# NeuroFlex: Lossless Element-Level ANN-SNN Co-Execution for Efficient Sparse Inference

arXiv: https://arxiv.org/abs/2609.14092 · HTML: https://mlsanity.com/paper/2609.14092/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 27 references that are in the map.
- First posted on arXiv 12 Sep 2026.

## Builds on (most cited first, 11 of 11)

- [Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding](https://mlsanity.com/paper/1510.00149/) (2015)
- [Efficient Processing of Deep Neural Networks: A Tutorial and Survey](https://mlsanity.com/paper/1703.09039/) (2017)
- [EIE: Efficient Inference Engine on Compressed Deep Neural Network](https://mlsanity.com/paper/1602.01528/) (2016)
- [Going Deeper in Spiking Neural Networks: VGG and Residual Architectures](https://mlsanity.com/paper/1802.02627/) (2018)
- [SCNN: An Accelerator for Compressed-sparse Convolutional Neural Networks](https://mlsanity.com/paper/1708.04485/) (2017)
- [Optimal ANN-SNN Conversion for High-accuracy and Ultra-low-latency Spiking Neural Networks](https://mlsanity.com/paper/2303.04347/) (2023)
- [A Hybrid ANN-SNN Architecture for Low-Power and Low-Latency Visual Perception](https://mlsanity.com/paper/2303.14176/) (2023)
- [LoAS: Fully Temporal-Parallel Dataflow for Dual-Sparse Spiking Neural Networks](https://mlsanity.com/paper/2407.14073/) (2024)
- [Prosperity: Accelerating Spiking Neural Networks via Product Sparsity](https://mlsanity.com/paper/2503.03379/) (2025)
- [PASCAL: Precise and Efficient ANN- SNN Conversion using Spike Accumulation and Adaptive Layerwise Activation](https://mlsanity.com/paper/2505.01730/) (2025)
- [Hybrid Layer-Wise ANN-SNN With Surrogate Spike Encoding-Decoding Structure](https://mlsanity.com/paper/2509.24411/) (2025)

## Abstract (arXiv)

Sparse DNN accelerators specialize in ANN or SNN execution, leaving energy or latency on the table when workload characteristics vary within a layer. Hybrid accelerator designs that switch modes at layer or tile granularity suffer from low PE utilization since one core type idles whenever the other is active. NeuroFlex is the first accelerator to assign every output element independently to ANN or SNN execution mode with zero accuracy loss. We extend integer-exact ANN-SNN equivalence from layers to individual output elements, thereby enabling mode switching with no conversion error. An offline cost-guided scheduler scores each element by its marginal energy-delay trade-off and packs work across PEs, achieving 97-99% PE utilization compared to 40-45% for layer-wise hybrids. NeuroFlex reduces EDP by 57-67% over a strong ANN-only baseline and delivers up to 2.5x speedup over a dual-sparse SNN-only baseline. Our cost-guided scheduler improves throughput by 16-19% over random element assignment across vision, language, and transformer workloads.