Search papers, labs, and topics across Lattice.
This paper explores the application of Tensor Networks, specifically matrix product states (MPS), in machine learning by introducing a global normalization condition to ensure MPS accurately represent quantum states. The authors evaluate two optimization techniques鈥攖raditional gradient descent and a modified Density Matrix Renormalization Group (DMRG) method鈥攁ssessing their performance in locating locally optimal tensors. Key findings reveal that the DMRG adaptation significantly enhances optimization effectiveness compared to standard gradient descent, marking a notable advancement in the intersection of quantum physics and machine learning.
A modified DMRG method outperforms gradient descent in optimizing tensor networks for quantum state representation, revealing new potential in machine learning applications.
Tensor Networks are a relatively new machine learning approach. The architectures proposed initially are inspired by approaches from quantum many-body physics simulations. One common layout is the matrix product state (MPS) also known as a tensor train optimized with gradient descent techniques. We introduce a global normalization condition, so that the MPS represents a quantum state. We investigate two optimization methods that find the locally optimal tensors and compare them regarding their effectiveness. One is based on gradient descent and the other on an adaptation of DMRG.
