Aidin Attar PhD student · Information Engineering

Publications

Research papers and academic theses.

Conference papers

Aidin Attar, Michele Rossi

1st Neuromorphic Physical Layer Signal Processing for Wireless Systems Workshop (NeuroPHY 2026), co-located with EWSN 2026 · 2026

DOIarXiv

Publication details

Predictive Suppression Layers for Communication-Efficient Spiking Neural Networks

Aidin Attar, Michele Rossi

Abstract

We introduce predictive coding layers for spiking neural networks that trade additional local processing for reduced inter-layer communication. Predictive suppression uses prediction error to selectively suppress predictable activity, yielding communication-efficient representations while preserving, and in our experiments improving, task performance.

Bibliographic information

Type
Conference paper
Venue
1st Neuromorphic Physical Layer Signal Processing for Wireless Systems Workshop (NeuroPHY 2026), co-located with EWSN 2026
Proceedings
Proceedings of EWSN 2026
Year
2026
Pages
309–314
DOI
10.3217/za2v-bn54
DOIarXiv

BibTeX

@inproceedings{attar2026predictive,
  title = {Predictive Suppression Layers for Communication-Efficient Spiking Neural Networks},
  author = {Attar, Aidin and Rossi, Michele},
  year = {2026},
  booktitle = {Proceedings of EWSN 2026},
  pages = {309--314},
  doi = {10.3217/za2v-bn54}
}

Preprints

Aidin Attar, Eleonora Cicciarella, Michele Rossi

Submitted to Elsevier Neurocomputing · 2026

arXivPDF

Publication details

Multi-Depth Temporal Fusion for Feedforward, Locally Trained Spiking Neural Networks

Aidin Attar, Eleonora Cicciarella, Michele Rossi

Abstract

We propose a new spiking neural network (SNN) design to process static images and event streams using time-to-first-spike (TTFS) latencies. Our key research question is which architectural choices best accommodate local and online learning in multi-layer convolutional SNNs. This question is addressed via an original framework combining residual-like connections with multi-depth feature aggregation and consensus. The full SNN pipeline features an early-vision front end, to convert raw visual data into sparse spike latencies, a four-layer convolutional backbone trained layerwise with unsupervised spike-timing-dependent plasticity (STDP), a deterministic Multi-Depth Temporal Fusion (MDTF) and a final classifier trained with reward-modulated spike-timing-dependent plasticity (R-STDP). Rather than replacing early features in deeper layers, the proposed MDTF preserves early temporal evidence, adding sparse residual events from intermediate layers, and incorporating deeper features only when they agree in time with earlier representations. The resulting architecture is experimentally validated across MNIST, Fashion-MNIST, CIFAR-10, and N-MNIST, delivering strong classification performance under a fully local learning regime. Selective multi-depth fusion significantly outperforms traditional STDP/R-STDP baselines on higher-variability visual tasks (achieving +18.2 pp on Fashion-MNIST and +29.2 pp on CIFAR-10). Furthermore, activity-budget analyses show that the network retains high accuracy even when removing a large fraction of late or weak spike events, confirming its high data efficiency and reduced event-processing requirements. The codebase is publicly available at https://github.com/aidinattar/multi-depth-temporal-fusion-snn.

Bibliographic information

Type
Preprint
Venue
Submitted to Elsevier Neurocomputing
Year
2026
Pages
23
arXivPDF

BibTeX

@misc{attar2026multidepthtemporalfusionfeedforward,
      title={Multi-Depth Temporal Fusion for Feedforward, Locally Trained Spiking Neural Networks}, 
      author={Aidin Attar and Eleonora Cicciarella and Michele Rossi},
      year={2026},
      eprint={2609.37047},
      archivePrefix={arXiv},
      primaryClass={cs.NE},
      url={https://arxiv.org/abs/2609.37047}, 
}

Theses

Master's thesis · University of Padova · 2024

Exploration of Deep Spiking Neural Network Architectures for Reward-Modulated STDP Learning

Master’s thesis on deep spiking neural networks, local plasticity, reward-modulated STDP, and architectural strategies for extending biologically inspired learning beyond shallow settings.

Read thesis

Bachelor's thesis · University of Padova

Search of gravitational waves signal due to post merger phase of neutron star binary merger

Bachelor’s thesis in Physics focused on the detection and reconstruction of post-merger gravitational-wave signals from binary neutron star systems. The work centers on the Coherent WaveBurst algorithm and on simulated injections in the LIGO-Virgo network, with attention to how well post-coalescence dynamics can be recovered under O4 sensitivity assumptions. The manuscript is written in Italian.

Read thesis