Research map / Architectures and efficiency / Spiking networks
Event2Vec: Processing neuromorphic events directly by representations in vector space
Wei Fang, Priyadarshini Panda
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| use parametric instead of standard spatial embedding (with conv temporal) (ablation) | Event2Vec · DVS Gesture | accuracy (%): 91.18 -> 97.57 | Improved |
| use parametric instead of standard spatial embedding (with sin temporal) (ablation) | Event2Vec · DVS Gesture | accuracy (%): 93.16 -> 96.56 | Improved |
| use convolutional temporal embedding on delta t instead of sinusoidal (parametric spatial) (ablation) | Event2Vec · DVS Gesture | accuracy (%): 96.56 -> 97.57 | Improved |
| use conv temporal embedding with standard spatial embedding (ablation) | Event2Vec · DVS Gesture | accuracy (%): 93.16 -> 91.18 | Hurt |
| use Batched K-Means++ clustered events instead of random events | Event2Vec · DVS-Lip | accuracy (%): 70.62 -> 75.88 | Improved |
| use batched K-Means++ instead of Faiss (iters=20) (ablation) | Event2Vec clustering · DVS-Lip | accuracy (%): 74.40 -> 75.88 | Improved |
| use batched K-Means++ instead of Scikit-learn (iters=300) (ablation) | Event2Vec clustering · DVS-Lip | accuracy (%): 75.08 -> 75.88 | Improved |
| use Forgetting Transformer instead of Gated Linear Attention (ablation) | Event2Vec · DVS-Lip | accuracy (%): 72.35 -> 75.88 | Improved |
| use Forgetting Transformer instead of Gated Linear Attention (ablation) | Event2Vec · DVS Gesture | accuracy (%): 96.67 -> 97.57 | Improved |
| share parameters in bidirectional self-attention (ablation) | Event2Vec · DVS Gesture | accuracy (%): 96.63 -> 97.57 | Improved |
| add Event2Vec parametric+conv embedding before frozen 5 Transformer layers for linear probing (ablation) | Linear probe · DVS Gesture | linear probing accuracy (%): 56.32 -> 86.94 | Improved |
| use Event2Vec parametric+conv embedding instead of raw events for linear probing (ablation) | Linear probe · ASL-DVS | linear probing accuracy (%): 12.65 -> 38.90 | Improved |
Closest papers by meaning
- DailyDVS-200: A Comprehensive Benchmark Dataset for Event-Based Action Recognition (ECCV 2024 · similarity 0.90)
- "FARSE-CNN: Fully Asynchronous, Recurrent and Sparse Event-Based CNN" (ECCV 2024 · similarity 0.89)
- EventZoom: A Progressive Approach to Event-Based Data Augmentation for Enhanced Neuromorphic Vision (AAAI 2025 · similarity 0.88)
- Leveraging Asynchronous Spiking Neural Networks for Ultra Efficient Event-Based Visual Processing (AAAI 2025 · similarity 0.88)
- E2HQV: High-Quality Video Generation from Event Camera via Theory-Inspired Model-Aided Deep Learning (AAAI 2024 · similarity 0.86)
- AE-NeRF: Augmenting Event-Based Neural Radiance Fields for Non-ideal Conditions and Larger Scenes (AAAI 2025 · similarity 0.86)
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.
