More Like this

1. Coupled ferroelectricity and superconductivity in bilayer Td-MoTe2

   https://doi.org/10.1038/s41586-022-05521-3  

   Jindal, Apoorv; Saha, Amartyajyoti; Li, Zizhong; Taniguchi, Takashi; Watanabe, Kenji; Hone, James C.; Birol, Turan; Fernandes, Rafael M.; Dean, Cory R.; Pasupathy, Abhay N.; et al (January 2023, Nature)
2. TD-DFTB study of optical properties of silver nanoparticle homodimers and heterodimers

   https://doi.org/10.1063/5.0025672  

   Liu, Zhen; Alkan, Fahri; Aikens, Christine M. (October 2020, The Journal of Chemical Physics)

   null (Ed.)
3. Distributed TD(0) With Almost No Communication

   https://doi.org/10.1109/LCSYS.2023.3287952  

   Liu, Rui; Olshevsky, Alex (January 2023, IEEE Control Systems Letters)
4. Geometric Insights into the Convergence of Non-linear TD Learning

   Brandfonbrener, David; Bruna, Joan (April 2020, International Conference on Learning Representations)

   While there are convergence guarantees for temporal difference (TD) learning when using linear function approximators, the situation for nonlinear models is far less understood, and divergent examples are known. Here we take a first step towards extending theoretical convergence guarantees to TD learning with nonlinear function approximation. More precisely, we consider the expected learning dynamics of the TD(0) algorithm for value estimation. As the step-size converges to zero, these dynamics are defined by a nonlinear ODE which depends on the geometry of the space of function approximators, the structure of the underlying Markov chain, and their interaction. We find a set of function approximators that includes ReLU networks and has geometry amenable to TD learning regardless of environment, so that the solution performs about as well as linear TD in the worst case. Then, we show how environments that are more reversible induce dynamics that are better for TD learning and prove global convergence to the true value function for well-conditioned function approximators. Finally, we generalize a divergent counterexample to a family of divergent problems to demonstrate how the interaction between approximator and environment can go wrong and to motivate the assumptions needed to prove convergence. 

   more » « less
5. Reducing the Cost of TD-CI Simulations of Strong Field Ionization

   https://doi.org/10.1021/acs.jpca.4c01732  

   Durden, Andrew S; Schlegel, H Bernhard (September 2024, The Journal of Physical Chemistry A)

   Strong field ionization of molecules by intense laser pulses can be simulated by time-dependent configuration interaction (TD-CI) with a complex absorbing potential (CAP). Standard molecular basis sets need to be augmented with several sets of diffuse functions for effective interaction with the CAP. This dramatically increases the number of configurations and the cost of the TD-CI simulations as the size of the molecules increases. The cost can be reduced by making use of spin symmetry and by employing an orbital energy cut-off to limit the number of virtual orbitals used to construct the excited configurations. Greater reductions in the number of virtual orbitals can be obtained by examining their interaction with the absorbing potential during simulations and their contributions to the strong field ionization rate. This can be determined from the matrix elements of the absorbing potential and the TD-CI coefficients from test simulations. Compared to a simple 3 hartree cut-off in the orbital energies, these approaches reduce the number of virtual orbitals by 20% - 35% for neutral molecules and 5%-10% for cations. As a result, the cost of simulations is reduced by 35% - 60% for neutral molecules and 5% - 10% for cations. The number of virtual orbitals needed can also be estimated by second-order perturbation theory without the need for test simulations. The number of virtual orbitals can be reduced further by adapting orbitals to the laser field using natural orbitals derived from test simulations. This is particularly effective for cations, yielding reductions of more than 20%. 

   more » « less