Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper investigates the relationship between entanglement geometry and the properties necessary for achieving variational quantum advantage through circuit cutting. The authors demonstrate that while certain matrix product state (MPS) and tree tensor network (TTN) circuits can be efficiently simulated, they do not provide asymptotic quantum advantage due to their classical simulability. By introducing a two-block circuit family that separates seam and intra-block entanglement, they show that using magic as a hardness resource allows for circuits that are both cuttable and trainable, circumventing the depth regime conflicts observed in MPS hardness and trainability.
Shallow Clifford+T circuits can be both efficiently cuttable and trainable, defying traditional limits imposed by entanglement geometry.
Circuit cutting promises to scale quantum computations beyond current hardware, but variational quantum advantage also requires low cutting overhead, classical hardness, and trainability. We show that these properties are strongly constrained by entanglement geometry. Matrix product state (MPS) and tree tensor network (TTN) circuits with constant seam bond dimension can be cut with \(O(1/\varepsilon^2)\) sampling overhead, but remain efficiently classically simulable, ruling out asymptotic quantum advantage within these families. By independently controlling seam and intra-block entanglement, we construct a two-block circuit family that remains cheaply cuttable while requiring a super-polynomial global MPS bond dimension, as supported numerically up to \(n=100\). However, MPS hardness and trainability require incompatible depth regimes, \(d=\omega(\log n)\) and \(d=O(\log n)\), respectively. Using magic rather than entanglement as the hardness resource avoids this conflict: shallow Clifford+\(T\) circuits remain cuttable and trainable while their stabiliser-simulation cost grows exponentially with the \(T\)-count.