Identifying and Controlling Important Neurons in Neural Machine Translation
arXiv:1811.01157
Abstract
Neural machine translation (NMT) models learn representations containing substantial linguistic information. However, it is not clear if such information is fully distributed or if some of it can be attributed to individual neurons. We develop unsupervised methods for discovering important neurons in NMT models. Our methods rely on the intuition that different models learn similar properties, and do not require any costly external supervision. We show experimentally that translation quality depends on the discovered neurons, and find that many of them capture common linguistic phenomena. Finally, we show how to control NMT translations in predictable ways, by modifying activations of individual neurons.
References in corpus (10)
- Convolutional Sequence to Sequence Learning
- Visualizing and Understanding Recurrent Networks
- Understanding Neural Networks through Representation Erasure
- Learning to Generate Reviews and Discovering Sentiment
- Fine-grained Analysis of Sentence Embeddings Using Auxiliary Prediction Tasks
- On the importance of single directions for generalization
- Convergent Learning: Do different neural networks learn the same representations?
- Revisiting the Importance of Individual Units in CNNs via Ablation
- Representation of linguistic form and function in recurrent neural networks
- Natural Language Multitasking: Analyzing and Improving Syntactic Saliency of Hidden Representations
Cited by in corpus (9)
- Rapid Learning or Feature Reuse? Towards Understanding the Effectiveness of MAML
- A Survey of Deep Learning for Scientific Discovery
- BERT Learns (and Teaches) Chemistry
- EX-RAY: Distinguishing Injected Backdoor from Natural Features in Neural Networks by Examining Differential Feature Symmetry
- Analysis Methods in Neural Language Processing: A Survey
- Neural Machine Translation: A Review and Survey
- Understanding and Controlling Memory in Recurrent Neural Networks
- How Do You Act? An Empirical Study to Understand Behavior of Deep Reinforcement Learning Agents
- Canoe : A System for Collaborative Learning for Neural Nets