Metraipolis › archive › dc_bd41371c71da

DADP: Dynamic Activity-Dependent Pruning, A Reverse Hebbian-Inspired Structural Pruning Method — arXiv:2610.11853

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.11853
License (as archived): CC0-1.0
Archived: 2026-10-09T06:23:33.853Z
SHA-256 of the archived text: 3a5748c7393b55224c57ce9d7eee1c395cffb77555bcb47e6749acebf01679ce
Stable URL of this page: https://metraipolis.com/documento/dc_bd41371c71da

How to cite

DADP: Dynamic Activity-Dependent Pruning, A Reverse Hebbian-Inspired Structural Pruning Method — arXiv:2610.11853 — arXiv (en), https://arxiv.org/abs/2610.11853 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bd41371c71da (Metraipolis archive dc_bd41371c71da, sha256 3a5748c7393b55224c57ce9d7eee1c395cffb77555bcb47e6749acebf01679ce)

BibTeX

@misc{dcbd41371c71da,
  title = {DADP: Dynamic Activity-Dependent Pruning, A Reverse Hebbian-Inspired Structural Pruning Method — arXiv:2610.11853},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.11853},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bd41371c71da; sha256 3a5748c7393b55224c57ce9d7eee1c395cffb77555bcb47e6749acebf01679ce},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11853. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Modern neural networks are heavily over-parameterized. This redundancy incurs substantial compute and memory overhead during training and inference. Existing pruning methods rely on post-hoc magnitude thresholds or static initialization heuristics. Consequently, they often require manual per-layer sparsity targets or expensive retraining cycles. We propose Dynamic Activity-Dependent Pruning (DADP), a biologically inspired structural plasticity mechanism. During training, DADP measures connection importance via the accumulated product of pre-synaptic activations and post-synaptic error gradients. Using a single global threshold instead of fixed layer budgets, DADP dynamically allocates sparsity across network depth while naturally inducing neuron- and channel-level pruning. Across MLP, VGG-16, ResNet-18, BiLSTM-CRF, and MiniBERT architectures, DADP matches or outperforms Magnitude, SNIP and RigL, retaining 73.67% accuracy (dense baseline: 76.06%) at 99% sparsity on ResNet-18. Finally, matrix-based Shannon entropy and effective rank measurements confirm that DADP preserves latent feature diversity at extreme sparsities without representation collapse.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.11853 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-09T05:50:31.686Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).