- Rating
- 3(+1)
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper proposes AGDL for learning fMRI dictionaries across individual brain geometries. It combines FGW graph dictionary learning with a pretrained ULOT solver and introduces α-conditioned dictionary atoms. Experiments on HCP evaluate contrast classification, subject identification, computational efficiency, and transfer across graph resolutions.
Critical strengths
Learning directly on individual geometries addresses a meaningful problem in fMRI analysis. The framework can preserve anatomical information and handle graphs with different numbers of nodes. The reported speedup makes graph dictionary learning more practical. Evaluation on held-out subjects and across graph resolutions provides useful evidence of generalization.
Critical weaknesses
The added contribution over GDL and ULOT is not sufficiently isolated. Both components are existing methods, and ULOT already supports differentiable transport prediction on fMRI graphs. The α-conditioned dictionary is a useful extension, but there is no comparison with separately trained fixed-α dictionaries using the same ULOT solver. Its additional benefit therefore remains unclear. The classification results do not clearly establish the benefit of joint modeling. Subject identification may largely rely on fixed anatomical differences, while Multi-AGDL increases embedding dimensionality and Multi-AGDL+ also includes baseline features. Although the paper includes the α=1 setting, comparisons with a simple anatomical representation and dimension-matched baselines are still needed to explain the gains.
Questions for authors
-
How does the α-conditioned dictionary compare with separately trained fixed-α dictionaries using the same pretrained ULOT, in terms of accuracy and total training cost?
-
Do the improvements remain with dimension-matched baselines? For subject identification, how does AGDL compare with a simple representation based only on anatomy?
3 comments
Thanks for the response and more experiment results. The fixed-α comparison helps support the benefit of conditioning, and the dimension-matched baseline addresses my concern about embedding size. Raised my score to 3.
We thank you for your constructive feedback, which helped us better motivate the -conditioned dictionary and clarify the Multi-AGDL results, and we thank you for increasing your score.
We thank you for your insightful comments. We try our best to answer your concerns below. We applied all the reviever's suggestions in red in the updated version of the paper.
Q1. Comparison with separately trained fixed-α dictionaries. To isolate the benefit of α-conditioning, we trained separate fixed-α dictionaries for α∈{0,0.25,0.5,0.75,1}, using the same pretrained ULOT model as AGDL. We directly optimize the atoms for each value of α. We compared each dictionary with the single α-conditioned AGDL model evaluated at the corresponding α.
Accuracy. AGDL achieves higher mean contrast-classification accuracy for all tested α<1, with improvements of 8–12 percentage points. AGDL also yields higher mean subject accuracy at the other tested values. The increase of performance of AGDL might come from the fact that training across α values that is in fact a multi-task learning problem may provide a regularizing effect through shared parameters and continuous dictionary variation. The only setting where fixed-α outperforms AGDL is for contrast classification at α=1 where contrast classification fails for both fixed-α and AGDL anyways. We showed that conditioning by α preserves the strong subject-identification performance of the geometry-only setting while improving contrast classification across the other tested settings.
Training cost. In our experiments, training one fixed-α dictionary takes approximately the same time as training the α-conditioned model. Let TU denote the shared ULOT pretraining cost and TD the cost of one dictionary-learning run. Training K separate dictionaries therefore costs approximately TU+KTD, compared with TU+TD for AGDL. For the five values tested here, this corresponds to approximately for K=5 five times the dictionary-training cost, which is 20 hours for AGDL and 100 hours for fixed α dictionary learning.
Finally, once trained, AGDL can also be evaluated at additional α values without retraining the dictionary. This enables finer selection of α, which is particularly useful given the sharp transition we observe near α=1: a coarse fixed grid could miss this behavior.
Q2 Embedding dimension matching Multi-AGDL concatenates embeddings of the same graph at three values of α, producing a 360-dimensional representation. We compared it with both the original 120-dimensional baseline and a 360-dimensional baseline.
At matched dimensionality, Multi-AGDL retains substantially higher subject-identification accuracy, while its contrast-classification accuracy remains slightly lower than the baseline’s. Its subject-identification advantage over this baseline therefore cannot be explained solely by the larger embedding dimension.
About anatomy only classification, at α=1, the activation-discrepancy term vanishes from the FGW reconstruction objective, leaving only the structural term. AGDL reaches 1.00 subject-identification accuracy in this setting which is interesting because it relies on the embedding estimated from ULOT solver so recovering perfect accuracy was not a given. Any other baseline considering only the anatomy would reach perfect performance also because subjects are characterized by their brain geometry. However, the strentgh of our contribution is the interpolation and multi-view representation between geometry and activation characterization. Our result demonstrates that the learned dictionary embeddings preserve these subject-specific anatomical differences, supporting the practical effectiveness of ULOT-based dictionary learning for capturing structural information.