Modularity in deep neural networks has been proposed as a means of improving both interpretability and training by promoting disentangled representations and reducing redundancy. In this work, we study the emergence of modular structure through competition dynamics between groups of neurons during training. We introduce a method that (i) maintains near-baseline accuracy, (ii) induces usage-based modularity by sparsely routing inputs to neuron groups, and (iii) encourages specialization of these modules, such that their activations are correlated with input classes. We evaluate the proposed approach on ImageNet-100 and CIFAR-100 and show that with it, specialized modules emerge without module-level supervision. These modules capture a meaningful high-level structure in the data, with individual modules responding to semantic categories (e.g., dogs or vehicles). We also study the emergence of a hierarchical partition of sub-tasks depending on the number of modules. Our results suggest that competitive dynamics can serve as a simple mechanism for inducing functional modularity in standard architectures.
As the feature representation changes, replay must preserve the earlier classes. However, only a bounded active exemplar set can be replayed. We propose Uniform Herding, which allocates the current active set across observed classes and uses a bounded candidate pool to refresh their chosen exemplars in the current representation. On CIFAR-100 with ten class-incremental tasks, a ResNet-18 backbone, active budget $M=2{,}000$, retrieval budget $b=64$, and three seeds, Uniform Herding obtains $44.00\pm0.51\%$ final average accuracy and $17.22\pm0.43\%$ forgetting, compared with $42.33\pm1.20\%$ and $24.87\pm1.11\%$ for iCaRL. Within the Uniform Herding protocol, final accuracy decreased when NME or herding was replaced with the tested alternatives, while forgetting increased when distillation was removed. Changing the retrieval budget has a smaller effect across the tested range than changing the active budget. The comparison with iCaRL is end-to-end. It does not isolate the effect of refresh from the other protocol differences. These results are limited to the tested protocol.
Knowledge distillation (KD) trains a compact student by attracting it towards a converged teacher. It is silent about which directions the teacher itself learned to suppress: repulsive and bias-aware objectives exist, but none exploits the teacher's own trajectory to identify what the student should avoid. We observe that the missing signal is already encoded in the teacher's optimization trajectory: features that an early-stage teacher emphasizes but that a converged teacher attenuates are precisely the shortcut directions worth pushing the student away from. We instantiate this observation as \textbf{A}nti-\textbf{S}hortcut \textbf{D}istillation (ASD), a push--pull KD framework that treats the converged teacher $\Tfinal$ as a positive semantic anchor and an early-checkpoint teacher $\Tearly$ as a temporal negative reference. ASD couples two losses: a temporal contrastive loss ($\Ltc$) that places the early-teacher feature as a same-sample negative against in-batch and memory-bank final-teacher features in an InfoNCE objective; and a shortcut suppression loss ($\Lss$) that penalizes student projection onto the top eigenvectors of $\E[\Dh\Dh^{\top}]$, the uncentered second-moment matrix of early-to-final feature displacements. Across 13 teacher--student pairs on CIFAR-100, ImageNet-100, and TinyImageNet, ASD attains the highest clean top-1 accuracy on more than 10 pairs and outperforms standard KD on 12. On CIFAR-100-C corruption robustness, ASD obtains the lowest mean Corruption Error ($86.1$\,mCE) on the most challenging cross-architecture pair (WRN-40-2$\to$ShuffleNet-V2). Mechanistic diagnostics confirm the intended geometry: the ASD student is systematically anti-aligned with the shortcut direction, while its projection onto the robust subspace is substantially larger ($0.45$ vs.\ $0.12$).
Knowledge distillation trains a smaller student to match the outputs of a larger teacher. Feature-based methods also align intermediate representations, but this extra constraint may affect students differently. We study this question on CIFAR-100 using a ResNet-50 teacher, a width-controlled CustomResNet family and MobileNetV2 as a cross-design comparison. For each student, we evaluate each feature method against a matched logit-KD run using the same teacher, optimizer settings, training schedule and seed. We repeat the main comparisons across multiple seeds. Logit KD improved every tested student over its scratch baseline. Attention Transfer showed no clear relationship with size inside the CustomResNet family, but its average effect was negative for that family and positive for MobileNetV2. FitNets was below logit KD in all 15 paired runs. Within the constant-depth width sweep, its gap increased for wider students, although the different-depth w=48 student did not follow this trend. Finally, the same auxiliary coefficient produced different gradient scales across students, showing that a fixed coefficient does not create a uniform training condition.
Tolgay Atinc Uzun, Radu Timofte, Dmitry Ignatovcs.LG cs.AI
Promising initial results in closed-loop large-language-model-based channel-configuration search demonstrated that neural-network widths can be optimized directly through executable code generation and accuracy feedback. However, those results were obtained from a relatively sparse set of valid evaluations, leaving open whether the observed optimization behavior transfers to a denser sampling regime and whether additional architectural regularities emerge when more generated networks are evaluated. To test this, the same search setting is scaled to 250 candidate networks per fine-tuning cycle. The analysis covers 2000 generated candidates from 8 complete cycles, yielding 462 verified CIFAR-100 evaluations after task and metadata filtering. Per-cycle mean accuracy exhibits a positive linear trend with slope 9.87e-4 (p=0.043), while the high-performing frontier improves more strongly: the best observed accuracy increases from 0.3144 to 0.3676, and both the top-5 and top-10 cycle-level means exhibit positive trends. The scaled run also reveals improved parameter efficiency. The best model reaches 0.3676 with 11.8M parameters, compared with an early high-performing model at 0.3144 with 166.5M parameters. Beyond accuracy, the larger sample exposes architectural regularities that were difficult to assess from sparse observations. Non-power-of-two channel widths occur in 41.8% of verified candidates, and the strongest models share structured channel-allocation patterns characterized by moderate early widths and expanded middle or later blocks. These findings indicate that the channel-search signal observed in the initial study transfers
Class-incremental learning requires a model to learn new classes while preserving decision regions for old ones. This is difficult when raw old samples are no longer available. We propose Prototype Latent World Model Replay, a memory-free framework that stores old classes as distributions over stable hidden states rather than as images. A frozen ImageNet-pretrained encoder maps each image into a latent state space. In this space, each class is summarized by several prototype-centered distributions with class-specific variances. When new classes arrive, the model samples old latent states from this prototype world model. It then trains a lightweight adapter and classifier using both sampled old states and real new-class features. We also add a supervised contrastive term in the adapter space to promote intra-class compactness and old-new class separation. On Split CIFAR-100, our method improves over fine-tuning under Inc5, Inc10, and Inc20 without storing raw exemplars. The full Ours-LWM+Con model raises LastAcc from 4.55% to 31.64%, from 9.06% to 37.06%, and from 16.96% to 43.10% in Inc5, Inc10, and Inc20, respectively. It also achieves AvgAcc of 45.86%, 52.19%, and 56.18%. Ablation and retention analyses show that stable latent-state replay is the main source of the gain. Contrastive separation further refines the old-new geometry. These results suggest that prototype latent memory preserves reusable class-state distributions, rather than only fitting the current classifier.
Catastrophic forgetting, the abrupt loss of previously acquired knowledge upon learning new information, remains the central challenge in Continual Learning. This project investigates whether the order in which a model learns information affects how well it retains knowledge. Specifically, we ask: does learning general categories first (like "animals" vs "vehicles") before learning specific classes (like "dog" vs "cat") reduce forgetting compared to learning all classes at once? We test three approaches on CIFAR-100: (1) Coarse-to-Fine: train on 2 super-classes, then expand to 10 specific sub-classes, (2) Fine-to-Coarse: train on 10 sub-classes, then group into 2 super-classes, and (3) Flat: train on all 10 classes from the start. We use Elastic Weight Consolidation (EWC) to prevent forgetting during transitions. Our hypothesis is that learning general patterns first creates a stable foundation that helps the model retain knowledge when learning more detailed distinctions. We evaluate using standard metrics (accuracy, precision, recall, F1) plus continual learning metrics like backward transfer and forgetting rates. This work could inform how we design learning sequences for real-world systems that need to learn incrementally.