Machine learning in the string landscape

A bstractWe utilize machine learning to study the string landscape. Deep data dives and conjecture generation are proposed as useful frameworks for utilizing machine learning in the landscape, and examples of each are presented. A decision tree accurately predicts the number of weak Fano toric threefolds arising from reflexive polytopes, each of which determines a smooth F-theory compactification, and linear regression generates a previously proven conjecture for the gauge group rank in an ensemble of 43×2.96×10755$$ \frac{4}{3}\times 2.96\times {10}^{755} $$ F-theory compactifications. Logistic regression generates a new conjecture for when E6 arises in the large ensemble of F-theory compactifications, which is then rigorously proven. This result may be relevant for the appearance of visible sectors in the ensemble. Through conjecture generation, machine learning is useful not only for numerics, but also for rigorous results.

[1]  W. Taylor,et al.  A Monte Carlo exploration of threefold base geometries for 4d F-theory vacua , 2016, Journal of High Energy Physics.

[2]  W. Taylor,et al.  Non-toric bases for elliptic Calabi-Yau threefolds and 6D F-theory vacua , 2015, 1504.07689.

[3]  W. Taylor,et al.  ℙ1$$ {\mathrm{\mathbb{P}}}^1 $$-bundle bases and the prevalence of non-Higgsable structure in 4D F-theory models , 2015, 1506.03204.

[4]  Compactifications of F-Theory on Calabi--Yau Threefolds -- I , 1996, hep-th/9602114.

[5]  W. Taylor,et al.  Geometric constraints in dual F-theory and heterotic string compactifications , 2014, 1405.2074.

[6]  Cost of seven-brane gauge symmetry in a quadrillion F-theory compactifications , 2016, 1610.08864.

[7]  W. Taylor,et al.  Non-Higgsable QCD and the standard model spectrum in F-theory , 2014, 1409.8295.

[8]  W. Taylor,et al.  6D F-theory models and elliptically fibered Calabi-Yau threefolds over semi-toric base surfaces , 2014, 1404.6300.

[9]  T. Watari.  Statistics of F-theory flux vacua for particle physics , 2015, 1506.08433.

[10]  W. Taylor.  On the Hodge structure of elliptically fibered Calabi-Yau threefolds , 2012, 1205.0952.

[11]  F. Denef,et al.  Computational complexity of the landscape II - Cosmological considerations , 2017, 1706.06430.

[12]  C. Lawrie,et al.  The Tate form on steroids: resolution and higher codimension fibers , 2012, 1212.2949.

[13]  S. Chekanov,et al.  Energy dependence of the charged multiplicity in deep inelastic scattering at HERA. , 2008 .

[14]  C. Lawrie,et al.  Box graphs and singular fibers , 2014, 1402.2653.

[15]  J. Shaneson,et al.  Non-Abelian Gauge Symmetry and the Higgs Mechanism in F-Theory , 2014, 1402.5962.

[16]  Quantization of four-form fluxes and dynamical neutralization of the cosmological constant , 2000, hep-th/0004134.

[17]  J. Shaneson,et al.  Matter from geometry without resolution , 2013, Journal of High Energy Physics.

[18]  J. D. Loera,et al.  Triangulations: Structures for Algorithms and Applications , 2010 .

[19]  C. Vafa.  Evidence for F theory , 1996, hep-th/9602022.

[20]  M. Cvetič,et al.  On the computation of non‐perturbative effective potentials in the string theory landscape – IIB/F‐theory perspective – , 2010, 1009.5386.

[21]  N. Nakayama.  On Weierstrass Models , 1988 .

[22]  F. Denef,et al.  Distributions of flux vacua , 2004, hep-th/0404116.

[23]  T. Watari,et al.  The vertical, the horizontal and the rest: anatomy of the middle cohomology of Calabi-Yau fourfolds and F-theory applications , 2014, 1408.6167.

[24]  Yang-Hui He,et al.  Deep-Learning the Landscape , 2017, 1706.02714.

[25]  W. Taylor,et al.  Classifying bases for 6D F-theory models , 2012, 1201.1943.

[26]  T. Watari.  Statistics of Flux Vacua for Particle Physics , 2015 .

[27]  W. Taylor,et al.  Calabi-Yau threefolds with large h2,1 , 2014, 1406.0514.

[28]  W. Taylor,et al.  Sections, multisections, and U(1) fields in F-theory , 2014, 1404.1527.

[29]  Sakura Schafer-Nameki,et al.  Box graphs and resolutions II: From Coulomb phases to fiber faces , 2015, 1511.01801.

[30]  J. Shaneson,et al.  Geometry and Topology of String Junctions , 2014, 1410.6817.

[31]  Counting flux vacua , 2003, hep-th/0307049.

[32]  Fabian Ruehle.  Evolving neural networks with genetic algorithms to study the string landscape , 2017, 1706.07024.

[33]  Thomas G. Dietterich.  What is machine learning? , 2020, Archives of Disease in Childhood.

[34]  R. Bousso,et al.  Fast optimization algorithms and the cosmological constant , 2017, 1706.08503.

[35]  W. Taylor,et al.  P-bundle bases and the prevalence of non-Higgsable structure in 4D F-theory models , 2015 .

[36]  J. Marsano,et al.  Yukawas, G-flux, and spectral covers from resolved Calabi-Yau’s , 2011, 1108.1794.

[37]  J. Halverson,et al.  Strong Coupling in F-theory and Geometrically Non-Higgsable Seven-branes , 2016, 1603.01639.

[38]  Sakura Schafer-Nameki,et al.  Box Graphs and Resolutions I , 2014, 1407.3520.

[39]  W. Taylor,et al.  Toric bases for 6D F‐theory models , 2012, 1204.0283.

[40]  M. Kreuzer,et al.  Classification of Reflexive Polyhedra in Three Dimensions , 1998 .

[41]  W. Taylor,et al.  The F-theory geometry with most flux vacua , 2015, 1511.03209.

[42]  W. Taylor,et al.  Non-Higgsable clusters for 4D F-theory models , 2014, 1412.6112.

[43]  Radford M. Neal.  Pattern Recognition and Machine Learning , 2007, Technometrics.

[44]  Michael R. Douglas,et al.  Computational complexity of the landscape I , 2006, ArXiv.