretical frameworks we invent in the mission to develop computational tools for systems and network analysis. In particular, we deal with: prediction of wiring in networks and multiscale-combinatorial marker design for quantification of topological modifications in complex networks. Our attention for precision biomedicine is aimed to topics with important impact from the economical point of view such as development of tools for disease biomarker discovery, drug repositioning and combinatorial drug therapy. 

  This talk will focus on two main theoretical innovation. Firstly, the development of machine learning for topological estimation of nonlinear relations in high-dimensional data1 (or in complex networks2) and its relevance for applications in big data. Secondly, we will discuss the Local Community Paradigm (LCP)4,5, which is a theory proposed to model local-topology-dependent link-growth in complex networks and therefore it is useful to devise topological methods for link prediction in monopartite and bipartite5 networks such as molecular drug-target interactionsand product-consumer networks.



  Carlo Vittorio Cannistraci is a theoretical engineer, head of the Biomedical Cybernetics Group and faculty of the Department of Physics in the Technical University Dresden, which is a member of the TU9 excellence-league (the nine most prestigious technical universities in Germany). Carlo’s area of research embraces information theory, machine learning and complex networks including also applications in systems biomedicine and neuroscience. Nature Biotechnology selected Carlo’s article (Cell 2010)7 on machine learning in developmental biology to be nominated in the list of 2010 notable breakthroughs in computational biology. Circulation Research featured Carlo’s work (Circulation Research 2012)8 on leveraging a cardiovascular systems biology strategy to predict future outcomes in heart attacks, commenting: “a space-aged evaluation using computational biology”. The Technical University Dresden honoured Carlo of the Young Investigator Award 2016 in Physics for his work on the local-community-paradigm theory and link prediction in bipartite networks5. In 2017, Springer-Nature scientific blog highlighted with an interview to Carlo his study on “How the brain handles pain through the lens of network science9. The American Heart Association covered this year on its website the recent chronobiology discovery of Carlo on how the sunshine affects the risk and time onset of heart attack10


References (* indicates first co-authorship)

1.           Cannistraci, C. V., Ravasi, T., Montevecchi, F. M., Ideker, T. & Alessio, M. Nonlinear dimension reduction and clustering by minimum curvilinearity unfold neuropathic pain and tissue embryological classes. Bioinformatics 26, i531–i539 (2010).

2.           Cannistraci, C. V., Alanis-Lobato, G. & Ravasi, T. Minimum curvilinearity to enhance topological prediction of protein interactions by network embedding. in Bioinformatics 29, (2013).

3.           Muscoloni, A., Thomas, J. M., ..,  & Cannistraci, C. V. Machine learning meets complex networks via coalescent embedding of networks in the hyperbolic space. Nature Communication (2017). 

4.           Cannistraci, C. V., Alanis-Lobato, G. & Ravasi, T. From link-prediction in brain connectomes and protein interactomes to the local-community-paradigm in complex networks. Sci. Rep. 3, 1–13 (2013).

5.           Daminelli, S., Thomas, J. M., Durán, C. & Vittorio Cannistraci, C. Common neighbours and the local-community-paradigm for topological link prediction in bipartite networks. New J. Phys. 17, 113037 (2015).

6.             Duran, C., …, Cannistraci, C.V. Pioneering topological methods for network-based drug–target prediction by exploiting a brain-network self-organization theoryBriefings in Bioinformatics, bbx041 (2017).

7.           Ravasi, T.*, Cannistraci C.V.*,  et al. An atlas of combinatorial transcriptional regulation in mouse and man. Cell 140, (2010).

8.           Ammirati, E.*, Cannistraci, C.V.*,  et al. Identification and predictive value of interleukin-6+ interleukin-10+ and interleukin-6-interleukin-10+ cytokine patterns in st-elevation acute myocardial infarction. Circ. Res. 111, 1336–1348 (2012).

9.             Narula, V., …, and Cannistraci, C.V. Can local-community-paradigm and epitopological learning enhance our understanding of how local brain connectivity is able to process, learn and memorize chronic pain? Applied Network Science 2 (1), 28        

10.           Cannistraci, C.V., et al., “Summer Shift”: A Potential Effect of Sunshine on the Time Onset of STElevation Acute Myocardial Infarction Journal of the American Heart Association 7 (8), e006878