- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 1
Read review
Paper summary
This paper proposes the existence-field diffusion model (EFDM), which associates each potential point with a continuous variable representing its degree of presence. EFDM uses Gaussian diffusion to jointly model point locations and cardinality through a simple continuous representation. Existence probabilities can increase or decrease in both forward and reverse diffusion, without explicit adding or deleting steps. The experiments demonstrate improved performance.
Critical strengths
- The proposed methodunifies coordinates, existence, and category attributes. By jointly diffusing existence variables and coordinates, and extending this to category attributes through shared existence probabilities, the model can simultaneously generate geometry, cardinality, and atomic types.
- The experiments are quite extensive, and the evaluation metrics are comprehensive.
Critical weaknesses
- EFDM applies Gaussian diffusion and MSE-based denoising to the existence logits, yet it ultimately derives discrete existence values via a Bernoulli distribution or an argmax operation. The paper does not provide a variational lower bound (ELBO) or establish a rigorous connection to the diffusion of categorical existence variables. The design—weighting coordinate losses by and using standard MSE for the existence loss—appears heuristic. It remains to be clarified whether this continuous relaxation effectively optimizes the likelihood of a valid discrete set of points, and whether the final inferred distribution aligns with the training objective.
- Although EFDM allows for continuous variation in cardinality during the diffusion process, the actual modeling space is a bounded cardinality space defined by . In experiments, is set to a value close to the maximum cardinality observed during training. This implies that the model cannot naturally extrapolate to cardinalities exceeding , and that both computational and memory requirements scale with .
- The proposed method is currently limited to spatial point processes. Whether it has the potential to be naturally extended to spatio temporal point processes, which is a more general and common scenario, remains to be discussed.
Questions for authors
- In Table.3, EFDM underperforms FDDM on the Count metric , but performs better on /. How do the authors explain this trade-off? Does this suggest that while the model improves the distribution within the conditional space, it does so at the expense of the marginal count distribution?
- How should cardinality be chosen? If the actual cardinality during testing exceeds the maximum used during training, how will the model fail?
- Is there a theoretical connection between Gaussian diffusion in logit space and the final Bernoulli/argmax readout?
- Can the model be naturally extended to spatio temporal point processes?
- In a multi marked point process, is the assumption of a shared existence probability too restrictive? How can EFDM be extended to accommodate scenarios where different attributes exhibit distinct missingness patterns? Furthermore, how significant is the discrepancy between the deterministic argmax readout and probabilistic sampling-based readout when applied to the QM9 dataset?
3 comments
Please respond
Hi iVCt,
Thanks for your work so far on this paper. Please respond to the authors before the end of the discussion period, even if to acknowledge their response.
I appreciate the authors for the effort and clarification. After carefully reading the author's response and other reviewers' comments, I decide to keep the rating.
Response to Reviewer iVCt
Question 1. Is there a trade-off between marginal and conditional accuracy?
FDDM/IDM's lower Count W1 is consistent with its explicit modeling of the cardinality marginal. Its dedicated count model gives it direct control over this marginal but does not by itself guarantee better conditional spatial fidelity. The differing metric rankings do not establish an inherent trade-off between marginal and conditional accuracy. Count W1, ICMMD$^2$, and JMMD$^2$ can all improve simultaneously.
Question 2. How should N be chosen, and what if a test configuration exceeds it?
With N slots, EFDM cannot generate more than N points. However, when test configurations contain far more points than any training configuration, every method must extrapolate beyond the observed cardinality range. No method can guarantee accurate generation in that regime from the training data alone. Increasing N raises this limit but does not by itself resolve the extrapolation problem. For our experiments, it suffices to choose N to cover the largest training configuration, with room for a modest margin when the computational budget permits.
Question 3. How does Gaussian diffusion relate to the final readout?
Yes. Both readouts follow from how we encode the training data for Gaussian diffusion. For binary presence, we encode each slot as a logit ui and keep it with probability σ(ui). Even with perfect denoising, this Bernoulli draw flips the clean label with probability δ.
On QM9, argmax recovers the correct atom type or absence from every clean ±1 code (Appendix A.3). The same encoding also supports sampling: since type k is encoded as +1 when present and −1 otherwise, its probability is qk=(u^k+1)/2 when u^k is the conditional mean of that code. The sum over atom types is the probability of keeping the slot.
Question 4. Can EFDM model spatiotemporal point processes?
Yes. In additional experiments on the Earthquakes dataset, we added time as an extra coordinate and trained EFDM with three random seeds, following the PSD setup. EFDM achieved lower mean SL and MMD than the published PSD results.
Question 5. Can attributes be missing individually, and how does sampling compare with argmax on QM9?
Mark-specific missingness is an interesting question that we have not tested. For conditional prediction, we could model attribute values given that the point exists and use a separate indicator for whether each attribute is observed. This would let us predict a missing attribute from the observed ones.
On the same 10,000 saved QM9 terminal states, sampling from the probabilities derived from the signed codes came within 0.37 percentage points of argmax on all three metrics.
Weakness 1. What does Eq. (4) optimize, and how does it relate to the generated point set?
Eq. (4) trains EFDM to denoise coordinates and existence jointly. For a given noisy input and time, the coordinate weight is a positive constant, so the optimal noise prediction under weighted MSE is still the conditional mean of the noise. The final readout defines a distribution over finite point sets. At the clean training codes, Bernoulli sampling recovers each binary label with probability 1−δ, and argmax recovers each QM9 category exactly. We do not claim that Eq. (4) is an ELBO for the decoded point sets.
Weakness 2. See Question 2 for how we choose N. Larger slot budgets increase computation and memory use.
Weakness 3. See Question 4 for our spatiotemporal Earthquakes experiment and comparison with PSD.