Showing 1–7 of 7 results for author: Lee, Y Y

Interpretable Neural PDE Solvers using Symbolic Frameworks

Authors: Yolanne Yi Ran Lee

Abstract: Partial differential equations (PDEs) are ubiquitous in the world around us, modelling phenomena from heat and sound to quantum systems. Recent advances in deep learning have resulted in the development of powerful neural solvers; however, while these methods have demonstrated state-of-the-art performance in both accuracy and computational efficiency, a significant challenge remains in their inter… ▽ More Partial differential equations (PDEs) are ubiquitous in the world around us, modelling phenomena from heat and sound to quantum systems. Recent advances in deep learning have resulted in the development of powerful neural solvers; however, while these methods have demonstrated state-of-the-art performance in both accuracy and computational efficiency, a significant challenge remains in their interpretability. Most existing methodologies prioritize predictive accuracy over clarity in the underlying mechanisms driving the model's decisions. Interpretability is crucial for trustworthiness and broader applicability, especially in scientific and engineering domains where neural PDE solvers might see the most impact. In this context, a notable gap in current research is the integration of symbolic frameworks (such as symbolic regression) into these solvers. Symbolic frameworks have the potential to distill complex neural operations into human-readable mathematical expressions, bridging the divide between black-box predictions and solutions. △ Less

Submitted 10 November, 2023; v1 submitted 31 October, 2023; originally announced October 2023.

Comments: Accepted to the NeurIPS 2023 AI for Science Workshop. arXiv admin note: text overlap with arXiv:2310.19763

Autoregressive Renaissance in Neural PDE Solvers

Authors: Yolanne Yi Ran Lee

Abstract: Recent developments in the field of neural partial differential equation (PDE) solvers have placed a strong emphasis on neural operators. However, the paper "Message Passing Neural PDE Solver" by Brandstetter et al. published in ICLR 2022 revisits autoregressive models and designs a message passing graph neural network that is comparable with or outperforms both the state-of-the-art Fourier Neural… ▽ More Recent developments in the field of neural partial differential equation (PDE) solvers have placed a strong emphasis on neural operators. However, the paper "Message Passing Neural PDE Solver" by Brandstetter et al. published in ICLR 2022 revisits autoregressive models and designs a message passing graph neural network that is comparable with or outperforms both the state-of-the-art Fourier Neural Operator and traditional classical PDE solvers in its generalization capabilities and performance. This blog post delves into the key contributions of this work, exploring the strategies used to address the common problem of instability in autoregressive models and the design choices of the message passing graph neural network architecture. △ Less

Submitted 30 October, 2023; originally announced October 2023.

Comments: Presented as a workshop poster at ICLR 2023

Observation of new intrinsic properties of VO2

Authors: Bongjin Simon Mun, Kai Chen, Youngchul Leem, Catherine Dejoie, Nobumichi Tamura, Martin Kunz, Zhi Liu, Michael E. Grass, Changwoo Park, Y. Yvette Lee, Honglyoul Ju

Abstract: The single crystal VO2, exihibiting a first-order metal-insulator transition (MIT) at 67.2 degrees C and an insulator-insulator transition (IIT) at ~49.7 degrees C, is grown. From synchrotron-based x-ray microdiffraction analysis, the IIT shows structural phase transition (SPT) of monoclinic M2 to M1 phases while the MIT displays M1 to rutile R phases. The IIT exhibits percolative SPT while the MI… ▽ More The single crystal VO2, exihibiting a first-order metal-insulator transition (MIT) at 67.2 degrees C and an insulator-insulator transition (IIT) at ~49.7 degrees C, is grown. From synchrotron-based x-ray microdiffraction analysis, the IIT shows structural phase transition (SPT) of monoclinic M2 to M1 phases while the MIT displays M1 to rutile R phases. The IIT exhibits percolative SPT while the MIT shows abrupt transition width of < 0.02 degrees C, supporting Mott's prediction. The MIT occurs non-percolatively with a sharp boundary between R and M1 phases. The MIT onset temperature shows significant variation. △ Less

Submitted 3 August, 2011; v1 submitted 27 July, 2010; originally announced July 2010.

Comments: This paper has been withdrawn by the authors. We would like to withdraw the paper since we arrived at different conclusion based on the re-interpretation of data. Thank you

An Improved Limit on the Muon Electric Dipole Moment

Authors: G. W. Bennett, B. Bousquet, H. N. Brown, G. Bunce, R. M. Carey, P. Cushman, G. T. Danby, P. T. Debevec, M. Deile, H. Deng, W. Deninger, S. K. Dhawan, V. P. Druzhinin, L. Duong, E. Efstathiadis, F. J. M. Farley, G. V. Fedotovich, S. Giron, F. E. Gray, D. Grigoriev, M. Grosse-Perdekamp, A. Grossmann, M. F. Hare, D. W. Hertzog, X. Huang , et al. (51 additional authors not shown)

Abstract: Three independent searches for an electric dipole moment (EDM) of the positive and negative muons have been performed, using spin precession data from the muon g-2 storage ring at Brookhaven National Laboratory. Details on the experimental apparatus and the three analyses are presented. Since the individual results on the positive and negative muon, as well as the combined result, d=-0.1(0.9)E-1… ▽ More Three independent searches for an electric dipole moment (EDM) of the positive and negative muons have been performed, using spin precession data from the muon g-2 storage ring at Brookhaven National Laboratory. Details on the experimental apparatus and the three analyses are presented. Since the individual results on the positive and negative muon, as well as the combined result, d=-0.1(0.9)E-19 e-cm, are all consistent with zero, we set a new muon EDM limit, |d| < 1.9E-19 e-cm (95% C.L.). This represents a factor of 5 improvement over the previous best limit on the muon EDM. △ Less

Submitted 26 July, 2009; v1 submitted 7 November, 2008; originally announced November 2008.

Comments: 19 pages, 15 figures, 7 tables

Intervalley splittings of Si quantum wells

Authors: S. -H. Park, Y. Y. Lee, Doyeol Ahn

Abstract: Multi-valley effective mass theory for silicon quantum well structure is studied taking into account the external fields and the quantum interfaces. It is found that the phenomenological delta function potential, employed to explain the valley splitting caused by the quantum well interface in the previous work [Ref. 10], can be derived self-consistently from the multi-valley effective mass theor… ▽ More Multi-valley effective mass theory for silicon quantum well structure is studied taking into account the external fields and the quantum interfaces. It is found that the phenomenological delta function potential, employed to explain the valley splitting caused by the quantum well interface in the previous work [Ref. 10], can be derived self-consistently from the multi-valley effective mass theory. Finite element method is used to solve the multi-valley effective equations. Theoretical predictions are in a reasonably good agreement with the recent experimental observation of valley splitting in a SiO_{2}/Si/SiO_{2} quantum well, which prove the validity of our approach. △ Less

Submitted 9 March, 2008; v1 submitted 5 November, 2007; originally announced November 2007.

Comments: Major revision including the correction of numerical errors

The NSNS High Energy Beam Transport Line

Authors: D. Raparia, J. Alessi, Y. Y. Lee, W. T. Weng

Abstract: In the National Spallation Neutron Source (NSNS) design, a 180 meter long transport line connects the 1 GeV linac to an accumulator ring. The linac beam has a current of 28 mA, pulse length of 1 ms, and 60 Hzへるつ rep rate. The high energy transport line consists of sixteen 60 degrees FODO cells, and accommodates a 90 degrees achromatic bend, an energy compressor, collimators, part of injection syste… ▽ More In the National Spallation Neutron Source (NSNS) design, a 180 meter long transport line connects the 1 GeV linac to an accumulator ring. The linac beam has a current of 28 mA, pulse length of 1 ms, and 60 Hzへるつ rep rate. The high energy transport line consists of sixteen 60 degrees FODO cells, and accommodates a 90 degrees achromatic bend, an energy compressor, collimators, part of injection system, and enough diagnostic devices to measure the beam quality before injection. To reduce the uncontrolled beam losses, this line has nine beam halo scrapers and very tight tolerances on both transverse and longitudinal beam dynamics under space charge conditions. The design of this line is presented. △ Less

Submitted 9 September, 1997; originally announced September 1997.

Comments: 3 pages, transfer line design

Journal ref: Conf.Proc.C970512:162-164,1997

Beam Dynamics problems in a muon collider

Authors: R. B. Palmer, J. C. Gallardo, R. C. Fernow, H. Kirk, I. Stumer, Y. Y. Lee, M. Syphers, Y. Torun, D. Winn, D. Neuffer, Y. Cho, J. Norem, N. Mokhov, R. Noble, A. Tollestrup, R. Scanlan, S. Caspi, O. Napoly

Abstract: We discuss the various beam dynamics problems in muon collider systems, starting from the proton accelerator needed to generate the muon beams and proceeding through the muon storage ring. We discuss the various beam dynamics problems in muon collider systems, starting from the proton accelerator needed to generate the muon beams and proceeding through the muon storage ring. △ Less

Submitted 30 March, 1995; originally announced March 1995.

Comments: Latex, 11 pages, submitted to ICFA Beam Dynamics Newsletter

Journal ref: ICFA Beam Dyn.Newslett.8:27-33,1995