Inference-Time Policy Steering through Human Interactions
This work addresses the challenge of enabling human guidance for autonomous policies in multimodal, long-horizon tasks, representing an incremental improvement by focusing on inference-time adjustments rather than fine-tuning.
The paper tackles the problem of aligning pre-trained generative policies with human intent during inference without causing distribution shift, proposing an Inference-Time Policy Steering framework that uses human interactions to bias sampling. The method achieves the best trade-off between alignment and distribution shift among tested strategies in simulated and real-world benchmarks.
Generative policies trained with human demonstrations can autonomously accomplish multimodal, long-horizon tasks. However, during inference, humans are often removed from the policy execution loop, limiting the ability to guide a pre-trained policy towards a specific sub-goal or trajectory shape among multiple predictions. Naive human intervention may inadvertently exacerbate distribution shift, leading to constraint violations or execution failures. To better align policy output with human intent without inducing out-of-distribution errors, we propose an Inference-Time Policy Steering (ITPS) framework that leverages human interactions to bias the generative sampling process, rather than fine-tuning the policy on interaction data. We evaluate ITPS across three simulated and real-world benchmarks, testing three forms of human interaction and associated alignment distance metrics. Among six sampling strategies, our proposed stochastic sampling with diffusion policy achieves the best trade-off between alignment and distribution shift. Videos are available at https://yanweiw.github.io/itps/.