- Rating
- 1
- Confidence
- 3
- Informativeness
- 1
- Presentation
- 1
- Soundness
- 2
Read review
Paper summary
The paper proposes GLADE, a framework for improving existing outlier detectors using auxiliary outlier category labels. It uses these labels to guide local learning on reference inliers and unlabeled target data, then combines the local scores with a global detector through adaptive shrinkage. The paper establishes finite-sample FDR control under exchangeability and symmetry conditions and gives conditional results on power gains and negative transfer. Simulations and four real data benchmarks show improvements over the corresponding global baselines.
Critical strengths
GLADE separates learning category structure from auxiliary data and learning detection scores from current target data. This is a reasonable approach when auxiliary and target anomaly distributions differ, since it allows the detector to adapt without directly transferring source trained detection scores.
Critical weaknesses
-
GLADE's specific methodological novelty is not sufficiently clear. The main difference appears to lie in the structural routing and local target learning in Section 2.2. However, the paper provides limited explanation and evidence of what distinct benefits the proposed weighted local fitting and aggregation (Eq. (2)) offer over existing ways of using auxiliary information.
-
The simplicity of individual local models does not establish the complexity or efficiency of GLADE as a whole. The paper emphasizes using simple local models to handle heterogeneous anomalies, but GLADE includes a router , local models, and repeated training and selection. Using simple logistic models alone is therefore insufficient to assess the overall method’s efficiency. An explicit complexity comparison with global models using the same information is needed.
-
The connection between the theoretical results and the actual GLADE implementation is limited. Theorem 1 establishes finite-sample FDR control, but its main argument relies on existing adaptive score conformal results (Marandon et al., 2024). It does not establish a distinct statistical advantage of GLADE’s proposed structural learning. Theorems 3 and 4 provide conditions for power gains and protection against negative transfer, but there is insufficient evidence that the implemented GLADE procedure satisfies these conditions and achieves these benefits.
-
The experiments do not identify which components of GLADE drive the performance gains. They show improvements from applying GLADE as a whole, but structural routing, local target learning, score combination, and selection are added together. Although the comparisons include baselines that already use auxiliary information, comparisons with alternative methods using the same information and component level ablations are limited. It is therefore difficult to determine whether the observed gains come from GLADE’s newly proposed components.
-
The main text does not bring together the explanations needed to understand the key contributions. It provides an overview of the method, but detailed comparisons with existing methods appear in Appendix A.3, and the full algorithm appears in Appendix A.1. The main text should more clearly distinguish existing components from new contributions and explain the full training procedure.
Questions for authors
-
Given existing methods that use auxiliary information and adaptive selection, what exactly is GLADE’s main novelty? In particular, what additional benefits do category specific weighted fitting and aggregation provide?
-
Does “simple model” refer to the form of each individual learner, or to the complexity or efficiency of the overall method? Can the authors compare GLADE with a global learner of similar overall model complexity? How do GLADE’s complexity and performance change as the number of categories increases?
-
What evidence shows that the implemented GLADE procedure satisfies the conditions in Theorems 3 and 4? Is there evidence that the minimum-gain threshold used in the implementation limits negative transfer?
-
Can the authors compare GLADE with an alternative that replaces the proposed category-specific weighted local fitting and aggregation with another learning approach, while keeping the same , , unlabeled pool, global-local shrinkage, selection, and conformal calibration? Can such comparisons identify which components of GLADE contribute to the performance gains?
6 comments
Model complexity and computational cost
Q2.
GLADE exploits the category structure of labeled auxiliary outliers to learn a flexible detector from limited current data. A global learner must discover where anomaly patterns differ and which feature combinations reveal each pattern. Learning both is often impractical when the unlabeled pool contains few examples of each type. GLADE uses auxiliary categories to guide the first task, allowing each local learner to estimate how the current pool differs from reference inliers in its region. Here, “simple” refers to the functional form of each local learner, such as a weighted logistic regression. Combining these fits yields a global score that can detect departures along region-specific directions. Without this structural guidance, fitting a single model flexible enough to capture the same heterogeneity can require substantially more current data than are available in practice.
Theorem 2 gives an exact example of this principle. In its Gaussian model with known groups, the group-specific linear scores together recover the likelihood-ratio ranking. When equally strong shifts point in different directions, this rule has strictly higher power than the best common linear direction. Under the conditions of Theorem 3, the fitted GLADE score retains a power advantage over its anchor when the population gain exceeds the errors from learning and selection.
To compare learners of similar size, we gave a single logistic regression the same features, global anchor, and routing weights, with routing-by-feature interactions and as many regression coefficients as the local models combined. We kept the same data and eight underlying anomaly types and merged auxiliary labels to obtain . We also measured full-pipeline CPU time, including routing, cross-fitting, selection, and final fitting.
Table R9. Power and cost as auxiliary labels distinguish more anomaly types. A fixed eight-type, 20-dimensional Gaussian task merges auxiliary labels to obtain . GLADE and a single interaction logistic learner use the same routing information and refinement coefficients, excluding the shared router and global anchor; they fit separate weighted local objectives and one routing-interaction objective, respectively. Part (a) reports mean powers and paired gains, with 95% Monte Carlo intervals; intervals are not truncated. Part (b) profiles the same task separately with one BLAS thread per process, covering routing, cross-fitting/selection, and final fitting. The share column is median cross-fit/selection CPU divided by median full-pipeline CPU. Part (c) reports mean FDP for the power comparison.
(a) Power (%) and paired gain [95% CI]
| Matched refinement coefficients | GLADE power | Interaction power | GLADE gain (pp) | |
|---|---|---|---|---|
| 2 | 44 | 0.89 [0.29,1.50] | 0.27 [-0.12,0.66] | +0.63 [-0.04,+1.30] |
| 4 | 88 | 9.34 [6.89,11.78] | 2.09 [0.94,3.25] | +7.24 [+5.10,+9.39] |
| 8 | 176 | 57.04 [54.56,59.52] | 33.36 [29.66,37.05] | +23.68 [+20.24,+27.12] |
(b) Median full-pipeline CPU time (seconds)
| GLADE | Interaction LR | Anchor | GLADE cross-fit/ selection share (%) | |
|---|---|---|---|---|
| 2 | 0.094 | 0.077 | 0.011 | 77.4 |
| 4 | 0.153 | 0.093 | 0.011 | 78.4 |
| 8 | 0.250 | 0.110 | 0.011 | 79.5 |
(c) Mean FDP (%) in the power comparison
| GLADE | Interaction LR | |
|---|---|---|
| 2 | 1.29 | 0.38 |
| 4 | 3.38 | 1.55 |
| 8 | 7.39 | 6.42 |
Table R9 shows how power and computation change with category resolution:
- Finer categories strengthen the power advantage. As auxiliary labels distinguish more anomaly types, GLADE gains more power over the interaction learner despite matching its refinement coefficient count. This supports using category structure through separate local fits.
- The measured cost remains modest. Full-pipeline times remain short over the category counts studied. The comparison includes the repeated fitting and selection steps, providing an assessment of the complete procedure.
The full procedure fits the router once and local models in each round of cross-fitting and in the final refit. With repetitions of -fold cross-fitting, and writing , , and for the router, anchor, and single-local-model fitting costs, the total fitting cost is approximately . All mixing weights reuse the same fitted scores.
Theoretical safeguards and empirical performance
Q3.
Thank you for asking us to connect the theoretical safeguards to their implementation.
(i) Theory and implementation. GLADE includes the exact anchor, a finite mixing grid, and the prescribed selection and tie rules. Theorem 3 shows when a structural gain survives learning and selection error. Theorem 4 gives utility protection on its uniform-accuracy event when , where and bounds candidate-evaluation error. These results remain conditional on the stated statistical assumptions, including control of the pooled-to-target discrepancy; the implementation does not certify this error bound. FDR control follows separately from Theorem 1 under its exchangeability and symmetry conditions.
(ii) Empirical support. To assess how selection balances power and protection, we compare adaptive and fixed weights across five dataset–anchor settings, holding the fitted candidates, data, and final conformal calibration fixed.
Table R10. Fixed versus adaptive mixing on the same real-data candidates. Terrain and terrain + soil are Covertype representations; CSE is CSE-CIC-IDS2018; Gas is Gas Sensor Drift. Each column fixes the dataset and global anchor. uses only the anchor, and uses only the local score. Underlined entries indicate the highest mean power among fixed-weight policies in each column, including ties at the displayed precision. Part (b) reports the percentage of all evaluated outcomes with fewer true discoveries than the matched anchor; it is distinct from FDP. Final conformal BH uses .
(a) Mean final-test power (%)
| Mixing policy | Terrain MCC | Terrain + soil MCC | UNSW-NB15 MCC | CSE MCC | Gas BIC |
|---|---|---|---|---|---|
| Fixed λ=0 (anchor) | 16.09 | 30.13 | 57.14 | 47.53 | 88.82 |
| Fixed λ=0.25 | 24.97 | 32.50 | 61.16 | 47.94 | 90.91 |
| Fixed λ=0.5 | 28.02 | 33.16 | 62.53 | ||
| Fixed λ=0.75 | 30.23 | 47.90 | 91.91 | ||
| Fixed λ=1 (local) | 60.76 | 47.06 | 86.77 | ||
| Adaptive (GLADE) | 32.59 | 33.15 | 62.49 | 47.98 | 91.69 |
(b) Harmful-outcome rate relative to the anchor (%)
| Mixing policy | Terrain MCC | Terrain + soil MCC | UNSW-NB15 MCC | CSE MCC | Gas BIC |
|---|---|---|---|---|---|
| Pure local (λ=1) | 0 | 4 | 10 | 76 | 51 |
| Adaptive (GLADE) | 0 | 1 | 0 | 23 | 4 |
Table R11. Adaptive shrinkage on Gas Sensor Drift/BIC and CSE-CIC-IDS2018/MCC. Local harms means pure-local scoring has fewer true discoveries than the matched anchor; other is its complement. Part (a) reports selected-weight frequencies within each group and the mean selected local weight. Part (b) reports mean power differences from the anchor within the harmful-local subset, with paired 95% Monte Carlo intervals. These conditional means are distinct from the overall harmful-outcome rates in Table R10(b). Groups are retrospective; test anomaly labels are used only for evaluation.
(a) Selected-weight frequencies (%) and mean
| Setting / anchor | Group | 0 | .25 | .5 | .75 | 1 | Mean λ̂ |
|---|---|---|---|---|---|---|---|
| Gas / BIC | Local harms | 27.5 | 13.7 | 45.1 | 13.7 | 0.0 | 0.363 |
| Gas / BIC | Other | 22.4 | 12.2 | 44.9 | 20.4 | 0.0 | 0.408 |
| CSE / MCC | Local harms | 25.0 | 40.8 | 14.5 | 13.2 | 6.6 | 0.339 |
| CSE / MCC | Other | 20.8 | 25.0 | 16.7 | 20.8 | 16.7 | 0.469 |
(b) Mean power difference from the anchor (pp) [95% CI]
| Setting / anchor | Pure local | Adaptive |
|---|---|---|
| Gas / BIC | -7.88 [-10.09,-5.66] | +1.58 [+0.99,+2.18] |
| CSE / MCC | -1.45 [-1.64,-1.26] | +0.19 [-0.15,+0.52] |
- Reduced negative transfer. Adaptive mixing substantially reduces harmful outcomes on Gas/BIC and CSE/MCC (Table R10(b)). Within the retrospective harmful-local subsets, smaller selected local weights accompany recovery of mean power above the anchor on Gas and approximately to its level on CSE (Table R11), supporting shrinkage as a way to limit unreliable local influence.
- Preserved detection gains. Adaptive mixing improves mean power over the anchor in every setting and stays close to the best fixed-weight mean, despite the preferred weight varying across datasets (Table R10(a)). Selection uses no test anomaly labels.
The matched routing and learning ablations in our response to Q4 assess these contributions while holding the remaining pipeline fixed (Tables R13 and R14).
Auxiliary categories and anomaly directions
We thank the reviewer for asking us to clarify GLADE's contribution and connect the theory more closely to its implementation. Our additional comparisons examine how auxiliary categories help target learning, which fitting choices matter, and what computational cost and adaptive safeguards accompany the gains.
Q1.
Our goal is to improve the power of OOD testing under FDR control by exploiting two complementary sources of structure: (a) the current test data contain information about departures from the inlier distribution in the target population, and (b) labeled auxiliary outliers reveal distinctions among anomaly types. AdaDetect (Marandon et al., Annals of Statistics, 2024) learns from the first source by comparing the current unlabeled pool with reference inliers. Integrative conformal p-values (ICP) use labeled outliers to reweight inlier-based evidence and then recalibrate the resulting score (Liang et al., JRSSB, 2024). The challenge is to bring these two sources of structure together in a single detection rule.
GLADE introduces a framework for learning a single detection score from structure in the test data and outlier category labels in auxiliary training data. GLADE brings together ICP’s strength of using labeled outliers and AdaDetect’s ability to adapt detection to the current test population. The auxiliary outlier categories indicate where to look for distinct departures. Within each such region, GLADE compares the current unlabeled pool with reference inliers to learn which features distinguish current anomalies. GLADE thus transfers what the auxiliary data reveal about anomaly types while learning from the current data how to detect those anomalies. To guard against negative transfer, GLADE uses the estimated gain in discoveries to determine how much weight to give this category-guided refinement and retains the global detector when the estimated gain is too small.
To examine when category-guided learning helps, we vary anomaly-direction separation in a fixed-norm Gaussian family, under balanced and imbalanced category proportions. GLADE and the parameter-matched interaction logistic learner receive the same routing information, without target category labels.
Table R8. Detection power as anomaly directions separate. Four Gaussian anomaly types in 20 dimensions have fixed target and auxiliary mean norms 4 and 4.6; reference, calibration, and test sizes are 2,000, 1,000, and 1,000. GLADE and interaction LR use the same auxiliary-category router; no method receives target category labels. Auxiliary size is 1,000, test anomaly prevalence is 20%, and final conformal BH uses α=0.10. Interaction LR uses routing-by-feature interactions with the same coefficient count as GLADE; both share the AdaDetect anchor. Cells in (a–c) are mean power (%) [95% Monte Carlo t CI]; bold means the largest power within each row. Part (d) reports a paired difference in percentage points. The direction sweep changes both routing distinguishability and scoring geometry.
(a) Balanced category proportions
(b) Category proportions (0.7,0.1,0.1,0.1)
(c) Minority-type power at 90∘ under imbalance
(d) Paired gain at orthogonal directions
Table R8 supports two benefits of auxiliary-guided target learning: