The Cascade Transformer: an Application for Efficient Answer Sentence Selection
By Luca Soldaini, Alessandro Moschitti. 2020
48 citations recorded by the source corpus.
Large transformer-based language models have been shown to be very effective in many classification tasks. However, their computational complexity prevents their use in applications requiring the classification of a large set of candidates. While previous works have investigated approaches to reduce model size, relatively little attention has been paid to techniques to improve batch throughput during inference. In this paper, we introduce the Cascade Transformer, a simple yet effective technique to adapt transformer-based models into a cascade of rankers. Each ranker is used to prune a subset of candidates in a batch, thus dramatically increasing throughput at inference time. Partial encodings from the transformer model are shared among rerankers, providing further speed-up. When compared to a state-of-the-art transformer model, our approach reduces computation by 37% with almost no impact on accuracy, as measured on two English Question Answering datasets.
Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Artificial Intelligence
- Jimmy J. Lin
- Rodrigo Nogueira
- Yaoliang Yu
- Ji Xin
- Yoshitomo Matsubara
- Marco Levorato
- Francesco Restuccia
- Zeynep Akkalyoncu Yilmaz
In Computer Vision and Pattern Recognition
- Jimmy J. Lin
- Rodrigo Nogueira
- Yoshitomo Matsubara
- Yaoliang Yu
- Ji Xin
- Francesco Restuccia
- Marco Levorato
- Zhuyun Dai
Citations in the corpus
- Early Exiting BERT for Efficient Document Ranking (2020, methodology): “We compare against Cascade Trans-former (Soldaini and Moschitti, 2020), a recently proposed technique to accelerate inference in BERT-based document ranking.”
- Split Computing and Early Exiting for Deep Learning Applications: Survey and Research Challenges (2021, methodology): “Specifically, these studies [24, 67, 87, 109, 132, 140, 141, 148, 150, 155, 157, 158, 168, 172] define a loss function for each of the classifier, and minimize the weighted sum of cross entropy losses per sample as follows: where 3) by treating the final classifier (last exit) as teacher model and…”