- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper introduces READit, a self-supervised representation learning model for irregularly sampled, multiband astronomical light curves. It applies DINO-style self-distillation to a compact Transformer using multi-scale temporal views and observation-specific data augmentations. Experiments on ZTF light curves report improved classification performance and faster pre-training convergence over existing methods, while also evaluating label efficiency and transfer to unseen classes and simulated LSST observations.
Critical strengths
-
The paper applies DINO-style self-distillation to irregular multiband light curves, improving classification performance over the contrastive learning baseline Astra-CLR. It also achieves better performance than supervised training from scratch in low-label regimes, demonstrating the potential of self-distillation for representation learning in this domain.
-
The evaluation extends beyond in-distribution classification to unseen astronomical classes and simulated LSST observations (PLAsTiCC). The few-shot results suggest that representations learned from ZTF observations can transfer to new astronomical classes and observational settings.
-
The paper provides ablations of observation-specific augmentations and multi-crop configurations. The results demonstrate the benefits of temporal masking and photometric perturbations, as well as combining global and local views rather than using only global views (Table 1).
Critical weaknesses
-
The comparison with existing contrastive learning approaches is insufficient. READit combines existing light-curve preprocessing and multi-crop strategies with DINO-style self-distillation. Although several design choices are examined through ablations, no controlled comparison of learning methods using the same data and backbone is provided, making it difficult to isolate the contribution of self-distillation. Additionally, the cross-survey transfer experiments compare only against random initialization, leaving the advantage over existing pre-trained methods unclear.
-
The claimed computational savings are not supported by a direct comparison of computational costs. The reported 18× speedup and approximately 95% compute reduction are based on comparing 200 epochs for Astra-CLR with 11 epochs for READit (Figure 4a). However, the methods differ in their training configurations and computational cost per epoch, making it difficult to establish the claimed reduction without a direct comparison in GPU-hours or FLOPs.
-
The construction of the evaluation dataset is unclear. Appendix A.1 states that 120K-Eval contains 10,000 light curves for each of 12 classes. However, according to Figure 5, some rare classes contain fewer than 10,000 samples even in the full dataset. For example, CEP accounts for only 0.059% of 2.6M-Full, corresponding to approximately 1,530 sources. It is therefore unclear how the balanced 120K-Eval dataset was constructed.
Questions for authors
-
How do contrastive learning and DINO-style self-distillation compare in classification performance and convergence speed when using the same data and backbone, with configurations appropriate for each learning method?
-
What is the actual computational speedup over Astra-CLR when measured in GPU-hours or FLOPs rather than pre-training epochs?
-
How were 10,000 samples per class obtained for the rare classes in 120K-Eval? If oversampling was used, were training and evaluation sets separated at the unique-source level, and how does performance compare when evaluated on unique sources without duplication?
5 comments
Response to Reviewer f8NB
Thank you again for your thorough review. To ensure our revisions are directly accessible within this thread, we have included our General Response below and updated our revised PDF (with all modified passages highlighted in blue).
In addition to addressing your specific questions regarding the controlled objective comparison (Appendix F, Table 8), pre-trained PLAsTiCC baselines (Section 5.2, Table 4), and strict unique-source segregation in 120K-Eval (Appendix A.1), we have also added a comprehensive benchmark on the human-vetted StarEmbed (ZTF-40K) catalog (Appendix E) along with direct comparisons against all baseline models on this catalog (Appendix G, Table 9), where READit demonstrates strong, zero-leakage performance.
General Response: Summary of Revisions & New Controlled Experiments
We thank the reviewers for their constructive feedback. We have revised our manuscript (modifications highlighted in blue in the updated PDF). Below is an executive summary of our new empirical additions and clarifications:
1. Technical Novelty & Controlled Compute Speedup (Appendix F, Table 8):
To isolate the self-distillation objective from architectural choices, we executed a strictly controlled ablation on identical hardware ( NVIDIA T4 GPUs) on 26K-Proxy, fixing the Tiny Transformer backbone (), 2 global views, and Cosine scheduler:
- Wall-Clock Speedup: Self-distillation reaches its loss basin in GPU-hours (Epoch 4) vs. GPU-hours (Epoch 123) for contrastive
NT-Xent—an empirical compute reduction. - Time-to-Parity: Contrastive learning requires more compute (h) merely to match the Macro-F1 that READit achieves in minutes, requiring additional GPU-hours for a marginal gain.
- Overcoming 2-View Ceilings via Temporal Anchoring: While 2-view distillation converges rapidly, it imposes an information bottleneck. READit introduces Temporal Anchoring (3 deterministic global views spanning start, mid, end + 6 multi-scale local views), raising downstream performance to Macro-F1 in just GPU-hours (Epoch 8, Figure 11).
- Linear Memory vs. OOM: Crucially, scaling to 9 views is computationally feasible only under self-distillation due to linear memory scaling . An equivalent 9-view contrastive setup triggers immediate Out-of-Memory (OOM) failures due to quadratic negative-pair matrices.
2. Expert-Vetted StarEmbed (ZTF-40K) Benchmark (Appendix E): We evaluated READit on the human-vetted StarEmbed benchmark. Using a 3D spherical -d tree matching radius (), we isolated a strictly non-overlapping, zero-leakage test set (). READit achieves virtually identical Linear Probing Macro-F1 on this clean split () compared to the full test set (), ruling out pre-training memorization, and demonstrates competitive performance against all baseline models on this catalog (Appendix G, Table 9).
3. Cross-Survey Transfer vs. Pre-Trained Baselines (Section 5.2, Table 4): On simulated LSST data (PLAsTiCC), READit strictly outperforms pre-trained Astra-CLR weights across all regimes:
- Tier 1 (Domain Shift, 3-Way): vs. Recall@1; vs. 5-shot Macro-F1.
- Tier 2 (Comprehensive Transfer with Supernovae, 5-Way): vs. Recall@1; vs. 5-shot Macro-F1.
4. Dataset Construction & Zero-Leakage Clarification (Appendix A.1):
In 120K-Eval, unique physical sources were strictly partitioned into disjoint training (80%) and validation (20%) splits prior to independent oversampling, guaranteeing zero source-level leakage.
5. Downstream Classifier Configurations (Appendix C):
We detailed classifier hyperparameters (Linear Probe: lbfgs, , max_iter=4000; -NN: ; RF: 200 trees, depth 20) and ASTROMER2/CHRONOS2 embedding extraction protocols.
Thank you for the detailed responses and additional experiments. The new results and clarifications have largely addressed my concerns regarding Q1, Q2, and Q3.
The controlled comparison (Table 8) shows that DINO-style self-distillation converges much faster, and the ablations (Table 1) show that multi-crop views improve representation quality. However, both are established techniques.
The additional gains from Temporal Anchoring and the specific 9-view configuration appear relatively modest: Temporal Anchoring vs. stochastic global views (0.383 vs. 0.378) and 9 vs. 6 views (0.383 vs. 0.381). While the manuscript motivates these design choices, the empirical analysis does not clearly establish how much they contribute to the improvements over prior methods.
I therefore remain unconvinced of the methodological contribution and will maintain my original rating.
Response to Reviewer f8NB
We thank Reviewer f8NB for their rigorous evaluation of our methodology, computational metrics, and dataset construction.
Regarding Pre-Trained Baselines on PLAsTiCC: We have updated Table 4 (Section 5.2) to directly benchmark against pre-trained Astra-CLR weights on PLAsTiCC. READit strictly outperforms pre-trained Astra-CLR across both Domain Adaptation (Tier 1: 60.16% vs. 53.59% 5-shot F1) and Comprehensive Transfer (Tier 2: 48.73% vs. 47.14% 5-shot F1). Please see Point 3 in our General Response.
Q1 & Q2: Controlled Comparison & Actual Hardware Speedup: To isolate the learning objective on identical hardware (2× NVIDIA T4 GPUs), we conducted a controlled ablation on the
26K-Proxydataset using the exact same Tiny backbone (L=4,d=256) and 2 global views (Table 8, Appendix F). Self-distillation reaches its loss basin in 0.38 GPU-hours, whereas the contrastive baseline (NT-Xent) requires 7.65 GPU-hours—an empirical 19.9× wall-clock speedup (95.0% compute reduction). Furthermore, contrastive learning requires 4–5× more compute simply to match READit's early performance. Please see Point 1 in our General Response for full details.Q3: Dataset Construction & Zero-Leakage Clarification: We guarantee zero data leakage at the unique-source level. Unique physical sources were strictly partitioned into disjoint training (80%) and validation (20%) splits prior to independent minority-class oversampling. Validation metrics are evaluated strictly on physical sources isolated from training. We have clarified this protocol in Appendix A.1 (see Point 4 in our General Response).