Fig 1.
DGASA architecture: lightweight subspace construction and drift compensation.
Fig 2.
Inference pipeline of the proposed DGASA framework.
GA modules are inserted in each Transformer block for task-specific adaptation.
Fig 3.
Architecture of the Gated Adapter (GA) module.
The gating mechanism uses global pooling followed by a linear layer with sigmoid activation to produce a scalar gate g(x), which modulates the adapter residual branch.
Table 1.
Comparison of different methods on five benchmark datasets. denotes average accuracy over incremental stages, and
is the accuracy after the last stage. The best performance is shown in bold. All methods are implemented without using exemplars.
Table 2.
Comparison of different methods with exemplar usage and accuracies on benchmark datasets.
Table 3.
Ablation Study Results on CIFAR-100 Inc5 Task.
Table 4.
Ablation study on the number and positions of Attention-Gated Adapters.
Table 5.
Ablation study on the weighting coefficient .
Fig 4.
t-SNE visualization comparison between (a) Adaptive Weighting by Instance-level Significance and (b) Fixed Weight Fusion by Subspace Prototypes.
Table 6.
Parameter growth analysis across incremental tasks.
Table 7.
Prototype memory cost across incremental tasks (CIFAR-100, 100 total classes).
Table 8.
Inference efficiency analysis across tasks (batch size = 64).
Table 9.
Break-even analysis: efficiency vs. accuracy trade-off. All results are reported on CIFAR-100 Inc5.
Table 10.
Forgetting and backward transfer analysis across benchmark datasets. Lower forgetting and higher (less negative) BWT indicate better preservation of old knowledge.
Fig 5.
Old vs. new class accuracy across incremental sessions on CIFAR-100 Inc5.
Table 11.
Statistical significance analysis (paired t-test, p-values). Values < 0.05 indicate significant improvement.
Fig 6.
Per-task accuracy matrix on CIFAR-100 Inc5.
Diagonal entries show performance on the current task; off-diagonal entries show retention of previous tasks.
Table 12.
Old vs. new class accuracy on CUB Inc10 (final stage).
Fig 7.
Confusion matrix comparison on CUB Inc10: (a) ADAM+Adapter, (b) DGASA.
Fig 8.
Distribution of adaptive weights: (a) samples from old classes, (b) samples from new classes.
Table 13.
Representative failure cases on CUB Inc10.
Table 14.
Calibration analysis on CUB Inc10.
Table 15.
Conditional analysis of adaptive weighting effectiveness.
Table 16.
Computational overhead of adaptive weighting (20 tasks).
Table 17.
Sensitivity analysis of the regularization parameter on CIFAR-100 Inc5 (20 tasks).
Table 18.
Impact of increment size and conditioning on drift compensation performance. Experiments conducted on CIFAR-100 with varying per-task class increments.
Table 19.
Stability analysis across long task sequences on CIFAR-100 (Inc5, 20 tasks total).