Michael Doron
ABSTRACT
Scientific computational models are crucial for analyzing and understanding complex real-life systems that are otherwise difficult for experimentation. However, the complex behavior and the vast input-output space of these models often make them opaque, slowing the discovery of novel phenomena. In this work, we present HINT (Hessian INTerestingness) -- a new algorithm that can automatically and systematically explore black-box models and highlight local nonlinear interactions in the input-output space of the model. This tool aims to facilitate the discovery of interesting model behaviors that are unknown to the researchers. Using this simple yet powerful tool, we were able to correctly rank all pairwise interactions in known benchmark models and do so faster and with greater accuracy than state-of-the-art methods. We further applied HINT to existing computational neuroscience models, and were able to reproduce important scientific discoveries that were published years after the creation of those models. Finally, we ran HINT on two real-world models (in neuroscience and earth science) and found new behaviors of the model that were of value to domain experts.
- Srdjan D. Antic, Wen Liang Zhou, Anna R. Moore, Shaina M. Short, and Katerina D. Ikonomu. 2010. The decade of the dendritic NMDA spike. Journal of Neuroscience Research (2010).Google Scholar
- Michael J Behrenfeld and Michael J Behrenfeld. 2017. Abandoning Sverdrup's Critical Depth Hypothesis on phytoplankton blooms. Ecology (2017).Google Scholar
- Francisco Bezanilla. 2008. Ion Channels: From Conductance to Structure. Neuron (2008).Google Scholar
- Daniela Bianchi, Addolorata Marasco, Alessandro Limongiello, Cristina Marchetti, Helene Marie, Brunello Tirozzi, and Michele Migliore. 2012. On the mechanisms underlying the depolarization block in the spiking dynamics of CA1 pyramidal neurons. Journal of Computational Neuroscience (2012). Google ScholarDigital Library
- E. Biton and H. Gildor. 2011. Stepwise seasonal restratification and the evolution of salinity minimum in the Gulf of Aqaba (Gulf of Eilat). Journal of Geophysical Research (2011).Google Scholar
- Gary F Bradshaw, P. W. Langley, and H. A. Simon. 1983. Studying Scientific Discovery by Computer Simulation. Science (1983).Google Scholar
- Nicholas T. Carnevale and Michael L. Hines. 2006. The NEURON Book. Cambridge University Press, Cambridge. Google ScholarDigital Library
- Michael Doron, Giuseppe Chindemi, Eilif Muller, Henry Markram, and Idan Segev. 2017. Timed Synaptic Inhibition Shapes NMDA Spikes, Influencing Local Dendritic Processing and Global I/O Properties of Cortical Neurons. Cell Reports (2017).Google Scholar
- Kai Du, Yu-Wei Wu, Robert Lindroos, Yu Liu, Balázs Rózsa, Gergely Katona, Jun B. Ding, and Jeanette Hellgren Kotaleski. 2017. Cell-type-specific inhibition of the dendritic plateau potential in striatal spiny projection neurons. (2017).Google Scholar
- A Edwards. 1999. Zooplankton Mortality and the Dynamical Behaviour of Plankton Population Models. Bulletin of Mathematical Biology (1999).Google Scholar
- J. Evans and A. Rzhetsky. 2010. Machine Science. Science (2010).Google Scholar
- Brian C. Falkenhainer and Ryszard S Michalski. 1986. Integrating quantitative and qualitative discovery: The ABACUS system. Machine Learning (1986). Google ScholarDigital Library
- Jerome H. Friedman. 2001. Greedy function approximation: A gradient boosting machine. The Annals of Statistics (2001).Google Scholar
- Jerome H. Friedman and Bogdan E. Popescu. 2008. Predictive learning via rule ensembles. The Annals of Applied Statistics (2008).Google Scholar
- C. M. Gray and D. A. McCormick. 1996. Chattering Cells: Superficial Pyramidal Neurons Contributing to the Generation of Synchronous Oscillations in the Visual Cortex. Science (1996).Google Scholar
- Brandon Greenwell, Bradley Boehmke, Jay Cunningham, and GBM Developers. 2019. gbm: Generalized Boosted Regression Models. R package version 2.1.5.Google Scholar
- T.J. Hastie and R.J. Tibshirani. 1990. Generalized additive models. Chapman & Hall/CRC.Google Scholar
- Etay Hay, Sean Hill, Felix Schürmann, Henry Markram, and Idan Segev. 2011. Models of neocortical layer 5b pyramidal cells capturing a wide range of dendritic and perisomatic active properties. PLoS Computational Biology (2011).Google Scholar
- Giles Hooker. 2004. Discovering additive structure in black box functions. In Proceedings of the 2004 ACM SIGKDD international conference on Knowledge discovery and data mining - KDD '04. ACM Press, New York, New York, USA. Google ScholarDigital Library
- Giles Hooker. 2007. Generalized Functional ANOVA Diagnostics for High-Dimensional Functions of Dependent Variables. Journal of Computational and Graphical Statistics (2007).Google Scholar
- Eugene M. Izhikevich. 2000. neural Excitability, Spiking And Bursting. International Journal of Bifurcation and Chaos (2000).Google Scholar
- Monika Jadi, Alon Polsky, Jackie Schiller, and Bartlett W. Mel. 2012. Location-Dependent Effects of Inhibition on Local Spiking in Pyramidal Neuron Dendrites. PLoS Computational Biology (2012).Google Scholar
- C E Jahr and C F Stevens. 1990. Voltage dependence of NMDA-activated macroscopic conductances predicted by single-channel kinetics. The Journal of neuroscience (1990).Google Scholar
- Scott L. Jones, Minh-Son To, and Greg J. Stuart. 2017. Dendritic small conductance calcium-activated potassium channels activated by action potentials suppress EPSPs and gate spike-timing dependent synaptic plasticity. eLife (2017).Google Scholar
- Bo Kang, Jefrey Lijffijt, Raú l Santos-Rodríguez, and Tijl De Bie. 2018. SICA: subjectively interesting component analysis. Data Mining and Knowledge Discovery (2018). Google ScholarDigital Library
- Ross D. King, Jem Rowland, Stephen G. Oliver, Michael Young, Wayne Aubrey, Emma Byrne, Maria Liakata, Magdalena Markham, Pinar Pir, Larisa N. Soldatova, Andrew Sparkes, Kenneth E. Whelan, and Amanda Clare. 2009. The Automation of Science. Science (2009).Google Scholar
- Angela M. Kuhn, Katja Fennel, and Jann Paul Mattern. 2015. Model investigations of the North Atlantic spring bloom initiation. Progress in Oceanography (2015).Google Scholar
- Alan U. Larkman. 1991. Dendritic morphology of pyramidal neurones of the visual cortex of the rat: III. Spine distributions. The Journal of Comparative Neurology (1991).Google Scholar
- Yin Lou. 2018. mltk: Machine Learning Tool Kit. https://github.com/yinlou/mltkGoogle Scholar
- Yin Lou, Rich Caruana, Johannes Gehrke, and Giles Hooker. 2013. Accurate intelligible models with pairwise interactions. In Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining - KDD '13. ACM Press, New York, New York, USA. Google ScholarDigital Library
- Guy Major, Matthew E. Larkum, and Jackie Schiller. 2013. Active Properties of Neocortical Pyramidal Neuron Dendrites. Annual Review of Neuroscience (2013).Google Scholar
- Christoph Molnar. 2019. Interpretable Machine Learning. https://christophm.github.io/interpretable-ml-book/.Google Scholar
- William F. Podlaski, Alexander Seeholzer, Lukas N. Groschner, Gero Miesenböck, Rajnish Ranjan, and Tim P. Vogels. 2017. Mapping the function of neuronal ion channels in model and experiment. eLife (2017).Google Scholar
- Astrid A. Prinz, Cyrus P. Billimoria, and Eve Marder. 2003. Alternative to Hand-Tuning Conductance-Based Models: Construction and Analysis of Databases of Model Neurons. Journal of Neurophysiology (2003).Google Scholar
- Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin. 2016. "Why Should I Trust You?": Explaining the Predictions of Any Classifier. (2016).Google Scholar
- Markus Schartau and Andreas Oschlies. 2003. Simultaneous data-based optimization of a 1D-ecosystem model at three locations in the North Atlantic: Part II-Standing stocks and nitrogen fluxes. Journal of Marine Research (2003).Google Scholar
- Jackie Schiller, Guy Major, Helmut J Koester, and Yitzhak Schiller. 2000. NMDA spikes in basal dendrites of cortical pyramidal neurons. Nature (2000).Google Scholar
- Jürgen Schmidhuber. 2010. Formal Theory of Creativity, Fun, and Intrinsic Motivation (1990--2010). IEEE Transactions on Autonomous Mental Development (2010). Google ScholarDigital Library
- Michael Schmidt and Hod Lipson. 2009. Distilling Free-Form Natural Laws from Experimental Data. Science (2009).Google Scholar
- I Segev and W Rall. 1988. Computational study of an excitable dendritic spine. Journal of Neurophysiology (1988).Google Scholar
- Greg J. Stuart and Michael Hausser. 2001. Dendritic coincidence detection of EPSPs and action potentials. Nature Neuroscience (2001).Google Scholar
- Greg J. Stuart and Bert Sakmann. 1994. Active propagation of somatic action potentials into neocortical pyramidal cell dendrites. Nature (1994).Google Scholar
- Yuriy Sverchkov, Mark Craven, FM Jones, PGK Reiser, CH Bryant, and SH Mugggleton. 2017. A review of active learning approaches to experimental design for uncovering biological networks. PLOS Computational Biology (2017).Google Scholar
- H. U. Sverdrup. 1953. On Conditions for the Vernal Blooming of Phytoplankton. ICES Journal of Marine Science (1953).Google Scholar
- J. Tabak, M. Tomaiuolo, A. E. Gonzalez-Iglesias, L. S. Milescu, and R. Bertram. 2011. Fast-Activating Voltage- and Calcium-Dependent Potassium (BK) Conductance Promotes Bursting in Pituitary Cells: A Dynamic Clamp Study. Journal of Neuroscience (2011).Google Scholar
- Simen Tennøe, Geir Halnes, and Gaute T. Einevoll. 2018. Uncertainpy: A Python Toolbox for Uncertainty Quantification and Sensitivity Analysis in Computational Neuroscience. Frontiers in Neuroinformatics (2018).Google Scholar
- Werner Van Geit, Michael Gevaert, Giuseppe Chindemi, Christian Rössert, Jean-Denis Courcol, Eilif B. Muller, Felix Schürmann, Idan Segev, and Henry Markram. 2016. BluePyOpt: Leveraging Open Source Software and Cloud Infrastructure to Optimise Model Parameters in Neuroscience. Frontiers in Neuroinformatics (2016).Google Scholar
- Jack Waters, Andreas Schaefer, and Bert Sakmann. 2005. Backpropagating action potentials in neurones: Measurement, mechanisms and potential functions. Progress in Biophysics and Molecular Biology (2005).Google Scholar
- Mary D Womack and Kamran Khodakhah. 2003. Somatic and Dendritic Small-Conductance Calcium-Activated Potassium Channels Regulate the Output of Cerebellar Purkinje Neurons. The Journal of Neuroscience (2003).Google Scholar
- Simon N Wood. 2017. Generalized Additive Models: An Introduction with R. Chapman and Hall/CRC.Google ScholarDigital Library
- Tailin Wu and Max Tegmark. 2018. Toward an AI Physicist for Unsupervised Learning. (2018). http://arxiv.org/abs/1810.10525Google Scholar
Index Terms
- Discovering Unexpected Local Nonlinear Interactions in Scientific Black-box Models
Recommendations
Efficient Discovery of the Most Interesting Associations
Self-sufficient itemsets have been proposed as an effective approach to summarizing the key associations in data. However, their computation appears highly demanding, as assessing whether an itemset is self-sufficient requires consideration of all ...
Neuroscience bridging scientific disciplines in health: Who builds the bridge, who pays for it?
The purpose of this study is to investigate the dynamics of cross-disciplinary research in health-related fields as affected by individual and institutional factors. We examine the topics of more than 500,000 doctoral dissertations from U.S. ...
Discovering unexpected documents in corpora
Text mining is widely used to discover frequent patterns in large corpora of documents. Hence, many classical data mining techniques, that have been proven fruitful in the context of data stored in relational databases, are now successfully used in the ...
Comments