- Rating
- 4
- Confidence
- 3
- Informativeness
- 3
- Presentation
- 3
- Soundness
- 3
Read review
Paper summary
This paper studies continual image forensics in the open world. Generators, compression methods, and propagation pipelines keep changing, so detection cues that worked in the past can stop working. A model then has to adapt to the current environment using only a few new labels. The central question is what information past experience should retain so that future labeling needs drop. In a Gaussian model, the authors prove that representations learned from history can retain information that the optimal past decisions left out, which reduces the labels required for the current task. They further analyze when a shared subspace keeps its discriminative power as generation and processing conditions change together. Building on this idea, the paper proposes CUE. The work offers a theoretical view of the long-term value of historical information, along with a workable method. The theoretical guarantees and the experimental results for the different implementations each hold within their own scope.
Critical strengths
- The problem matters. Detection cues that worked before can fail, and the paper directly asks what a model should keep and what it should update.
- The theoretical analysis is illuminating. Under a clear Gaussian setting, the authors prove that retaining historical information can help future adaptation more than retaining only the optimal past decisions.
- The statistical version of CUE is clearly designed. The coordinate transform and the classifier come from one objective, and updates leave the visual encoder untouched.
- The experiments give concrete support. The statistical version performs strongly on six benchmarks, and matched controls show that historical statistics improve few-shot adaptation.
Critical weaknesses
- Different experiments use different implementations, so a unified test of all continual learning abilities with one statistical version of CUE is still missing.
- In the six-benchmark comparison, CUE uses current labels while the other detectors do not, so the result is not a direct advantage at an equal supervision budget.
- The theoretical conclusions rest on specific assumptions, and the history reuse experiments cover few new generators, so wider applicability still needs verification.
I support acceptance. What draws me most is the analysis of which information in history deserves to be kept, and the attempt to turn that idea into a statistical update method.
Questions for authors
- Can the main continual learning experiments be run with one and the same statistical version of CUE?
- Does the CUE advantage survive once the strongest baseline also receives 16 labels per class?
- When the historical and current distributions differ a lot, can the method lower the weight of history automatically?
- Are the gains from history reuse stable when the set of historical generators changes?
- With only one or two labels per class, is class mean estimation the main bottleneck?
- How much do a few incorrect audit labels affect the update?
- After removing a window, can you report an error comparison against rebuilding the model from the remaining data alone?
- Is the classifier selection stable for single-image or small-batch prediction?
Dear reviewers,
As the discussion period approached the end, please take a moment to read the authors’ responses and share your thoughts on any remaining points of discussion. If the discussion has changed your assessment, please update your review accordingly.
Thank you for your time and contributions!