Search papers, labs, and topics across Lattice.
XGenAct enhances world action models by representing diverse modalities (RGB, actions, depth, normals, segmentation) as videos via deterministic codecs, enabling unified temporal prediction with a single diffusion transformer and objective. This geometry-enhanced approach, trained with cross-task generation, significantly improves robot manipulation success rates on RLBench tasks compared to RGB-only models and baselines that rely on separate perception modules. XGenAct also demonstrates superior accuracy in predicting future depth and segmentation maps.
Unified prediction of future RGB, depth, normals, and segmentation from actions unlocks substantial gains in robot manipulation success.
World action models (WAMs) have advanced robot control by predicting how observations and actions evolve over time. Despite this progress, RGB and action based future prediction does not explicitly address the spatial understanding needed for robot manipulation. Existing efforts often add a limited set of spatial prediction tasks through specialized heads or branches, leaving both the range of spatial supervision and the model architecture fragmented. We introduce XGenAct, a world action model that represents RGB observations, robot actions, metric depth, surface normals, and functional role segmentation as RGB videos through deterministic codecs. By sampling perception and action tasks during training, XGenAct uses one video diffusion transformer and one objective to learn temporal prediction across these spaces without modality specific learned heads. On held out RLBench tasks, structured perception training improves average closed loop success over RGB only training, and XGenAct achieves 52% success in the five task external comparison, versus 26% for the strongest evaluated baselines. It also predicts future depth and segmentation more accurately than the evaluated pipelines that generate RGB first and then apply a frozen perception expert.
