- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
The paper jointly adapts a scene generator and a PointNav policy. Policy improvement on fixed before/after episodes provides scene-level weights for updating a diffusion generator through LoRA. Experiments evaluate navigation across five distributions and analyze changes in generated layouts and policy behavior.
Critical strengths
-
The external-domain results are promising. In Table 1, MP3D success improves from 73.94% to 83.64%, and SPL from 51.38% to 59.79%, over Frozen Generator with the same policy training budget. This supports the practical value of adapting training environments.
-
Table 2 includes selection-only and shuffled-feedback controls. Full performs better on both external domains, providing useful evidence that the scene-specific feedback matters.
Critical weaknesses
-
Tables 1–2 do not report the number of independent training runs or variation across seeds. Both policy learning and scene generation are stochastic, so I am unsure how reliable the gains are. The bootstrap analyses elsewhere do not measure training-run variation.
-
The central generator update needs more detail. Eqs. 10–11 do not clearly explain how the target noise is constructed from the retained generation record. The utility weights and clipping settings are also missing, making the method difficult to reproduce.
Questions for authors
-
Clarification and supporting evidence: How many independent training seeds were used for Tables 1–2? Can you report mean and standard deviation for Full, Frozen Generator, and Shuffled-feedback on Gibson and MP3D?
-
Clarification: How are the noisy state and target noise in Eqs. 10–11 constructed? Are generated layouts forward-noised, or are reverse-sampling states reused? Please also specify the utility weights and clipping settings.
2 comments
Thanks for the rebuttal. I think the three-seed results and the clarified generator update address both of my concerns. I think I would raise my rating to 3.
Point-to-Point Response to Reviewer U1eT
We thank reviewer U1eT for the valuable time and constructive feedback. We provide point-to-point responses below.
Q1: Additional training seeds and result variability.
A1: The results in Tables 1–2 were obtained from one training run per method. We conduct two additional independent runs for Full, Frozen Generator, and Shuffled-feedback, bringing the total to three training seeds per method. The table below reports mean ± sample standard deviation across training runs on Gibson and MP3D.
Full achieves the highest mean SR and SPL on both external domains, with lower standard deviations than both controls across all four metrics. These results support the consistency of the gains across independent training runs. We have added the results to supp. §C.2, Table 11.
Q2: Clarification on generator updates, utility weights, and clipping settings.
A2: We forward-noise the completed generated layouts using freshly sampled noise, which serves as the denoising target. Reverse-sampling states are not reused. The SR, SPL, and normalized-DTG coefficients are 1, 0.5, and 0.1, respectively. We detail the construction and weighting settings below.
First, generated layouts are forward-noised; intermediate reverse-sampling states are not reused. We retain each completed generated layout, x0, and its condition, c. At each adaptation update, we sample a timestep t and fresh standard Gaussian noise ϵ∼N(0,I), and construct the noisy input as
xt=αˉtx0+1−αˉtϵ.
Here, αˉt is the cumulative signal-retention coefficient at timestep t. The sampled ϵ is the target noise. The network's v-prediction is converted to a noise prediction before computing the feedback-weighted noise MSE. We have clarified this construction and the retained records in §3.5 and supp. §A.4.
Second, we specify the utility weights and clipping settings. The navigation score is
Qi=SRi+0.5SPLi−0.110clip(DTGi,0,10).
SR and SPL are in [0,1]. The term DTGi is the mean distance-to-goal in meters across the scene's evaluation episodes. The function clip(value, lower, upper) limits the value to the specified interval. Feedback and weights are computed in the following order:
where P is the eligible fitting pool, and the denominator in the final expression is the mean of uj over that pool. Normalization precedes the final weight clipping, so the final weights need not have exactly unit mean.
The coefficients were selected on the ProcTHOR validation set. The selection procedure and sensitivity results are provided in supp. §C.4. We have also specified the feedback-weighting settings and operation order in §3.5, supp. §A.4, and Table 8.