Learning generalizable robotic manipulation policies remains a key challenge due to the scarcity of diverse real-world training data. While recent approaches have attempted to mitigate this through self-supervised representation learning, most either rely on 2D vision pretraining paradigms such as masked image modeling, which primarily focus on static semantics or scene geometry, or utilize large-scale video prediction models that emphasize 2D dynamics, thus failing to jointly learn the geometry, semantics, and dynamics required for effective manipulation. In this paper, we present DynaRend, a representation learning framework that learns 3D-aware and dynamics-informed triplane features via masked reconstruction and future prediction using differentiable volumetric rendering. By pretraining on multi-view RGB-D video data, DynaRend jointly captures spatial geometry, future dynamics, and task semantics in a unified triplane representation. The learned representations can be effectively transferred to downstream robotic manipulation tasks via action value map prediction. We evaluate DynaRend on two challenging benchmarks, RLBench and Colosseum, as well as in real-world robotic experiments, demonstrating substantial improvements in policy success rate, generalization to environmental perturbations, and real-world applicability across diverse manipulation tasks.

The rising number of innovative start-up operations working within the domain of AI powered tools and services is one of the key factors driving the growth within the global artificial intelligence as a service market. The solutions offered by the players and vendors functioning within the global artificial intelligence as a service market are utilized in a number of end use industry verticals, such as healthcare and life sciences, telecommunications, manufacturing, education, transportation, media and entertainment, banking, financial services, and insurance or BFSI, retail, government and defence, energy, and agriculture, among others. Some of the key technologies used by the players in the global artificial intelligence as a service market include deep learning, natural language processing or NLP, and machine learning or ML. The rising demand from the BFSI industry vertical is positively influencing the growth in the global artificial intelligence as a service market. On the other hand, healthcare and life sciences end use industry vertical is also expected to contribute heavily in the development of the global artificial intelligence as a service market in coming years.

The solutions offered by the players and vendors functioning within the global artificial intelligence as a service market are utilized in a …