报告摘要： I introduce an exact, non‑perturbative framework for transforming many‑body operators, $O_{\eta} = e^{-\eta} O e^{\eta}$, that overcomes the limitations of the Baker–Campbell–Hausdorff expansion by leveraging the continuous exact solution of few‑particle sectors. The generator $\eta$ is constructed directly from the few‑body physics, with a strictly bounded norm $∥\eta∥<\alpha$, ensuring full control over the unitary flow. This procedure systematically eliminates high‑energy off‑diagonal couplings, producing a transformed Hamiltonian that is maximally sparse while preserving the full spectrum. Crucially, the flow rotates into a basis where many‑body eigenstates exhibit drastically reduced entanglement and are strongly localized in the Slater determinant basis, effectively killing entanglement entropy between distant sites. This sparsity and low entanglement make the Hamiltonian exceptionally well‑suited for large‑scale numerical methods such as DMRG and FCI‑QMC. I benchmark the approach on the Hubbard model and discuss its natural extensions to other quantum problems.

个人简介：Abolhassan Vaezi is a faculty member at Sharif University of Technology. He received his Bachelor's and Master's degrees from Sharif University of Technology and his PhD from MIT, where he worked with Prof. Xiao-Gang Wen on topological order and strongly correlated systems. After a year as a Research Scientist at Sharif University and the Institute for Research in Fundamental Sciences (IPM), he joined Cornell University as a Bethe Postdoctoral Fellow, working with Prof. Eun-Ah Kim. He later became a Moore Postdoctoral Fellow at Stanford University, where he worked with Profs. Shoucheng Zhang and Xiaoliang Qi. He returned to Sharif University of Technology as a faculty member in 2019. His research spans theoretical and computational condensed matter physics, with interests in topological quantum computing and topological order, quantum Hall physics, high-temperature superconductivity, and strongly correlated electron systems. He also develops numerical methods for quantum many-body problems, including DMRG, quantum Monte Carlo, and machine learning approaches.

邀请人：孙孝奇 xqsun@iphy.ac.cn