Machine Learning Assisted Exploration for Affine Deligne–Lusztig Varieties

Abstract

This paper presents a novel, interdisciplinary study that leverages a Machine Learning (ML) assisted framework to explore the geometry of affine Deligne–Lusztig varieties (ADLV). The primary objective is to investigate the non-emptiness pattern, dimension, and enumeration of irreducible components of ADLV. Our proposed framework demonstrates a recursive pipeline of data generation, model training, pattern analysis, and human examination, presenting an intricate interplay between ML and pure mathematical research. Notably, our data-generation process is nuanced, emphasizing the selection of meaningful subsets and appropriate feature sets. We demonstrate that this framework has a potential to accelerate pure mathematical research, leading to the discovery of new conjectures and promising research directions that could otherwise take significant time to uncover. We rediscover the virtual dimension formula and provide a full mathematical proof of a newly identified problem concerning a certain lower bound of dimension. Furthermore, we extend an open invitation to the readers by providing the source code for computing ADLV and the ML models, promoting further explorations. This paper concludes by sharing valuable experiences and highlighting lessons learned from this collaboration.

Keywords

Affine Deligne–Lusztig varieties / Affine Weyl groups / Loop groups / AI-assisted math research

References

[1] Aitken AC. IV.—on least squares and linear combination of observations. Proc. R. Soc. Edinb., 1936, 55: 42-48

[2] Bach, F., Jenatton, R., Mairal, J., Obozinski, G.: Convex optimization with sparsity-inducing norms. In: Optimization for Machine Learning, pp. 19–53. MIT Press, Cambridge, MA (2011)

[3] Beazley ET. Codimensions of Newton strata for SL3(F)\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$\rm {SL}_3(F)$$\end{document} in the Iwahori case. Math. Z., 2009, 263(3): 499-540

[4] Berglund, P., Campbell, B., Jejjala, V.: Machine learning Kreuzer–Skarke Calabi–Yau threefolds. arXiv:2112.09117 (2021)

[5] Brenti F , Fomin S, Postnikov A. Mixed Bruhat operators and Yang-Baxter equations for Weyl groups. Int. Math. Res. Not. IMRN, 1999, 1999(8): 419-441

[6] Brown N, Sandholm T. Superhuman AI for multiplayer poker. Science, 2019, 365(6456): 885-890

[7] Bruhat, F., Tits, J.: Groupes réductifs sur un corps local, II. Schémas en groupes. Existence d’une donnée radicielle valuée. Publ. Math. l’Inst. Hautes Études Sci. 60, 5–184 (1984)

[8] Brunton SL, Proctor JL, Kutz JN. Discovering governing equations from data by sparse identification of nonlinear dynamical systems. Proc. Natl. Acad. Sci. USA, 2016, 113(15): 3932-3937

[9] Burges CJC. A tutorial on support vector machines for pattern recognition. Data Min. Knowl. Disc., 1998, 2(2): 121-167

[10] Chai C-L. Newton polygons as lattice points. Am. J. Math., 2000, 122(5): 967-990

[11] Chmiela S, Tkatchenko A, Sauceda HE, Poltavsky I, Schütt KT, Müller K-R. Machine learning of accurate energy-conserving molecular force fields. Sci. Adv., 2017, 3(5): e1603015

[12] Cortes C, Vapnik V. Support-vector networks. Mach. Learn., 1995, 20: 273-297

[13] Davies A, Velicković P, Buesing L, Blackwell S, Zheng D, Tomasev N, Tanburn R, Battaglia P, Blundell C, Juhász Aet al.. Advancing mathematics by guiding human intuition with AI. Nature, 2021, 600(7887): 70-74

[14] Degrave J, Felici F, Buchli J, Neunert M, Tracey B, Carpanese F, Ewalds T, Hafner R, Abdolmaleki A, de Las Casas Det al.. Magnetic control of tokamak plasmas through deep reinforcement learning. Nature, 2022, 602(7897): 414-419

[15] Deligne P, Lusztig G. Representations of reductive groups over finite fields. Ann. Math., 1976, 103(1): 103-161

[16] Görtz U, Haines TJ, Kottwitz RE, Reuman DC. Dimensions of some affine Deligne–Lusztig varieties. Ann. Sci. l’École Normale Supérieure, 2006, 39(3): 467-511

[17] Görtz U, Haines TJ, Kottwitz RE, Reuman DC. Affine Deligne–Lusztig varieties in affine flag varieties. Compos. Math., 2010, 146(5): 1339-1382

[18] Görtz U, He X. Dimension of affine Deligne–Lusztig varieties in affine flag varieties. Doc. Math., 2010, 15: 1009-1028

[19] Görtz U, He X, Nie S. P-alcoves and nonemptiness of affine Deligne–Lusztig varieties. Ann. Sci. l’École Normale Supérieure, 2015, 48(3): 647-665

[20] Hamacher P. The geometry of Newton strata in the reduction modulo p\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$p$$\end{document} of Shimura varieties of PEL type. Duke Math. J., 2015, 164(15): 2809-2895

[21] Hamacher P , Viehmann E. Irreducible components of minuscule affine Deligne–Lusztig varieties. Algebra Number Theory, 2018, 12(7): 1611-1634

[22] Han, S., Mao, H., Dally, W.J.: Deep compression: compressing deep neural networks with pruning, trained quantization and Huffman coding. arXiv:1510.00149 (2015)

[23] He X. Minimal length elements in some double cosets of Coxeter groups. Adv. Math., 2007, 215(2): 469-503

[24] He X. A subalgebra of 0-Hecke algebra. J. Algebra, 2009, 322(11): 4030-4039

[25] He X. Geometric and homological properties of affine Deligne–Lusztig varieties. Ann. Math., 2014, 179(1): 367-404

[26] He, X.: Hecke algebras and p\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$p$$\end{document}-adic groups. In: Current Developments in Mathematics 2015, pp. 73–135. International Press, Somerville, MA (2016)

[27] He, X.: Some results on affine Deligne–Lusztig varieties. In: Proceedings of the International Congress of Mathematicians—Rio de Janeiro 2018. Vol. II. Invited Lectures, pp. 1345–1365. World Scientific, Hackensack, NJ (2018)

[28] He X . Cordial elements and dimensions of affine Deligne–Lusztig varieties. Forum Math. Pi, 2021, 9e9

[29] He X, Nie S. Minimal length elements of extended affine Weyl groups. Compos. Math., 2014, 150(11): 1903-1927

[30] He, X., Nie, S., Yu, Q.: Affine Deligne–Lusztig varieties with finite Coxeter parts. arXiv:2208.14058 (2022)

[31] He X, Yu Q. Dimension formula for the affine Deligne–Lusztig variety X(μ,b)\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$X(\mu , b)$$\end{document}. Math. Ann., 2021, 379(3–4): 1747-1765

[32] Ho, J., Jain, A., Abbeel, P.: Denoising diffusion probabilistic models. In: Advances in Neural Information Processing Systems 33 (NeurIPS 2020), pp. 6840–6851. Curran Associates Inc., Red Hook, NY (2020)

[33] Jia, W., Wang, H., Chen, M., Lu, D., Lin, L., Car, R., E, W., Zhang, L.: Pushing the limit of molecular dynamics with ab initio accuracy to 100 million atoms with machine learning. In: SC20: International Conference for High Performance Computing, Networking, Storage and Analysis, pp. 1–14. IEEE (2020)

[34] Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Zidek, A., Potapenko, A., et al.: Highly accurate protein structure prediction with AlphaFold. Nature 596(7873), 583–589 (2021)

[35] Kates-Harbeck J, Svyatkovskiy A, Tang W. Predicting disruptive instabilities in controlled fusion plasmas through deep learning. Nature, 2019, 568(7753): 526-531

[36] Kingma, D.P., Ba, J.: Adam: a method for stochastic optimization. arXiv:1412.6980 (2014)

[37] Kottwitz RE. Isocrystals with additional structure. Compos. Math., 1985, 56(2): 201-220

[38] Lam T, Shimozono M. Quantum cohomology of G/P\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$G/P$$\end{document} and homology of affine Grassmannian. Acta Math., 2010, 204(1): 49-90

[39] Lenart C , Naito S, Sagaki D, Schilling A, Mark S. A uniform model for Kirillov-Reshetikhin crystals I: lifting the parabolic quantum Bruhat graph. Int. Math. Res. Not. IMRN, 2015, 2015(7): 1848-1901

[40] Long Z, Lu Y, Dong B. PDE-Net 2.0: Learning PDEs from data with a numeric-symbolic hybrid deep network. J. Comput. Phys., 2019, 399108925

[41] Long Z, Lu Y, Ma X, Dong B. PDE-Net: learning PDEs from data. Proc. Mach. Learn. Res., 2018, 80: 3208-3216

[42] Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. arXiv:1711.05101 (2017)

[43] Miethke M, Pieroni M, Weber T, Brönstrup M, Hammann P, Halby L, Arimondo PB, Glaser P, Aigle B, Bode HBet al.. Towards the sustainable discovery and development of new antibiotics. Nat. Rev. Chem., 2021, 5(10): 726-749

[44] Mildenhall B, Srinivasan PP, Tancik M, Barron JT, Ramamoorthi R, Ng R. NeRF: representing scenes as neural radiance fields for view synthesis. Commun. ACM, 2021, 65(1): 99-106

[45] Milićević E. Maximal newton points and the quantum Bruhat graph. Mich. Math. J., 2021, 70(3): 451-502

[46] Milićević E, Viehmann E. Generic Newton points and the Newton poset in Iwahori-double cosets. Forum Math. Sigma, 2020, 8e50

[47] Nair, V., Hinton, G.E.: Rectified linear units improve restricted Boltzmann machines. In: Proceedings of the 27th International Conference on Machine Learning (ICML’10), pp. 807–814. Omnipress, Madison, WI (2010)

[48] Nelder JA, Wedderburn RWM. Generalized linear models. J. R. Stat. Soc. Ser. A, 1972, 135(3): 370-384

[49] Nie S. Irreducible components of affine Deligne–Lusztig varieties. Camb. J. Math., 2022, 10(2): 433-510

[50] Raissi M, Perdikaris P, Karniadakis GE. Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. J. Comput. Phys., 2019, 378: 686-707

[51] Rapoport M. A guide to the reduction modulo p\documentclass[12pt]{minimal}\usepackage{amsmath}\usepackage{wasysym}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{amsbsy}\usepackage{mathrsfs}\usepackage{upgreek}\setlength{\oddsidemargin}{-69pt}\begin{document}$$p$$\end{document} of Shimura varieties. Astérisque, 2002, 298: 271-318

[52] Rosenblatt F . The perceptron: a probabilistic model for information storage and organization in the brain. Psychol. Rev., 1958, 65(6): 386-408

[53] Ruder, S.: An overview of gradient descent optimization algorithms. arXiv:1609.04747 (2016)

[54] Rumelhart DE, Hinton GE, Williams RJ. Learning representations by back-propagating errors. Nature, 1986, 323(6088): 533-536

[55] Sadhukhan A. Affine Deligne–Lusztig varieties and quantum Bruhat graph. Math. Z., 2023, 303(1): 21

[56] Schremmer, F.: Generic Newton points and cordial elements. arXiv:2205.02039 (2022)

[57] Silver D, Huang A, Maddison CJ, Guez A, Sifre L, Van Den Driessche G, Schrittwieser J, Antonoglou I, Panneershelvam V, Lanctot Met al.. Mastering the game of go with deep neural networks and tree search. Nature, 2016, 5297587484-489

[58] Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R. Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res., 2014, 15: 1929-1958

[59] Stokes JM, Yang K, Swanson K, Jin W, Cubillos-Ruiz A, Donghia NM, MacNair CR, French S, Carfrae LA, Bloom-Ackermann Zet al.. A deep learning approach to antibiotic discovery. Cell, 2020, 180(4): 688-702

[60] Viehmann E. The dimension of some affine Deligne–Lusztig varieties. Ann. Sci. l’École Normale Supérieure, 2006, 39(3): 513-526

[61] Viehmann E. Truncations of level 1 of elements in the loop group of a reductive group. Ann. Math., 2014, 179(3): 1009-1040

[62] Viehmann, E.: On the geometry of the Newton stratification. In: Shimura Varieties, pp. 192–208. Cambridge University Press, Cambridge (2020)

[63] Viehmann E. Minimal Newton strata in Iwahori double cosets. Int. Math. Res. Not. IMRN, 2021, 2021(7): 5349-5365

[64] Vinyals O, Babuschkin I, Czarnecki WM, Mathieu M, Dudzik A, Chung J, Choi DH, Powell R, Ewalds T, Georgiev Pet al.. Grandmaster level in StarCraft II using multi-agent reinforcement learning. Nature, 2019, 575(7782): 350-354

[65] Von Lilienfeld OA, Burke K. Retrospective on a decade of machine learning for chemical discovery. Nat. Commun., 2020, 11(1): 4895

[66] Williamson, G.: Is deep learning a useful tool for the pure mathematician? Bull. Am. Math. Soc. 61(2), 271–286 (2024)

[67] Zhang L, Han J, Wang H, Car R, E W. Deep potential molecular dynamics: a scalable model with the accuracy of quantum mechanics. Phys. Rev. Lett., 2018, 12014143001

[68] Zhao, W.X., Zhou, K., Li, J., Tang, T., Wang, X., Hou, Y., Min, Y., Zhang, B., Zhang, J., Dong, Z., et al.: A survey of large language models. arXiv:2303.18223 (2023)

[69] Zhou R, Zhu Y. Twisted orbital integrals and irreducible components of affine Deligne–Lusztig varieties. Camb. J. Math., 2020, 81149-241

[70] Zhu X. Affine Grassmannians and the geometric Satake in mixed characteristic. Ann. Math. (2), 2017, 185(2): 403-492

RIGHTS & PERMISSIONS

The Author(s)