Spiking Convolutional Neural Networks for Text Classification
This work addresses energy efficiency and robustness in AI for natural language processing, though it is incremental as it builds on existing SNN conversion techniques.
The paper tackles the challenge of applying spiking neural networks (SNNs) to text classification by proposing a conversion and fine-tuning method with spike-based word encoding, achieving comparable accuracy to deep neural networks while significantly reducing energy consumption and improving robustness to adversarial attacks.
Spiking neural networks (SNNs) offer a promising pathway to implement deep neural networks (DNNs) in a more energy-efficient manner since their neurons are sparsely activated and inferences are event-driven. However, there have been very few works that have demonstrated the efficacy of SNNs in language tasks partially because it is non-trivial to represent words in the forms of spikes and to deal with variable-length texts by SNNs. This work presents a "conversion + fine-tuning" two-step method for training SNNs for text classification and proposes a simple but effective way to encode pre-trained word embeddings as spike trains. We show empirically that after fine-tuning with surrogate gradients, the converted SNNs achieve comparable results to their DNN counterparts with much less energy consumption across multiple datasets for both English and Chinese. We also show that such SNNs are more robust to adversarial attacks than DNNs.