I would like to extend my sincere appreciation for the remarkable work you have conducted in your paper.
Upon careful review of your work, I notice the relevance of our previous research endeavors in relation to the topic you have explored. In our studies [1, 2], , we primarily focuses on scenarios where the dataset consists predominantly of sub-optimal trajectories. In such cases, a straightforward application of the policy might inadvertently lead to the imitation of suboptimal actions. To address this, we proposed a sampling strategy designed as a plug-in mechanism. This approach effectively constrains the policy, ensuring it is guided by “good data” rather than uniform sampling.
Given the apparent synergy between our works, I kindly request the inclusion of citations to our papers in your publication. I believe that acknowledging these references will enrich the context of your work and provide a comprehensive perspective to your readers.
[1] Yue, Yang, et al. "Boosting Offline Reinforcement Learning via Data Rebalancing." NIPS 2022, Offline RL Workshop.
[2] Yue, Yang, et al. "Offline Prioritized Experience Replay." arXiv preprint arXiv:2306.05412 (2023).
I would like to extend my sincere appreciation for the remarkable work you have conducted in your paper.
Upon careful review of your work, I notice the relevance of our previous research endeavors in relation to the topic you have explored. In our studies [1, 2], , we primarily focuses on scenarios where the dataset consists predominantly of sub-optimal trajectories. In such cases, a straightforward application of the policy might inadvertently lead to the imitation of suboptimal actions. To address this, we proposed a sampling strategy designed as a plug-in mechanism. This approach effectively constrains the policy, ensuring it is guided by “good data” rather than uniform sampling.
Given the apparent synergy between our works, I kindly request the inclusion of citations to our papers in your publication. I believe that acknowledging these references will enrich the context of your work and provide a comprehensive perspective to your readers.
[1] Yue, Yang, et al. "Boosting Offline Reinforcement Learning via Data Rebalancing." NIPS 2022, Offline RL Workshop.
[2] Yue, Yang, et al. "Offline Prioritized Experience Replay." arXiv preprint arXiv:2306.05412 (2023).