Papers › KERMIT: Complementing Transformer Architectures with Encoders of Explicit Syntactic...

KERMIT: Complementing Transformer Architectures with Encoders of Explicit Syntactic Interpretations

1 Nov 2020EMNLP 2020 11archive 2025-07-28

Fabio Massimo Zanzotto, Andrea Santilli, Leonardo Ranaldi, Dario Onorati, Pierfrancesco Tommasino, Francesca Fallucchi

Syntactic parsers have dominated natural language understanding for decades. Yet, their syntactic interpretations are losing centrality in downstream tasks due to the success of large-scale textual representation learners. In this paper, we propose KERMIT (Kernel-inspired Encoder with Recursive Mechanism for Interpretable Trees) to embed symbolic syntactic parse trees into artificial neural networks and to visualize how syntax is used in inference. We experimented with KERMIT paired with two state-of-the-art transformer-based universal sentence encoders (BERT and XLNet) and we showed that KERMIT can indeed boost their performance by effectively embedding human-coded universal syntactic representations in neural networks

PaperPDFCode

Code

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Representation LearningSentenceSyntax RepresentationText Classification

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

Introduced by this paper: Syntax Heat Parse Tree

Absolute Position EncodingsAdamBERTDense ConnectionsDropoutPosition-Wise Feed-Forward LayerSyntax Heat Parse TreeTransformerXLNet

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections