Papers for

privacy protection teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Membership inference fails using image frequency detail to detect memorization

Frequency-Domain Membership Inference Does Not Beat Confidence Thresholding: A Band-Resolved Analysis of Why

Abstract: A widely repeated account of membership inference holds that models lean on fine-grained detail for samples they have memorized, so degrading an input's resolution should expose membership through how quickly a prediction falls apart. We test that account band by band. A monotone family of zero-phase low-pass filters is applied to each query image, the target's confidence on its original prediction is read at every step, and the resulting decay is scored in a dozen ways. Two federated ResNet-18 targets are studied, one trained normally and one pushed to a 45.1-point generalization gap. The account does not survive. On the normally trained target the best profile score reaches an AUC of 0.513 against 0.537 for a single confidence threshold, and on the overfit target 0.735 against 0.943. A search over 1206 candidate scores, with direction chosen on a held-out half and a fitted combination as an upper bound, produces no score that beats the threshold in either regime. The failure has a mechanism: the mean confidence gap between members and non-members grows as detail is removed, exactly as the hypothesis predicts, but the spread grows faster, so discriminability falls from the first step onward. Decomposing each image into phase and magnitude shows why no band-localized operation can work, since the advantage travels with image structure, surviving at 0.567 when the phase is kept and collapsing to 0.492 when only the magnitude is. Membership is a whole-image identity effect rather than an accumulation of spectral energy in any band, and defenses aimed at high-frequency reliance are aimed at the wrong target.

Sat 17 OctArtificial Intelligence
The gist
Membership inference tries to guess whether a certain data point was used to train a machine learning model by seeing how the model’s confidence changes when parts of the input are removed. The authors tested if removing high-frequency details in images helps spot memorized samples better. They found this approach does not work better than simply looking at how confident the model is overall. The difference between members and non-members shrinks as image detail is lost, making the method less reliable.
Open → 2610.24643v1

Black-box method improves face photo privacy by smartly using small changes

SSM-Attack: Learning Where to Perturb for Black-Box Face Privacy

Abstract: Face recognition systems are deployed widely enough that a photograph shared online can be matched without its subject's consent. Generator-based protection methods rely on heavy pretrained backbones, multi-stage inference or visible semantic edits. We introduce SSM-Attack, a black-box face privacy framework built on Selective State Space Models. Rather than synthesizing the perturbation, the state-space model predicts in one linear-time pass a spatial allocation of the perturbation budget, marking which facial locations carry identity evidence; a short projected refinement then sets the values within this per-pixel budget. Against published benchmark results, SSM-Attack attains the highest average black-box attack success rate on all three evaluation sets: 87.10 percent on CelebA-HQ, 86.90 on LADN and 81.15 on FFHQ, the last from a network that never saw that dataset. The allocation network holds 5.73 million parameters, or 16.13 million with the frozen networks used by its quality terms. Under the same cap of 0.10 on the maximum per-pixel change, allocation improves structural similarity by 0.044, peak signal-to-noise ratio by 1.62 decibels and Frechet Inception Distance by 11.2 points over a uniform map while spending 37 percent of its capacity. Strong facial privacy protection therefore does not require a large generative prior; it requires knowing where to spend a small one.

Mon 12 OctComputer Vision and Pattern RecognitionCryptography and SecurityMachine Learning
The gist
Face recognition software can identify people from photos shared online, sometimes without their permission. The authors present SSM-Attack, a way to protect privacy by figuring out exactly where on a face small changes should be made to fool these systems. Instead of creating an entirely new image, their method smartly decides which facial spots to alter a little, making recognition much harder while keeping the picture looking natural. Their approach works well even on photos and systems it has never seen before.
Open → 2610.17323v1