- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 1
- Soundness
- 2
Read review
Paper summary
The paper proposes TAPQ-Net for medical image classification under backdoor attacks. The model builds prototypes with EM clustering and does not use labels for this step. These prototypes give each training sample a trust weight. A tree-structured quantum encoder produces the image features. The key idea is a gate that checks the geometry of the representations and decides whether to update the prototypes. The experiments use seven datasets. However, the paper does not clearly show what the quantum encoder adds. It also does not show that the gate helps beyond simply running more EM iterations.
Critical strengths
- The design fits together well. The label-free prototypes add a signal that the classification loss does not provide. Soft weighting keeps every sample in training instead of dropping suspected ones. The paper also separates how confident a cluster assignment is from whether a sample is clean. This is a useful distinction.
- The evaluation of patch attacks is broad. It covers seven datasets and several poisoning rates. The authors also analyze trust scores, compression, and triggers.
- The paper includes relevant controls and is open about its limits. The comparison with a matched classical tree is useful. The authors state that all quantum results come from simulation and that some controls are still missing.
Critical weaknesses
- The value of the quantum module is unclear. Every defense component could run with a classical encoder. The theory gives no robustness mechanism that is specific to quantum computing. Table 3 shows that the chosen TQE beats some other designs. The matched classical tree, however, appears in only one setting, and the paper does not fully describe how it is built. The authors should compare against stronger classical encoders in more settings.
- The novelty and effect of the gate are not well supported. Clustering, prototype supervision, sample reweighting, and dynamic prototype updates all exist in prior work. The main new element is that geometry decides when to refresh. In Table 2, though, the effect of gating cannot be separated from the effect of extra EM iterations. Three controls would separate them: always refreshing, starting from fully converged EM, and matching the number of updates.
- The gate may freeze permanently. The balance and activity checks rely on EM assignments. These assignments do not change when a refresh is rejected. So once a check fails, it seems that no later update can pass. The authors should say whether this happens and how often.
- The random-class ASR results need explanation. Suppose each test target is drawn uniformly from the wrong classes. Then the expected ASR equals the triggered error rate divided by C − 1. Several values in Table 12 are well above this level. The authors should explain how targets are chosen and how the metric is computed.
- The theory does not show that predictions are robust. Proposition 4.1 simplifies the EM weights, although in practice they depend on the data. Proposition 4.3 gives only a general bound on alignment. Neither result shows that predictions stay stable or that ASR stays low. That would need extra assumptions about trigger directions, classifier norms, or margins.
- Some details needed for reproduction and evaluation are missing. The paper does not give key hyperparameter values, the exact classical tree function, or how much clean data each baseline uses. Model selection uses a clean validation set. The statistical tests run over related conditions, so their results should be read with care.
Questions for authors
- Which quantum property explains the gains? Do the gains hold against fully described, strong classical encoders in several settings?
- How does the gate compare with always refreshing, with converged EM at the start, and with schedules that match the number of updates? How many updates does the gate accept?
- Can refresh restart after a balance or activity check fails? If not, how often does the gate freeze permanently?
- How are the random-class targets generated? How do the reported ASR values follow from the stated definition?
- Can the authors give full configurations, the clean data budget of each baseline, and the ASR without any defense to confirm that the attacks work?
- What evidence links alignment to stable outputs? Useful evidence would include measured residuals, classifier norms, or how the logits change between clean and triggered inputs.
4 comments
Thank you for the added experiments and explanations. Reviewer oE5X and reviwer 12t3 also questioned the gate’s contribution, and I share this concern. Using a fixed training epoch addresses the clean-validation concern. The attack targets are now clearer, and the new adaptive attack results are useful.
My main concern still remain the gate’s contribution and the need for the quantum circuit. The gate is meant to stop potentially harmful prototype update, but it never reject an update in any primary run. How do these experiments show that it help? Since the update schedule is identical to fixed-interval refresh, why do the results differ? The full model and control use different circuit-evaluation pathways, so the current comparison do not answer this question. Please keep the implementation the same and compare results with the gate on and off. The classical model is also competitive in the new experiments. It is still unclear what extra benefit the quantum circuit provides.
Thank you for recognizing the fixed epoch-15 clarification, the target definitions and the added adaptive-attack results. We address the gate and encoder questions separately below.
Part 1/2: Gate contribution, refresh-schedule results and the requested gate-on/off comparison
We present the gate as a conditional prototype-refresh mechanism and distinguish its observed decisions from the performance of the complete training configuration. The primary runs accept every eligible check, so their gains cannot be attributed to refresh vetoes.
A. What the experiments establish about the gate
-
Observed behavior in the primary runs. All 126 full-model runs accept all 882 eligible checks, giving seven refreshes per run at epochs 2, 4, 6, 8, 10, 12 and 14. The observed premature-stop count is 0/126. These runs therefore follow the fixed-interval refresh schedule (§5.3, p. 8; Appendix D.2, p. 23).
-
Conditional behavior across encoder configurations. Under the acceptance rule in Eq. (8), p. 4, Table 19, p. 23, reports mean accepted-update counts of 7.0 on each focal dataset for TQE and exact recurrence, and 0.0 for scalar tree, MLP and random Fourier features. Other means include 6.7 for no-CNOT and flat-VQC configurations on OASIS, 6.3 for the four-qubit tree on DermaMNIST, and 6.7/6.0 for the deeper tree on PathMNIST/OASIS. These are acceptance trajectories; the counts do not identify the failed criterion or establish that a rejected update was harmful.
-
Behavior under the joint adaptive attack. The classifier-and-gate attack retains seven accepted refreshes per seed and dataset. TAPQ-Net ASRs are 0.48/0.64/1.18% on PathMNIST/DermaMNIST/OASIS, versus 99.97/99.74/100.00% undefended. The reported three-dataset means are 0.77% and 99.90%. These results measure the complete defense while the gate continues to pass (Table 6, p. 10; Appendix E.2, Eq. (23), p. 24).
The gate tests embedding silhouette and mixture occupancy before another EM pass. As specified in §3.5, p. 4, an occupancy failure is absorbing without a restart, whereas a silhouette rejection can recover as the encoder changes. We report these operational properties without treating acceptance as proof of semantic correctness.
B. Why the full-model and fixed-interval scores differ
Table 4, p. 9, documents uncached statevector evaluation for TAPQ-Net and cached circuit evaluation for the refresh controls. Appendix B.7, p. 19, records shared PathMNIST features/PCA, initial encoder/head and mixture states, poison indices and minibatch ordering. The comparison therefore matches these initial records but changes the evaluation pathway. The reported results do not isolate the numerical cause of the endpoint gap.
C. Numerical comparison and status of the requested control
The following are three-seed mean Macro-F1 / ASR (%) under A2A 10% poisoning. Seeds are 11, 22 and 33. U is the number of PathMNIST post-initialization updates. The first row uses uncached evaluation; the listed controls use cached evaluation.
| Configuration | PathMNIST | DermaMNIST | OASIS | PathMNIST U |
|---|---|---|---|---|
| TAPQ-Net | 87.13 / 0.61 | 57.78 / 0.57 | 39.35 / 0.96 | 7 |
| Fixed interval | 85.91 / 0.45 | 58.39 / 0.59 | 40.27 / 0.90 | 7 |
| Always refresh | 86.21 / 0.31 | 58.54 / 0.45 | 39.71 / 0.46 | 15 |
| Matched pass count | 86.58 / 0.18 | 60.96 / 0.54 | 38.50 / 0.61 | 7 |
| Alignment, frozen EM | 86.61 / 0.40 | 58.52 / 0.40 | 40.00 / 0.57 | 0 |
| Converged initialization | 86.66 / 0.35 | 57.45 / 0.49 | 39.81 / 0.28 | 0 |
| Shuffled monitor | 86.70 / 0.48 | 57.53 / 0.31 | 40.21 / 0.49 | 7 |
| Silhouette only | 86.46 / 0.40 | 56.02 / 0.41 | 39.29 / 0.67 | 7 |
| Balance/activity only | 85.61 / 0.41 | 60.14 / 0.44 | 39.00 / 0.56 | 7 |
Locations: Table 4, p. 9; Appendix D.2, Table 17, p. 23.
Table 18, p. 23, additionally reports PathMNIST no-balance and no-activity scores of 86.82/0.46 and 86.74/0.43, respectively, with seven accepted updates in every seed.
Part 2/2: Motivation and measured contribution of the tree encoder
We use the TQE to specify a compact hierarchy of interactions among the eight compressed image-feature coordinates. Seven learned angles couple local coordinates into progressively larger branches, with eight bounded readouts available for prototype alignment and classification. Four readouts preserve individual-coordinate cosines; two combine pairs, one combines four inputs, and the root combines all eight. This defines the feature family using a small number of learned coefficients (§3.3, pp. 3–4; Appendix A.1, p. 14).
Including the biased linear head, the encoder has gradient parameters: 88/70/43 on PathMNIST/DermaMNIST/OASIS. The scalar tree has the same counts. The compact MLP has four hidden outputs and aligns to the first four prototype coordinates; it is an alternative representation, not an exact equal-capacity substitution (Appendix B.4, pp. 17–18).
The comparisons show task-dependent utility and ASR trade-offs. Scores below are Macro-F1 / ASR (%) under A2A 10% poisoning, averaged across seeds 11, 22 and 33. Parameter counts exclude the fitted mixture and frozen backbone; U gives mean post-initialization refreshes. Triplets use PathMNIST/DermaMNIST/OASIS order.
| Encoder | PathMNIST | DermaMNIST | OASIS | Gradient parameters | U |
|---|---|---|---|---|---|
| TAPQ-Net | 87.13 / 0.61 | 57.78 / 0.57 | 39.35 / 0.96 | 88/70/43 | 7.0/7.0/7.0 |
| Exact recurrence | 86.31 / 0.57 | 59.79 / 0.48 | 40.26 / 0.98 | 88/70/43 | 7.0/7.0/7.0 |
| Scalar tree | 87.32 / 0.41 | 58.93 / 0.54 | 37.90 / 0.54 | 88/70/43 | 0.0/0.0/0.0 |
| MLP | 87.14 / 0.46 | 59.28 / 0.28 | 37.57 / 0.83 | 81/71/56 | 0.0/0.0/0.0 |
| Random Fourier | 86.77 / 0.54 | 58.65 / 0.35 | 38.40 / 0.75 | 81/63/36 | 0.0/0.0/0.0 |
| No CNOT | 85.95 / 0.43 | 58.49 / 0.28 | 39.66 / 0.84 | 88/70/43 | 7.0/7.0/6.7 |
| Flat VQC | 87.11 / 0.48 | 59.93 / 0.32 | 38.54 / 0.45 | 88/70/43 | 7.0/7.0/6.7 |
| Four-qubit tree | 86.56 / 0.34 | 59.85 / 0.46 | 39.57 / 0.51 | 48/38/23 | 7.0/6.3/7.0 |
| Deeper tree | 86.17 / 0.56 | 56.88 / 0.43 | 38.98 / 0.26 | 95/77/50 | 6.7/7.0/6.0 |
Locations: Table 5, p. 9, for scores; Appendix D.3, Table 19, p. 23, for parameter counts and mean refresh counts.
On OASIS, TQE's Macro-F1 exceeds scalar tree, MLP and Fourier features by 1.45/1.78/0.95 points, with higher ASR. Classical alternatives are competitive on PathMNIST and stronger on DermaMNIST. The deeper tree has lower Macro-F1 on all three datasets, accompanied by lower ASR. These are results for the trained configurations and their resulting refresh trajectories; acceptance counts do not establish superior semantic geometry.
Quantum execution is not required for this feature map. Equation (12), p. 5, gives its exact seven-merge classical recurrence. Appendix A.1, p. 14, verifies it against independent statevector evaluation over 100 random forward cases and 70 finite-difference gradient columns, with maximum discrepancies and , respectively. It is the same function at identical inputs and parameters; independently trained endpoint differences do not demonstrate a quantum-execution benefit.
Appendix B.4, p. 17, counts 35 multiplications and 14 additions/subtractions for the merges, excluding trigonometric evaluation and the common head. This gives a classical evaluation route without statevector simulation; it is not a zero-cost or measured deployment-speed result.
The noise analysis has a specified scope. Proposition A.3, p. 16, shows that a common positive final-readout contraction preserves cosine alignment and silhouette at fixed parameters, targets and assignments with inactive norm floors. A classifier bias can still change predictions. Appendix F, p. 25, and Figure 6, p. 27, separately evaluate simulated 256/1,024/8,192-shot and gate-level channel configurations. Those measurements do not constitute hardware validation or general noise immunity.
The evidence therefore supports evaluating this seven-angle feature family within prototype-guided training. It does not establish a requirement for quantum execution or consistent superiority over the classical alternatives. Thank you for helping us make that contribution and the gate-attribution boundary precise.
Thank you for the questions. Percentages below average seeds 11, 22 and 33. F1 denotes macro-F1.
Q1. Quantum mechanism and classical encoders. The TQE provides a fixed hierarchy of bounded trigonometric interactions controlled by seven learned angles. We test this feature-map design against fully specified classical alternatives on three datasets under A2A 10%:
The exact classical recurrence matches circuit expectations to 8.88×10−16 at identical parameters; the table compares independently trained endpoints. Classical alternatives are competitive. We specify all nine encoders' functions, dimensions, operation/parameter counts and refresh counts.
Revision: Sec. 5.4, pp. 8–9; Table 5, p. 9; Appendix A.1, p. 14; Appendix B.4, pp. 17–18; Table 19, p. 23.
Q2. Gate attribution and EM controls. All 126 primary full runs accept every eligible update (7/7). Their refresh epochs match the fixed interval, so the primary-grid gains cannot be attributed to vetoes. PathMNIST A2A 10% gives:
These compare complete configurations with shared initialization records and different circuit-evaluation pathways. Converged initialization adds 142/148/139 initial EM passes (14.2/14.8/13.9 s). The gate implements the stated geometric acceptance rule.
Revision: Sec. 5.3, pp. 7–8; Table 4, p. 9; Appendix D.2, pp. 22–23; Tables 17–19, p. 23.
Q3. Restart and permanent freezing. There is no restart after an occupancy failure, and zero premature permanent freezes occur in the 126 primary full-model runs (882/882 checks accepted). Retaining the mixture preserves hard assignments and their balance/activity statistics, so such a rejection is absorbing. Silhouette-only rejection can recover as the encoder changes. Alternative encoders' 0–7 accepted updates are reported without inferring which criterion failed.
Revision: Sec. 3.5, p. 4; Appendix D.2, pp. 22–23; Table 19, p. 23.
Q4. Targets and ASR. Training and triggered testing share cyclic g(y)=(y+1)modC for A2A or target 0 for A2O. Targets are deterministic; the independently uniform wrong-target identity does not apply. A2A uses all test images; A2O excludes true class 0.
Revision: Sec. 3.1, p. 3; Appendix B.6, p. 19; Table 9, p. 18.
Q5. Configurations, clean data and attack validity. TAPQ-Net uses no trusted clean training subset and evaluates epoch 15 without validation selection. The nominal 5% clean subsets for NP/SAU/TSBD/SCAn and every method's stages are specified; configurations were fixed, with zero post-hoc tuning runs. A2O 10% attack ASRs averaged over the three focal datasets are:
Revision: Tables 8–10, pp. 17–19; Appendix B.5, pp. 18–19; Appendix B.7, p. 19; Appendices E.1–E.2, p. 24; Tables 6 and 20, pp. 10 and 25.
Q6. Alignment and output stability. At the auxiliary epoch-14 encoder/head checkpoint, 99.05%–99.43% of paired predictions are unchanged, using frozen epoch-15 mixture targets:
Table 7 gives exact changed counts per seed. Both inputs use the same epoch-14 encoder/head; the primary endpoint remains epoch 15. Proposition A.1 (p. 15) gives prediction preservation when every clean logit margin exceeds ∥wy−wk∥2Dz; Dz uses alignment residuals, target-direction and norm changes. The prototype bound retains fitted data-dependent weights. Statistics average related conditions within seeds (Appendix H, pp. 26–27).
Revision: Sec. 4, p. 5; Appendix A.2, pp. 14–15; Appendices A.3–A.4 and Table 7, p. 15.