Pre-trained Text-to-Image Diffusion Models Are Versatile Representation Learners for Control
This addresses the challenge of insufficient fine-grained scene understanding in embodied AI agents, which is vital for control tasks, by leveraging diffusion models for improved representation learning.
The paper tackles the problem of enabling embodied AI agents to achieve fine-grained scene understanding for control by using pre-trained text-to-image diffusion models to create Stable Control Representations, resulting in policies that are competitive with state-of-the-art methods and achieve state-of-the-art performance on the OVMM benchmark.
Embodied AI agents require a fine-grained understanding of the physical world mediated through visual and language inputs. Such capabilities are difficult to learn solely from task-specific data. This has led to the emergence of pre-trained vision-language models as a tool for transferring representations learned from internet-scale data to downstream tasks and new domains. However, commonly used contrastively trained representations such as in CLIP have been shown to fail at enabling embodied agents to gain a sufficiently fine-grained scene understanding -- a capability vital for control. To address this shortcoming, we consider representations from pre-trained text-to-image diffusion models, which are explicitly optimized to generate images from text prompts and as such, contain text-conditioned representations that reflect highly fine-grained visuo-spatial information. Using pre-trained text-to-image diffusion models, we construct Stable Control Representations which allow learning downstream control policies that generalize to complex, open-ended environments. We show that policies learned using Stable Control Representations are competitive with state-of-the-art representation learning approaches across a broad range of simulated control settings, encompassing challenging manipulation and navigation tasks. Most notably, we show that Stable Control Representations enable learning policies that exhibit state-of-the-art performance on OVMM, a difficult open-vocabulary navigation benchmark.