- Rating
- 2(+1)
- Confidence
- 3
- Informativeness
- 1
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper proposes BiaWDRO, a conservative surrogate model learning framework for Offline Model-Based Optimization (Offline MBO), aimed at mitigating the problem of surrogate models producing overly optimistic predictions on out-of-distribution (OOD) data, which can lead the optimization process toward spurious optima. The core idea of BiaWDRO is to define and control the Worst-case Predictive Bias, which measures the maximum increase in the mean prediction of a surrogate model under distributional shifts constrained by the Wasserstein distance. The authors incorporate this metric as a regularization term into the surrogate model's training objective to suppress unreliable, excessively high predictions. Specifically, by leveraging Wasserstein duality, they transform the challenging distributional optimization problem into an adversarial sample search problem with a transportation cost penalty. Adversarial samples are generated through gradient ascent, while the surrogate model parameters and dual variables are updated alternately to train a more conservative surrogate model. Furthermore, the authors theoretically analyze the trade-off between conservatism and model fitting accuracy and conduct experiments on five Design-Bench tasks. The experimental results demonstrate that BiaWDRO achieves competitive performance on offline design optimization tasks.
Critical strengths
-
The paper addresses surrogate overestimation on out-of-distribution (OOD) samples in offline model-based optimization. It introduces a Wasserstein DRO-based conservative objective and provides a theoretical analysis of the trade-off between conservatism and prediction accuracy.
-
The method is evaluated on five Design-Bench tasks covering both continuous and discrete design spaces. BiaWDRO achieves a competitive overall average rank, suggesting its potential effectiveness across different optimization settings.
Critical weaknesses
1.The proposed method has substantial overlap with existing work, particularly Conservative Objective Models (COMs), which already use gradient-based adversarial samples and conservative regularization to mitigate surrogate overestimation. Moreover, Wasserstein duality, transport-cost penalties, and distributionally robust adversarial training have been extensively studied in prior work, such as Sinha et al. (2018). The main contribution appears to be introducing a Wasserstein transport penalty into a COMs-style training framework, with a dual variable controlling adversarial exploration. However, the paper does not clearly establish what fundamentally new capability this formulation provides over existing approaches, making the claimed methodological novelty somewhat limited.
2.More importantly, the current ablation studies do not adequately validate the core technical contribution. Although the paper examines different transport costs and Wasserstein radii, these experiments mainly investigate design choices rather than demonstrate the necessity of the Wasserstein DRO formulation. In particular, comparisons against COMs augmented with simple distance regularization, fixed transport penalties, or adaptive transport penalties are missing. It therefore remains unclear whether the improvements arise from the proposed DRO formulation or simply from regularizing the distance traveled by adversarial samples. Furthermore, the contributions of the adaptive dual variable and the adaptive conservatism weight are not separately evaluated. Without these comparisons, the empirical evidence is insufficient to establish the benefits of the proposed formulation over simpler alternatives.
3.The effectiveness of the proposed worst-case control mechanism remains insufficiently validated. While the theoretical objective considers the worst-case prediction increase within a Wasserstein uncertainty set, the practical algorithm approximates this objective using finite-step local gradient ascent. As the surrogate evolves, new spurious high-prediction regions may emerge, and it is unclear whether the adversarial search can reliably identify and suppress them. Moreover, controlling prediction increases does not necessarily reduce overestimation errors relative to the true objective. Although Figure 1 illustrates the intended suppression of spurious peaks, the experiments mainly report aggregate optimization scores and exploration distances, without directly demonstrating that this mechanism occurs in practice.
Questions for authors
- What is the essential technical contribution beyond COMs and existing Wasserstein DRO methods?
Since COMs already uses adversarial samples for conservative surrogate training, and transport-cost-based adversarial optimization has been studied previously, what fundamentally distinguishes the proposed method? In particular, could similar improvements be achieved by simply introducing a distance penalty into COMs?
- Can the authors provide ablations that directly isolate the benefits of the Wasserstein DRO formulation?
A particularly important comparison would be between COMs with simple distance regularization, fixed transport penalties, and the proposed adaptive transport penalty. Additionally, can the authors independently ablate the adaptive dual variable and the adaptive conservatism weight? These experiments would help determine whether the gains genuinely come from the proposed formulation rather than simpler regularization or adaptive weighting mechanisms.
- Can the authors provide more direct evidence that the method suppresses spurious prediction peaks?
Since the worst-case search relies on finite-step local gradient ascent, how sensitive are the results to the number of ascent steps and the use of multiple initializations? Could the authors also provide a low-dimensional synthetic example showing the ground-truth objective, surrogate prediction landscapes before and after conservative training, and optimization trajectories? Together with a direct evaluation of OOD overestimation errors, this would help verify whether the method genuinely suppresses spurious peaks rather than merely shifting their locations.
5 comments
Response to Reviewer c8tT (2/3)
Q2 Separating the effects of transport regularization and adaptive weighting
Thank you for this suggestion. We added COMs with simple distance regularization and independently ablated the transport dual variable and conservatism weight . All settings use eight random seeds. Scores are the highest true normalized score among the final 128 candidates.
(1) Independent ablation of adaptive . We fix while retaining adaptive and all other BiaWDRO settings. Our method achieves the best mean rank, as shown in Table 9 (revised PDF, Appendix C.4.1, p. 19). Although some fixed weights achieve higher scores on individual tasks, adaptive transport weighting achieves better average performance across the evaluated tasks than fixed weights.
| Method | TFBind8 | TFBind10 | Superconductor | Ant | D’Kitty | Mean rank ↓ |
|---|---|---|---|---|---|---|
| Adaptive | 0.987 ± 0.007 | 0.713 ± 0.061 | 0.513 ± 0.013 | 0.952 ± 0.012 | 0.956 ± 0.007 | 2.4 |
| Fixed γ=2 | 0.992 ± 0.001 | 0.651 ± 0.013 | 0.505 ± 0.004 | 0.952 ± 0.004 | 0.954 ± 0.001 | 4.0 |
| Fixed γ=5 | 0.991 ± 0.000 | 0.667 ± 0.018 | 0.495 ± 0.009 | 0.959 ± 0.006 | 0.949 ± 0.001 | 4.8 |
| Fixed γ=0.3 | 0.981 ± 0.004 | 0.634 ± 0.013 | 0.509 ± 0.004 | 0.950 ± 0.011 | 0.956 ± 0.003 | 5.0 |
| Fixed γ=1.5 | 0.990 ± 0.001 | 0.645 ± 0.007 | 0.508 ± 0.004 | 0.946 ± 0.004 | 0.953 ± 0.005 | 5.8 |
| Fixed γ=3 | 0.992 ± 0.001 | 0.633 ± 0.012 | 0.501 ± 0.005 | 0.947 ± 0.006 | 0.953 ± 0.001 | 5.8 |
| Fixed γ=0.1 | 0.976 ± 0.005 | 0.618 ± 0.015 | 0.512 ± 0.005 | 0.946 ± 0.005 | 0.958 ± 0.003 | 6.0 |
| Fixed γ=1 | 0.981 ± 0.004 | 0.659 ± 0.015 | 0.506 ± 0.004 | 0.941 ± 0.004 | 0.956 ± 0.004 | 5.8 |
| Fixed γ=0.5 | 0.987 ± 0.002 | 0.629 ± 0.007 | 0.509 ± 0.004 | 0.942 ± 0.004 | 0.951 ± 0.003 | 7.0 |
| Fixed γ=10 | 0.991 ± 0.000 | 0.665 ± 0.012 | 0.488 ± 0.004 | 0.937 ± 0.008 | 0.947 ± 0.001 | 7.2 |
| COMs | 0.941 ± 0.048 | 0.647 ± 0.028 | 0.465 ± 0.114 | 0.933 ± 0.030 | 0.946 ± 0.007 | 10.0 |
(2) Independent ablation of adaptive . We fix while retaining adaptive and all other settings. Although some fixed weights achieve higher scores on individual tasks, our method offers a better balance of performance across tasks and achieves the best mean rank. Table 10 (revised PDF, Appendix C.4.2, p. 19) reports the results, with ranks computed within these four settings from task means; ties receive the minimum shared rank.
| Conservatism-weight setting | TFBind8 | TFBind10 | Superconductor | Ant | D’Kitty | Mean rank ↓ |
|---|---|---|---|---|---|---|
| Full adaptive method | 0.987 ± 0.007 | 0.713 ± 0.061 | 0.513 ± 0.013 | 0.952 ± 0.012 | 0.956 ± 0.007 | 1.2 |
| Fixed λ=0.1 | 0.700 ± 0.036 | 0.632 ± 0.011 | 0.433 ± 0.013 | 0.946 ± 0.007 | 0.961 ± 0.005 | 2.6 |
| Fixed λ=0.01 | 0.980 ± 0.003 | 0.618 ± 0.009 | 0.500 ± 0.005 | 0.945 ± 0.005 | 0.950 ± 0.001 | 2.8 |
| Fixed λ=1 | 0.741 ± 0.031 | 0.666 ± 0.031 | 0.299 ± 0.030 | 0.915 ± 0.010 | 0.706 ± 0.010 | 3.4 |
(3) Simple distance regularization versus BiaWDRO with fixed transport weights. The complete comparison is provided in the table in our response to Q1 above (Table 11, Appendix C.4.3, p. 20). With both transport weights fixed and adaptive retained, BiaWDRO achieves a better mean rank than distance-regularized COMs. Its advantage therefore does not depend entirely on adaptive transport weighting.
Together, these comparisons support the practical value of the DRO-derived training scheme beyond simple distance regularization and separately demonstrate the contributions of the two adaptive updates.
Response to Reviewer c8tT (3/3)
Q3 Direct evaluation of overestimation, search sensitivity, and a low-dimensional example
Thank you for your suggestions. We directly evaluate overestimation along optimization trajectories and examine search sensitivity and a low-dimensional example.
(1) Direct evaluation along optimization trajectories. We examine whether reduced overestimation accompanies improvement in true scores during optimization. On D’Kitty, we compare BiaWDRO and COMs over eight seeds, starting from the same 128 designs within each seed. At each step, both models evaluate the pooled designs generated by both methods, enabling prediction-error comparisons on identical inputs. We also evaluate the true objective values of each method’s 128 candidates at every step and report their mean to track progress in actual design quality.
Figure 8 (revised PDF, Appendix C.6.1, p. 23) presents three complementary views.
- Panel (a) shows the mean true score of each method’s candidates: COMs plateaus after an initial improvement, whereas BiaWDRO continues to improve and reaches a higher final value.
- Panel (b) shows the difference in mean prediction error, COMs minus BiaWDRO, on identical inputs, which grows during optimization.
- Panel (c) directly measures positive overestimation on these same inputs, showing that COMs’ overestimation increases more rapidly.
Together, these results show that BiaWDRO achieves lower overestimation alongside continued improvement in true scores, supporting the practical benefit of balancing conservatism and exploration through predictive-bias regularization.
(2) Sensitivity to the number of inner steps. Across five tasks, we vary the number of gradient-ascent steps in the training inner loop, keeping the other training settings and final design-optimization procedure unchanged. Every configuration uses eight random seeds.
The following table reports the mean across seeds of the highest true normalized score among the final 128 candidates for different numbers of inner steps, .
The inner-step sensitivity results are reported in Table 12 (revised PDF, Appendix C.5, p. 20) and reproduced below.
| Task | |||||
|---|---|---|---|---|---|
| Ant | 0.95106 | 0.95202 | 0.95073 | 0.95812 | — |
| D’Kitty | 0.94653 | 0.95576 | 0.96228 | 0.95309 | — |
| Superconductor | 0.50743 | 0.51322 | 0.48208 | 0.48125 | — |
| TFBind8 | 0.66029 | 0.96055 | 0.99253 | 0.98672 | 0.95637 |
| TFBind10 | 0.52463 | 0.65508 | 0.66869 | 0.71321 | 0.64984 |
Performance is relatively stable under moderate changes in the number of inner steps, and increasing the step count does not consistently improve performance.
(3) Sensitivity to the number of search starts. We compare one, three, and five search starts at the default number of steps. The first start is the source design; additional starts receive independent Gaussian perturbations with standard deviation . Each start returns its final iterate, and the inner objective selects among these results. Transport costs are always calculated relative to the original source design.
The multistart sensitivity results are reported in Table 13 (revised PDF, Appendix C.5, p. 20) and reproduced below.
| Task | Single start | Three starts | Five starts |
|---|---|---|---|
| Ant | 0.95202 | 0.94498 | 0.93840 |
| D’Kitty | 0.95576 | 0.96100 | 0.96199 |
| Superconductor | 0.51322 | 0.50266 | 0.49472 |
| TFBind8 | 0.98672 | 0.98759 | 0.98582 |
| TFBind10 | 0.71321 | 0.70265 | 0.70070 |
Across all five tasks, mean true scores vary by less than 0.02 across the multistart settings, indicating limited sensitivity to this setting.
(4) Low-dimensional example: visualizing the relationship between the true objective and search directions. We conducted a 2D toy experiment on the Branin function. The surrogate is trained on approximately 4,000 offline samples (sampled from 5,000 points with the top and bottom 10% removed), intentionally leaving the high-value regions as OOD. Figure 9 (revised PDF, Appendix C.6.2, p. 23) plots optimization trajectories of both COMs and BiaWDRO on the true objective contour. As shown, BiaWDRO finds better optima than COMs while avoiding being misled into the low-value OOD regions (bottom 10%).
Response to Reviewer c8tT (1/3)
Thank you for your careful review and constructive comments, which we address point by point below.
Q1 Fundamental contribution relative to COMs and existing Wasserstein DRO methods
We propose a distributionally robust conservative learning criterion for offline model-based optimization. In supervised learning, WDRO minimizes the maximum expected prediction loss under distribution shift. When inputs are perturbed while their labels are retained, this objective encourages predictions at the perturbed inputs to remain close to the original labels. However, optimizing a design can change its true objective value, so this label preservation can suppress prediction differences that guide the search for better designs. Our key modeling insight is that the maximum increase in expected surrogate prediction over an ambiguity set of input distributions admits a DRO formulation. We use this quantity as a conservative regularizer alongside data fitting, thereby obtaining a DRO learning objective for offline MBO.
Relative to COMs, BiaWDRO uses a Wasserstein ambiguity set to specify the distribution shifts over which its prediction difference regularizer is maximized. This allows more flexible adjustment of conservatism: the transport cost and radius specify the geometry and extent of the distribution shifts being regularized, while the regularization weight controls the penalty strength.
To evaluate the effectiveness of this DRO-derived training scheme, we compare BiaWDRO with COMs augmented by a distance penalty. Specifically, we add −βc(x~,x) to COMs' training adversarial objective, where c(x~,x)=∥x~−x∥22/d, and compare it with fixed-γ BiaWDRO. Both methods use nine fixed penalty weights: 0.1, 0.3, 0.5, 1, 1.5, 2, 3, 5, and 10.
The complete results are reported in Table 11 (revised PDF, Appendix C.4.3, p. 20) and reproduced below. Scores are the highest true normalized score among the final 128 candidates, averaged across eight seeds. Mean ranks are computed jointly among the 18 settings.
The six settings with the best mean ranks are all fixed-γ BiaWDRO variants. The best mean rank is 5.2 for BiaWDRO, compared with 8.2 for distance-regularized COMs. These results support the effectiveness of BiaWDRO's DRO-derived training scheme beyond simply adding a fixed distance penalty to COMs.