Towards Antisymmetric Neural Ansatz Separation

We study separations between two fundamental models (or Ansätze ) of antisymmetric functions, that is, functions f of the form f ( x σ (1) , . . . , x σ ( N ) ) = sign ( σ ) f ( x 1 , . . . , x N ) , where σ is any permutation. These arise in the context of quantum chemistry, and are the basic modeling tool for wavefunctions of Fermionic systems. Specifically, we consider two popular antisymmetric Ansätze: the Slater representation, which leverages the alternating structure of determinants, and the Jastrow ansatz, which augments Slater determinants with a product by an arbitrary symmetric function. We construct an antisymmetric function that can be more efficiently expressed in Jastrow form, yet provably cannot be approximated by Slater determinants unless there are exponentially (in N 2 ) many terms. This represents the first explicit quantitative separation between these two Ansätze.

[1]  Joan Bruna,et al.  Exponential Separations in Symmetric Neural Networks , 2022, ArXiv.

[2]  J. Stokes,et al.  Fermionic wave functions from neural-network constrained hidden states , 2021, Proceedings of the National Academy of Sciences of the United States of America.

[3]  G. Pfander,et al.  Quantitative approximation results for complex-valued neural networks , 2021, SIAM J. Math. Data Sci..

[4]  J. M. Landsberg,et al.  Geometry of backflow transformation ansatz for quantum many-body fermionic wavefunctions , 2021, 2111.10314.

[5]  Michael A. Osborne,et al.  Universal Approximation of Functions on Sets , 2021, J. Mach. Learn. Res..

[6]  Marcus Hutter,et al.  On Representing (Anti)Symmetric Functions , 2020, ArXiv.

[7]  Michael A. Osborne,et al.  On the Limitations of Representing Functions on Sets , 2019, ICML.

[8]  Yee Whye Teh,et al.  Set Transformer , 2018, ICML.

[9]  Bryan K. Clark,et al.  Backflow Transformations via Neural Networks for Quantum Many-Body Wave Functions. , 2018, Physical review letters.

[10]  Max Welling,et al.  Spherical CNNs , 2018, ICLR.

[11]  Razvan Pascanu,et al.  A simple neural network module for relational reasoning , 2017, NIPS.

[12]  Alexander J. Smola,et al.  Deep Sets , 2017, 1703.06114.

[13]  Amit Daniely,et al.  Depth Separation for Neural Networks , 2017, COLT.

[14]  Ohad Shamir,et al.  The Power of Depth for Feedforward Neural Networks , 2015, COLT.

[15]  Yoshua Bengio,et al.  Unitary Evolution Recurrent Neural Networks , 2015, ICML.

[16]  Jiang Zeng,et al.  Generalizations of Cauchy's determinant and Schur's Pfaffian , 2004, Adv. Appl. Math..

[17]  E. Langmann.  A method to derive explicit formulas for an elliptic generalization of the Jack polynomials , 2005, math-ph/0511015.

[18]  Attila Maróti,et al.  On elementary lower bounds for the partition function. , 2003 .

[19]  Yoshua Bengio,et al.  Convolutional networks for images, speech, and time series , 1998 .

[20]  Thomas S. Sundquist,et al.  Two Variable Pfaffian Identities and Symmetric Functions , 1996 .

[21]  A. Szabó,et al.  Modern quantum chemistry : introduction to advanced electronic structure theory , 1982 .

[22]  R. Jastrow.  Many-Body Problem with Strong Forces , 1955 .