FlowCred: Hierarchical Credit Assignment and Functional Coordination for Joint Text-to-Video Post-Training

Key points

• Prompt optimization and generator post-training are two complementary ways to improve text-to-video (T2V) generation, and recent systems increasingly train both modules jointly.1
• However, optimizing a prompt policy and a video generator from the same terminal reward creates two ambiguities: prompt quality is confounded with stochastic generation, and the two modules may realize overlapping local improvements.1
• First, hierarchical credit assignment uses matched prompt-seed rollout grids to separate prompt-level credit from prompt-conditioned generator residuals under shared randomness.1
• Second, functional coordination probes alternative current-policy prompts at states from actual generation trajectories, providing generator-aware evidence for prompt optimization while selectively attenuating generator updates already supported by conditioning changes.1

Sources (1)

• [1]FlowCred: Hierarchical Credit Assignment and Functional Coordination for Joint Text-to-Video Post-Training
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 13, 10:48 AM

  We introduce FlowCred, a two-level framework for joint T2V post-training.

  Prompt optimization and generator post-training are two complementary ways to improve text-to-video (T2V) generation, and recent systems increasingly train both modules jointly.

Extractive summary: sentences quoted from the sources.