QFlowNet: Fast, Diverse, and Efficient Unitary Synthesis with Generative Flow Networks
1 st Inhoe Koo, 2 nd Hyunho Cha, 3 rd Jungwoo Lee, Inhoe Koo, Hyunho Cha, and Jungwoo Lee · 2026
Read QFlowNet: Fast, Diverse, and Efficient Unitary Synthesis with Generative Flow Networks by 1 st Inhoe Koo, 2 nd Hyunho Cha, 3 rd Jungwoo Lee, Inhoe Koo, Hyunho Cha, and Jungwoo Lee (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-03045.txt.
Description
Inhoe Koo, Hyunho Cha, and Jungwoo Lee build quantum circuits with a learner that prizes variety as well as success. Their 2026 framework, QFlowNet, tackles unitary synthesis: breaking a target quantum operation into a short gate sequence. Reinforcement learners often chase one best answer and struggle when rewards are rare.
A generative flow network instead learns to sample solutions in proportion to their reward, yielding many diverse circuits. A transformer encoder reads the target operation's nonlocal structure and compresses it for the policy. On a three-qubit benchmark, the agent succeeded on nearly every case across short circuit lengths. It also produced a varied set of compact circuits, offering compilers real choices rather than a single take-it-or-leave-it answer.
Catalog record
| CREATOR | 1 st Inhoe Koo, 2 nd Hyunho Cha, 3 rd Jungwoo Lee, Inhoe Koo, Hyunho Cha, and Jungwoo Lee |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 31350 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.03045