发表机构
Imperial College London; Aalborg University; Centre for Frontier AI Research (CFAR); Institute of Advanced Intelligence and Computing (IAIC); A*STAR; University of Copenhagen(帝国理工学院; 奥尔堡大学; 前沿人工智能研究中心; 先进智能与计算研究所; 新加坡科技研究局; 哥本哈根大学)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
提出基于离散Wasserstein几何的KL梯度流框架,通过马尔可夫跳跃实现有限状态空间上的一步生成建模,验证了动力学一致性与生成器跟踪能力。
AI 中文摘要
我们提出了一种在有限状态空间上进行一步生成建模的新框架。为了将漂移扩展到连续域之外,我们利用离散Wasserstein几何,在可逆马尔可夫核的转移上定义了一个目标相对KL梯度流。我们通过马尔可夫跳跃在粒子层面实现这一概率流,并将由此产生的传输更新摊销到一个潜在条件生成器中,使得迭代动力学仅在训练期间需要,而推理保持一步完成。在可以精确计算底层分布和传输动力学的受控设置中,我们验证了KL耗散、粒子动力学与概率流之间的一致性,以及预测的数值标度。我们进一步证明,有限容量的神经生成器能够跟踪这些精确的传输目标,同时保持一步生成。这些结果验证了基本构造,并为将离散漂移扩展到结构化离散数据奠定了基础。
英文摘要
We introduce a new framework for one-step generative modelling on finite state spaces. To extend drifting beyond continuous domains, we use discrete Wasserstein geometry to define a target-relative KL gradient flow over the transitions of a reversible Markov kernel. We realize this probability flow at the particle level through Markov jumps and amortize the resulting transport updates into a latent-conditioned generator, so that the iterative dynamics are required only during training while inference remains one-step. In a controlled setting where the underlying distributions and transport dynamics can be computed exactly, we verify KL dissipation, consistency between the particle dynamics and the probability flow, and the predicted numerical scaling. We further show that a finite-capacity neural generator can track these exact transport targets while retaining one-step generation. These results validate the basic construction and provide a foundation for scaling Discrete Drifting to structured discrete data.