Probing Linguistic Features of Sentence-Level Representations in Relation Extraction
By Christoph Alt, Aleksandra Gabryszak, Leonhard Hennig. 2020
36 citations recorded by the source corpus.
Despite the recent progress, little is known about the features captured by state-of-the-art neural relation extraction (RE) models. Common methods encode the source sentence, conditioned on the entity mentions, before classifying the relation. However, the complexity of the task makes it difficult to understand how encoder architecture and supporting linguistic knowledge affect the features learned by the encoder. We introduce 14 probing tasks targeting linguistic properties relevant to RE, and we use them to study representations learned by more than 40 different encoder architecture and linguistic feature combinations trained on two datasets, TACRED and SemEval 2010 Task 8. We find that the bias induced by the architecture and the inclusion of linguistic features are clearly expressed in the probing task performance. For example, adding contextualized word representations greatly increases performance on probing tasks with a focus on named entity and part-of-speech information, and yields better results in RE. In contrast, entity masking improves RE, but considerably lowers performance on entity type related probing tasks.
Subfields: Representation & Self-Supervised Learning, Information Extraction
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Representation & Self-Supervised Learning
- Yonatan Belinkov
- Yanai Elazar
- Eduard H. Hovy
- Abhilasha Ravichander
- Shauli Ravfogel
- Yoav Goldberg
- Frank Rudzicz
- Alon Jacovi
In Information Extraction
- Eduard H. Hovy
- Yonatan Belinkov
- Yoav Goldberg
- Frank Rudzicz
- Abhilasha Ravichander
- Zining Zhu
- Eliyahu Kiperwasser
- Joakim Nivre
Citations in the corpus
- Amnesic Probing: Behavioral Explanation with Amnesic Counterfactuals (2020, methodology): “By using the probing technique, different linguistic phenomenon such as POS, dependency information and NER (Tenney et al., 2019a; Liu et al., 2019a; Alt et al., 2020) have been found to be “easily extractable” (typically using linear probes).”
- Probing the Probing Paradigm: Does Probing Accuracy Entail Task Relevance? (2020, methodology): “body of work uses this approach to pinpoint the information models rely on to do a task (Alt et al., 2020; Giulianelli et al., 2018; Saleh et al., 2020).”
- An Information Theoretic View on Selecting Linguistic Probes (2020, background): “Recently, various probing tasks have queried the representations of, e.g., contextualized word embeddings (Tenney et al., 2019a,b) and sentence embeddings (Linzen et al., 2016; Chen et al., 2019; Alt et al., 2020; Kassner and Sch¨utze, 2020; Maudslay et al., 2020; Chi et al., 2020).”