LSFormer uses local structure-aware spiking self-attention and spiking response pooling to cut global attention bottlenecks, delivering 4.3% and 8.6% accuracy gains on Tiny-ImageNet and N-CALTECH101 over prior transformer-based SNNs.
Towards artificial general intelligence with hybrid tianjic chip architecture
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3representative citing papers
A PPO policy trained in Isaac Sim with domain randomization balances and steers a bicycle in sim (99.9% success) and transfers to real hardware.
Introduces circulate-firing neurons, time-step-wise learnable surrogate gradients, and balanced loss for direct SNN training, reporting competitive results on datasets and Transformers.
citing papers explorer
-
Breaking Global Self-Attention Bottlenecks in Transformer-based Spiking Neural Networks with Local Structure-Aware Self-Attention
LSFormer uses local structure-aware spiking self-attention and spiking response pooling to cut global attention bottlenecks, delivering 4.3% and 8.6% accuracy gains on Tiny-ImageNet and N-CALTECH101 over prior transformer-based SNNs.
-
CycleRL: Sim-to-Real Deep Reinforcement Learning for Robust Autonomous Bicycle Control
A PPO policy trained in Isaac Sim with domain randomization balances and steers a bicycle in sim (99.9% success) and transfers to real hardware.
-
Advancing Direct Training for Spiking Neural Networks with Circulate-Firing Neurons and Learnable Gradients
Introduces circulate-firing neurons, time-step-wise learnable surrogate gradients, and balanced loss for direct SNN training, reporting competitive results on datasets and Transformers.