Search papers, labs, and topics across Lattice.
University of Tübingen, Technical University of Munich
3
0
6
Models trained on LAION-BVD achieve state-of-the-art performance in multimodal tasks, showcasing the dataset's potential to redefine video understanding.
DF-MoE achieves superior deepfake detection performance by harnessing a diverse set of multimodal features, setting a new standard in the field.
The dream of universal representations across modalities may be just that: scaling up datasets and relaxing constraints reveals that models trained on different modalities learn rich, but fundamentally different, representations of the world.
