- Rating
- 3
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
GMD learns user-specific Gaussian mixture priors via a common codebook in the framework of cross-domain sequential recommendation. It employs domain-specific log density difference-based denoising. Three pairs of Amazon domains are utilized in experiments on users with scarce or zero interactions in the target domain.
Critical strengths
Personalized initialization is a useful direction for cross-domain adaptation, and the prior comparisons, ablations, and runtime measurements provide evidence of gains in both accuracy and efficiency.
Critical weaknesses
-
The stated mixture prior does not correspond with the described sampling method. Equation (4) describes a Gaussian mixture, but Equations (8)–(9) weight each component's mean and variance by the weight of the component and concatenate parallel trajectories from Gaussian distributions. During inference, the output of each component is averaged. These processes are typically different from sampling the mixture in Equation (4). The authors need to state which distribution the sampling method employs, since the use of a mixture prior is key to the paper.
-
The reduction in denoising steps requires better proof. The number of steps is compared in Figure 3, but the noise schedule and final noise level are not described well enough to confirm that the comparison is apples-to-apples. Comparison to faster standard-prior samplers at similar accuracy would also demonstrate whether the learned prior provides an advantage over other sampling methods.
Questions for authors
Would you mind explaining how the weight-scaled Gaussian trajectory Eqs. (8)–(9) relate to the mixture prior of Eq. (4)? Also, does Fig. 3 use consistent noise schedules and terminal noise for both priors?
4 comments
Response to Reviewer 8Ffz(2)
W1 & Q The stated mixture prior does not correspond with the described sampling method. Equation (4) describes a Gaussian mixture, but Equations (8)–(9) weight each component's mean and variance by the weight of the component and concatenate parallel trajectories from Gaussian distributions. During inference, the output of each component is averaged. These processes are typically different from sampling the mixture in Equation (4). The authors need to state which distribution the sampling method employs, since the use of a mixture prior is key to the paper. Would you mind explaining how the weight-scaled Gaussian trajectory Eqs. (8)–(9) relate to the mixture prior of Eq. (4)? Also, does Fig. 3 use consistent noise schedules and terminal noise for both priors?
Response (2/2):
To make this clear, we have rewritten Eq. (8) as
where both operations refer to the same component selected by GS according to , so the mean and variance always come from one sampled sub-distribution. Its expanded form (Eq. (9)) is
where is the Gaussian noise.
For the denoising sampling process (Eq. (17)), the starting point is drawn from the Gaussian mixture given in Appendix Eq. (42):
This expression likewise denotes GS sampling of the distribution of . Both the noising and the denoising processes therefore use GS sampling of the Gaussian mixture.
The same applies to the initialization in Algorithm 2. Its original form,
is likewise intended to denote sampling from the Gaussian mixture rather than scaling all components by the mixture weights. We have corrected it to
The core properties of the diffusion process are fully preserved under the revised formulation. The timestep is uniformly sampled from . When , and thus . When , exactly (see W2 & Q), and thus
which is consistent with both the starting point of the denoising process in Eq. (17) and the corrected initialization in Algorithm 2.
Finally, all formulas involving the Gaussian mixture (e.g., Eqs. (4), (8)--(9) and Algorithm 2) follow the sampling procedure given in Eq. (68) of Appendix B.5. We have clarified this in the revised paper.
On Fig. 3. Yes, both priors use exactly the same noise schedule and terminal noise level () at every ; details are given in our response to W2 & Q.
Revision. Sec. 4.2.1--4.2.2 (Eqs. (8)--(9)), Appendix A (Alg. 1--2), Appendix B and B.5 (Table 4), in blue in the revised PDF.
Response to Reviewer 8Ffz(3)
W2 The reduction in denoising steps requires better proof. The number of steps is compared in Figure 3, but the noise schedule and final noise level are not described well enough to confirm that the comparison is apples-to-apples. Comparison to faster standard-prior samplers at similar accuracy would also demonstrate whether the learned prior provides an advantage over other sampling methods.
Response:
We thank the reviewer for this constructive comment. Yes, Figure 3 uses matched noise schedules and terminal noise levels for both priors.
Noise schedule. We use a linear schedule that increases from 0 to 1 over steps:
The interval between consecutive steps depends on the total number of steps : the smaller is, the more noise is added at each step.
Matched terminal noise. Since for every , holds exactly for all step budgets. In Figure 3, the number of diffusion steps is a hyperparameter of each setting: for each value of , a separate model is trained with only changed and all other hyperparameters kept fixed. DP and SP use exactly the same schedule at every , and differ only in the prior (Sec. 5.4, lines 456--458). Therefore, the gap between the two curves comes from the learned prior itself.
Comparison with accelerated samplers. The purpose of Figure 3 is to verify the effect of the prior itself, i.e., that a learned prior alone already accelerates denoising. Accelerated samplers are complementary to the prior: GMD can also adopt DDIM for further acceleration. For a fair comparison, we therefore equip both priors with DDIM (KF-FR, NDCG@10):
| 5 | 10 | 15 | 20 | 30 | 50 | |
|---|---|---|---|---|---|---|
| Mixture prior (DP) | 0.172 | 0.190 | 0.194 | 0.205 | 0.201 | 0.207 |
| Mixture prior + DDIM | 0.186 | 0.204 | 0.205 | 0.206 | 0.206 | 0.207 |
| Standard prior (SP) | 0.079 | 0.089 | 0.096 | 0.103 | 0.120 | 0.145 |
| Standard prior + DDIM | 0.105 | 0.116 | 0.126 | 0.133 | 0.141 | 0.143 |
DDIM improves both priors at small step counts, and the mixture prior with DDIM performs best at every step budget, reaching 0.204 with only 10 steps. Under the same sampler, the mixture prior consistently outperforms the standard prior by a large margin: even with DDIM and 50 steps, the standard prior (0.143) remains below the mixture prior with only 5 steps (0.172). This shows that the advantage comes from the prior itself rather than from the sampling procedure, and that the two are complementary: combining the mixture prior with DDIM further accelerates GMD. We have added these details to the revised paper.
Revision. New Appendix C.3 (Table 7), Sec. 5.4 and new Appendix C.8 (Table 13, DDIM), in blue in the revised PDF.
Response to Reviewer 8Ffz(4)
Summary of revisions for Reviewer 8Ffz (changes are marked in blue in the revised PDF)
| Issue | Location | Revision |
|---|---|---|
| W1 & Q | Sec. 4.2.1–4.2.2, App. A, B, B.5 | Eqs. (8)–(9) rewritten as sampling from the mixture in Eq. (4); "aggregation" revised; Table 4 |
| W2 | Sec. 5.4, App. C.3, C.8 | Table 7 (matched schedule and terminal noise); Table 13 (DDIM for both priors) |
Response to Reviewer 8Ffz(1)
Response (1/2):
We thank the reviewer for this careful reading. The process uses the Gaussian mixture in Eq. (4). The parallel trajectories in Eqs. (8)--(9) carry the parameters of each mixture component, i.e., the mixing weights π1:NX, means μ1:NGX and standard deviations σ1:NGX of all N components, which preserves the mixture structure (Appendix B.5, line 939). Gumbel-Softmax (GS) then selects one component according to the mixing weights.
On the per-component expressions in Appendix B.5. The per-component expressions in Appendix B.5 (lines 939--943), N(πnμn,(πnσn)2I), are another way of writing the Gaussian mixture. By the linear transformation property of the Gaussian distribution, if xn∼N(μn,σn2I), then πnxn∼N(πnμn,(πnσn)2I); this expression therefore represents each component together with its mixing weight in a parallel form. When the weights are realized by the Gumbel-Softmax selection wn, which assigns 1 to the selected component and 0 to the others, we have ∑nwnxn=xn∗ with xn∗∼N(μn∗,σn∗2I), i.e., exactly a sample from the Gaussian mixture in Eq. (68). We have revised these expressions together with Eq. (8) and Algorithm 2 to make this correspondence explicit.
On the term "aggregation" in Appendix B.5 and C.1. We apologize that the word "aggregation" in Appendix B.5 (lines 944--950) and C.1 (lines 1046--1048) is imprecise and may have caused confusion. It refers to the Gumbel-Softmax combination in Eq. (12) (Eq. (72) in Appendix B.5), whose weights π^n are relaxed one-hot sampling weights rather than fixed averaging coefficients. It is therefore a differentiable relaxation of selecting one component, not a fixed average of the components, and we have revised the wording accordingly.
Setting gn=gm=0 is applied only at inference. During training, Gumbel noise gn∼Gumbel(0,1) is always drawn, so the model is learned by sampling components according to the mixing weights π1:NX, exactly as in Eq. (68). At inference, the Gumbel noise is set to zero solely to avoid randomness in the test results and make them reproducible. This does not change what the model has learned; the core of GMD remains sampling from the Gaussian mixture.
To verify that this inference-only choice does not affect the results, we compare it with keeping the default Gumbel noise at inference, i.e., stochastic sampling, averaged over five runs (NDCG@10):
The two settings yield nearly identical results: the differences (0.0008--0.0011) are all within one standard deviation of stochastic sampling. This confirms that setting gn=gm=0 at inference only removes evaluation randomness and does not change the behavior of the model.
Specifically, the model prediction based on the forward (noising) process in Eq. (12),
x^pre=∑m=1Nexp(logπmX+gm)∑n=1N[exp(logπnX+gn)⋅x^0,n],gn,gm∼Gumbel(0,1),
implements the sampling process of Eqs. (8)--(9). The formulation of Eq. (8) in the paper,
q(xt∣xt−1):=N(αtxt−1+(1−αt)π1:NX⋅μ1:NGX, βt(π1:NX⋅σ1:NGX)2I),
may have been misleading. Here π1:NX⋅μ1:NGX and π1:NX⋅σ1:NGX only denote a parallel structure that represents the Gaussian mixture; they are not computed as weighted sums. Only the single component obtained by sampling participates in the computation.