- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 1
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
The paper studies discrete Schrödinger-bridge neural samplers and generalizes the DASBS controller target. DASBS relies on a deterministic cyclic translation between reference endpoint kernels; this paper replaces it with any transport-compatible coupling. For coordinate-local transitions whose endpoint kernels share a context marginal, it derives a coupling-free, fiberwise Rao–Blackwellized target with cost. Validation covers Ising and 4-state Potts models at three temperatures, a target-variance and ablation study, and one non-additive Potts reference.
Critical strengths
-
An exact, lower-variance controller target at no extra network cost. Thm 4.2 and Eq. (15) show that is unbiased and removes both auxiliary randomness sources. Fig. 2 shows roughly 1–2 orders of magnitude variance reduction at early bridge times. Eq. (16) uses the same local energy ratios and the same single corrector forward pass as DASBS, so it is a drop-in change with the same training time (Table 2).
-
Correct and transparent theory. Thms 4.1–4.2, A.3 and C.1–C.4 are short and check out. Coupling-independence (C.4) and the reduction to factorized kernels (C.3) are stated precisely, which makes the scope of the result easy to audit.
-
Strong sample quality on the standard benchmarks. TAM is best or near-best among uniform-based samplers in most cells of Tables 2–3. At criticality it improves even on the values reported in the DASBS paper (e.g., critical Ising EW2 2.3 vs 12.1; critical Potts EW2 1.0 vs 20.0). Fig. 1 shows that this agreement holds at the distribution level, not just for scalar metrics.
Critical weaknesses
-
Removing the cyclic structure seems incremental, and the transport framing adds limited value. Extending the DASBS adjoint-matching construction to other compatible group actions or kernels appears straightforward. Generalizing beyond cyclic maps therefore seems incremental, and I am not convinced that the transport framing provides a substantial conceptual advance or that the proposed kernels offer a clear practical advantage.
-
The practical advantage over DASBS is not convincingly established. The DASBS row in Tables 2–3 differs from its original paper, while every other baseline row is unchanged. Examples include critical Ising ( here vs originally), (53.7 vs 12.1), and critical Potts ( vs ). The paper does not clarify whether DASBS was rerun, its training budget (3.0k vs 3.75k steps), or its seed count. Against the published DASBS results, some gains shrink or reverse: high-temperature Ising is for DASBS versus for TAM, while critical Potts is versus . A direct comparison between TAM and the DASBS deterministic-coupling target under the same pipeline, budget, and multiple seeds is also missing; Table 4 compares only against independent coupling, a weaker reference point. Finally, variance reduction should improve training efficiency, but no convergence curves or step-matched comparisons establish this benefit.
-
The experimental scale and baseline coverage are limited. The experiments reuse DASBS’s main benchmarks: Ising (, ) and Potts (, ), with approximately 144k parameters and no scaling study in lattice or alphabet size. Since the fiberwise target averages over symbols, larger alphabets are particularly relevant, yet neither its variance gains nor its evaluation cost is tested beyond . All non-DASBS baseline results—LEAPS, UDNS, DNFS, MDNS, and MH—are copied from DASBS without reruns or variability estimates. MDNS still performs better on high-temperature Ising ( vs ) and (0.1 vs 1.9), low-temperature Ising ( vs ), and low-temperature Potts ( vs ). MH is better on all three high-temperature Ising metrics, so the “best” claims depend on restricting comparisons to uniform-based neural samplers. Newer uniform-diffusion samplers cited by DASBS, such as PDNS (Guo et al., ICLR 2026), are absent. Moreover, all tested references remain coordinate-factorized, including the non-additive example, leaving the broader transport claim insufficiently explored.
I would be happy to increase my score if the authors convincingly demonstrate that the proposed generalization to broader group actions or kernels is practically useful. For example, by enabling useful new reference dynamics, improving scalability or computational efficiency, or showing a clear training advantage over DASBS under matched conditions.
Questions for authors
-
Novelty and practical usefulness. Could the authors clarify what TAM adds beyond extending DASBS adjoint matching to other compatible group actions or kernels? Which useful reference dynamics does this generalization enable, and can they improve scalability, computational efficiency, or training performance? In particular, what new capability does the non-additive experiment in Table 5 demonstrate compared with existing DASBS constructions?
-
Fair comparison and training speed. Was DASBS rerun, and with which hyperparameters, training budget, and random seeds? Why do its results differ from the original paper—for example, critical Ising of versus ? Please compare TAM with the original DASBS deterministic-coupling target under the same pipeline and training budget, report the mean and standard deviation across seeds, and include training curves such as versus gradient steps. Does TAM converge faster, or does it only achieve slightly better results at a fixed budget?
-
Scaling and stronger baselines. Do TAM's gains over DASBS persist, grow, or disappear with larger lattices (e.g., Ising ) and larger alphabets (e.g., Potts )? How does TAM compare with MDNS at the same training cost in these settings? These results would show whether the method offers a useful advantage beyond the small DASBS benchmarks or whether its benefits are limited to those settings.
3 comments
Thank you for the clarification and additional experiments. The Ehrenfest example clarifies which reference dynamics TAM can support, and the matched DASBS comparison shows the practical benefit of the fiberwise training target. These results address my main concerns, and I am raising my score to 3.
We sincerely thank the reviewer for their thoughtful assessment, constructive feedback, and time devoted to reviewing our work. We greatly appreciate the reviewer's recognition of our contributions and willingness to raise the rating.
We are grateful for the reviewer's careful consideration and support.
Response to Reviewer VCAh
We thank the reviewer for the constructive assessment. We clarify TAM's relationship to DASBS and add matched controller-target comparisons, training curves, larger-scale benchmarks, and non-additive references.
1. Generalization beyond deterministic cyclic shifts
The DASBS deterministic cyclic-shift target is a special case of TAM: choosing the coupling induced by that shift recovers its endpoint correspondence. TAM generalizes this construction through the marginal compatibility condition KxCtx→y=Ky. Theorem 4.1 shows that any Markov coupling satisfying this condition, deterministic or stochastic, yields the exact population identity for the optimal-control potential ratio. This identity requires neither a group structure nor coordinate-local transitions.
For coordinate-local transitions whose endpoint kernels share a context marginal, we further derive a coupling-free, fiberwise Rao--Blackwellized target. It marginalizes both the auxiliary transport variable and the local endpoint symbol, preserves the population identity, and has O(N) evaluation cost without constructing a transport matrix.
The non-additive Potts experiment demonstrates the practical value of this generalization. With the nonuniform refresh distribution ρ=(0.1,0.2,0.3,0.4), endpoint probabilities depend on the absolute symbol, so cyclic translation no longer aligns the endpoint kernels. TAM remains computable from local endpoint probabilities and corrector ratios. Its energy Wasserstein distance is 1.4±0.2, compared with 220.79±1.64 for DASBS-DM. We also evaluate a 64-site ordinal chain with five states per site and an Ehrenfest birth--death reference, whose state-dependent adjacent-symbol transitions do not admit the original cyclic-shift construction. With matched reference dynamics, architecture, and budget, TAM obtains EW2=0.34±0.02, compared with 24.58±0.30 for DASBS-DM.
2. Matched DASBS comparison and training behavior
The discrepancy in the critical-point DASBS values arises from different versions of that work. The reviewer quotes the arXiv results (EW2=12.1 for critical Ising and 20.0 for critical Potts), whereas Tables 2--3 of our submission use the published ICML 2026 values (53.7 and 14.6, respectively). Other baseline results in these tables are unchanged.
We reproduced the DASBS cyclic-shift target under the same environment, architecture, training configuration, and three seeds as TAM; the results are reported in the Appendix. The direct controller-target ablation further isolates the target construction. On critical Ising, DASBS obtains EW2=54.68±54.35, compared with 2.3±0.1 for fiberwise TAM; on critical Potts, the corresponding values are 16.94±9.75 and 1.0±0.04.
We added five-stage training curves. On critical Ising (L=48), TAM's EW2 decreases from 21.53 to 12.45, while DASBS remains above 600. On eight-state Potts, TAM reaches 3.27 after the second stage and remains below 3.3, whereas DASBS decreases from 341.45 to 31.42 over five stages. End-to-end training times are comparable: 41.56±0.12 versus 44.04±0.06 minutes on Ising, and 7.79±0.02 versus 7.82±0.06 minutes on Potts.
3. Scaling and target-variance evidence
We extended the matched-budget comparison to Ising (L=48) and eight-state Potts (N=8). Over three seeds, TAM reduces EW2 from 727.92±97.13 to 12.45±1.58 on Ising, and from 31.42±13.47 to 2.45±1.01 on Potts, with comparable training times.
The variance plot compares the sampled-transport, transport-marginalized, and fiberwise targets on critical Ising and Potts. Each marginalization further reduces target variance at the evaluated bridge times. In controlled log-potential tests, we extend the alphabet-size study to N=2,4,8,16. Relative to the DASBS cyclic-shift target, the fiberwise target reduces variance by 87.98--88.95% at t=0.2 and 47.09--51.86% at t=0.8. With learned eight-state Potts correctors, the reductions at t=0.2,0.5,0.8 are 98.14±0.03%, 94.44±0.03%, and 81.26±0.13%, respectively. TAM target-construction time remains nearly constant as N increases from 2 to 16 (approximately 0.109--0.113 ms per call).
We also cite PDNS (Guo et al., 2026). Our best-performance claims refer only to uniform-based discrete neural samplers, not all classes of sampling methods.
We hope these clarifications and additional results address the concerns about novelty, practical usefulness, comparison fairness, and experimental scale, and respectfully ask the reviewer to reconsider the rating.