# Otters++: A Time-to-first-spike Based Energy Efficient Optical Spiking Transformer

arXiv: https://arxiv.org/abs/2606.13016 · HTML: https://mlsanity.com/paper/2606.13016/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 20 of its 43 references that are in the map.
- First posted on arXiv 11 Jun 2026.

## Builds on (most cited first, 20 of 20)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [TinyBERT: Distilling BERT for Natural Language Understanding](https://mlsanity.com/paper/1909.10351/) (2019)
- [TernaryBERT: Distillation-aware Ultra-low Bit BERT](https://mlsanity.com/paper/2009.12812/) (2020)
- [Pushing Large Language Models to the 6G Edge: Vision, Challenges, and Opportunities](https://mlsanity.com/paper/2309.16739/) (2023)
- [How Hungry is AI? Benchmarking Energy, Water, and Carbon Footprint of LLM Inference](https://mlsanity.com/paper/2505.09598/) (2025)
- [BiT: Robustly Binarized Multi-distilled Transformer](https://mlsanity.com/paper/2205.13016/) (2022)
- [QKFormer: Hierarchical Spiking Transformer using Q-K Attention](https://mlsanity.com/paper/2403.16552/) (2024)
- [SpikingBERT: Distilling BERT to Train Spiking Language Models Using Implicit Differentiation](https://mlsanity.com/paper/2308.10873/) (2023)
- [SpikeLM: Towards General Spike-Driven Language Modeling via Elastic Bi-Spiking Mechanisms](https://mlsanity.com/paper/2406.03287/) (2024)
- [SpikeLLM: Scaling up Spiking Neural Network to Large Language Models via Saliency-based Spiking](https://mlsanity.com/paper/2407.04752/) (2024)
- [Spiking Vision Transformer with Saccadic Attention](https://mlsanity.com/paper/2502.12677/) (2025)
- [Reconsidering the Energy Efficiency of Spiking Neural Networks Inference from Analytical Perspectives](https://mlsanity.com/paper/2409.08290/) (2024)
- [Spiking Transformer:Introducing Accurate Addition-Only Spiking Self-Attention for Transformer](https://mlsanity.com/paper/2503.00226/) (2025)
- [LitE-SNN: Designing Lightweight and Efficient Spiking Neural Network through Spatial-Temporal Compressive Network Search and Joint Optimization](https://mlsanity.com/paper/2401.14652/) (2024)
- [Sorbet: A Neuromorphic Hardware-Compatible Transformer-Based Spiking Language Model](https://mlsanity.com/paper/2409.15298/) (2024)
- [Otters: An Energy-Efficient SpikingTransformer via Optical Time-to-First-Spike Encoding](https://mlsanity.com/paper/2509.18968/) (2025)
- [Human-Inspired Computing for Robust and Efficient Audio-Visual Speech Recognition](https://mlsanity.com/paper/2408.16564/) (2024)
- [Efficiently Training Time-to-First-Spike Spiking Neural Networks from Scratch](https://mlsanity.com/paper/2410.23619/) (2024)
- [A Data-Driven Dynamic Execution Orchestration Architecture](https://mlsanity.com/paper/2602.17119/) (2026)

## Abstract (arXiv)

Spiking neural networks (SNNs) are promising for energy-efficient inference, and time-to-first-spike (TTFS) coding is especially attractive because each neuron fires at most once. In practice, however, this benefit is often reduced by the cost of computing a temporal decay term and multiplying it by the synaptic weight. We address this issue by turning a physical hardware "bug," the natural signal decay in optoelectronic devices, into the main computation of TTFS, named Otters++. Specifically, we use the measured decay of a custom In2O3 optoelectronic synapse to directly realize the TTFS temporal term, removing the need for explicit digital decay computation. To scale this idea to Transformer models, we establish a layer-wise functional equivalence between the Otters++ and a quantized neural network (QNN), and develop a hybrid training method that uses device-faithful SNN computation in the forward pass and QNN straight-through gradients through the equivalent QNN path in the backward pass, together with model distillation. This avoids differentiation through discrete first-spike events and reduces the over-sparsity problem in direct TTFS-SNN training. We further make training aware of measured device noise by sampling run-to-run variation, and refine the system-level energy model by accounting for device sharing and multi-hop communication. On GLUE dataset, Otters++ improves the average score to 84.17% while maintaining a clear energy advantage over prior spiking Transformer baselines. These results show that physically grounded TTFS computing can be efficient, trainable, and robust under realistic hardware effects.