Enriching text embedding with flexibility and resilience through stochastic modeling enhances text-video retrieval performance.
