From Balustrades to Pierre Vinken: Looking for Syntax in Transformer Self-Attentions
By David Marecek, Rudolf Rosa. 2019
57 citations recorded by the source corpus.
We inspect the multi-head self-attention in Transformer NMT encoders for three source languages, looking for patterns that could have a syntactic interpretation. In many of the attention heads, we frequently find sequences of consecutive states attending to the same position, which resemble syntactic phrases. We propose a transparent deterministic method of quantifying the amount of syntactic information present in the self-attentions, based on automatically building and evaluating phrase-structure trees from the phrase-like sequences. We compare the resulting trees to existing constituency treebanks, both manually and by computing precision and recall.
Subfields: Artificial Intelligence, Information Systems
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
- Yonatan Belinkov
- Qun Liu
- Alessandro Raganato
- Zhiyong Wu
- Yun Chen
- Jörg Tiedemann
- B. Kao
- Yves Scherrer
In Information Systems
Citations in the corpus
- Self-Attention Attribution: Interpreting Information Interactions Inside Transformer (2020, neutral): “, 2019) and attention weights (Marecek and Rosa, 2019).”
- Probing Classifiers: Promises, Shortcomings, and Alternatives (2021, background): “These scores may come from attention weights (Raganato and Tiedemann 2018; Clark et al. 2019; Mareˇcek and Rosa 2019; Htut et al. 2019) or from distances between word representations, perhaps including perturbations of the input sentence ( Wu et al. 2020).”
- Denoising Pre-Training and Data Augmentation Strategies for Enhanced RDF Verbalization with Transformers (2020, unknown): “Mareˇcek and Rosa (2019) has shown that the self-attention layers embed implicitly syntactic parse trees.”
- Fixed Encoder Self-Attention Patterns in Transformer-Based Machine Translation (2020, background): “The first line of research focuses on the interpretation of the network, in particular on the analysis of attention mechanisms and the interpretability of the weights and connections (Raganato and Tiedemann, 2018; Tang et al., 2018; Mareček and Rosa, 2019; Voita et al., 2019a; Brunner et al., 2020).”
- Perturbed Masking: Parameter-free Probing for Analyzing and Interpreting BERT (2020, background): “In their later study (Mareˇcek and Rosa, 2019), they propose a bottom-up algorithm to extract constituent trees from transformer-based NMT encoders and evaluate their results on three languages.”