Research map / Architectures and efficiency / Spiking networks
Self-Supervised Learning of Event-Based Optical Flow with Spiking Neural Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add explicit recurrent connections (FireFlowNet to FireNet) (ablation) | FireNet · MVSEC indoor_flying1 dt=1 | AEE: 1.37 -> 0.89 | Helped |
| add explicit recurrence to LIF FireFlowNet (ablation) | LIF-FireNet · MVSEC indoor_flying1 dt=1 | AEE: 1.22 -> 0.98 | Helped |
| add explicit recurrence to leaky FireFlowNet (ablation) | Leaky-FireNet · MVSEC indoor_flying1 dt=1 | AEE: 0.97 -> 0.90 | Mixed |
| use ConvGRU instead of vanilla ConvRNN | EV-FlowNet · MVSEC indoor_flying1 dt=1 | AEE: 0.62 -> 0.60 | Helped |
| add spiking LIF neurons (vs leaky ANN) | Leaky-EV-FlowNet · MVSEC indoor_flying1 dt=1 | AEE: 0.71 -> 0.71 | No clear effect |
| add postsynaptic adaptive threshold (ALIF) | LIF-EV-FlowNet · MVSEC indoor_flying1 dt=1 | AEE: 0.71 -> 1.00 | Hurt |
| add postsynaptic adaptive threshold (ALIF) | LIF-FireNet · MVSEC indoor_flying1 dt=1 | AEE: 0.98 -> 1.04 | Hurt |
| add presynaptic adaptation (PLIF) | LIF-FireNet · MVSEC indoor_flying1 dt=1 | AEE: 0.98 -> 0.90 | Helped |
| add presynaptic adaptation (PLIF) | LIF-EV-FlowNet · MVSEC indoor_flying1 dt=1 | AEE: 0.71 -> 0.75 | No clear effect |
| add presynaptic threshold adaptation (XLIF) | LIF-EV-FlowNet · MVSEC indoor_flying1 dt=1 | AEE: 0.71 -> 0.73 | No clear effect |
| narrower surrogate gradient (SuperSpike gamma 10/100 vs aTan') (ablation) | LIF-FireNet · training loss | AEE | Hurt |
| add activity regularization (5% active fraction) (ablation) | LIF-FireNet with SuperSpike · training loss | AEE | Mixed |
Closest papers by meaning
- SpikeCLR: Self-Supervised Contrastive Learning for Visual Representations with Spiking Neural Networks (ICML 2026 · similarity 0.91)
- Self-cross Feature based Spiking Neural Networks for Efficient Few-shot Learning (ICML 2025 · similarity 0.89)
- Enhanced Self-Distillation Framework for Efficient Spiking Neural Network Training (NeurIPS 2025 · similarity 0.89)
- Leveraging Asynchronous Spiking Neural Networks for Ultra Efficient Event-Based Visual Processing (AAAI 2025 · similarity 0.89)
- Spiking Neural Network as Adaptive Event Stream Slicer (NeurIPS 2024 · similarity 0.88)
- Temporal Flexibility in Spiking Neural Networks: Towards Generalization Across Time Steps and Deployment Friendliness (ICLR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.