New method learns shared and unique patterns across multiple AI data sources

Original: RePercENT: Scaling Disentangled Representation Learning Beyond Two Modalities

Source: arxiv.org ↗