Probing Pretrained Language Models for Lexical Semantics
By Ivan Vulic, Edoardo Maria Ponti, Robert Litschko, Goran Glavas, Anna Korhonen. 2020
286 citations recorded by the source corpus.
The success of large pretrained language models (LMs) such as BERT and RoBERTa has sparked interest in probing their representations, in order to unveil what types of knowledge they implicitly capture. While prior research focused on morphosyntactic, semantic, and world knowledge, it remains unclear to which extent LMs also derive lexical type-level knowledge from words in context. In this work, we present a systematic empirical analysis across six typologically diverse languages and five different lexical tasks, addressing the following questions: 1) How do different lexical knowledge extraction strategies (monolingual versus multilingual source LM, out-of-context versus in-context encoding, inclusion of special tokens, and layer-wise averaging) impact performance? How consistent are the observed effects across tasks and languages? 2) Is lexical knowledge stored in few parameters, or is it scattered throughout the network? 3) How do these representations fare against traditional static word vectors in lexical tasks? 4) Does the lexical information emerging from independently trained monolingual LMs display latent similarities? Our main results indicate patterns and best practices that hold universally, but also point to prominent variations across languages and tasks. Moreover, we validate the claim that lower Transformer layers carry more type-level lexical knowledge, but also show that this knowledge is distributed across multiple layers.
Subfields: Artificial Intelligence, Natural Language Processing, Computational Linguistics & Semantics
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
- Ivan Vulic
- Iryna Gurevych
- Sebastian Ruder
- Jonas Pfeiffer
- Mohit Bansal
- Graham Neubig
- Luke S. Zettlemoyer
- Anders Søgaard
In Natural Language Processing
- Ivan Vulic
- Jonas Pfeiffer
- Iryna Gurevych
- Sebastian Ruder
- Anders Søgaard
- Phillip Rust
- Ellie Pavlick
- Goran Glavas
In Computational Linguistics & Semantics
Citations in the corpus
- AdapterDrop: On the Efficiency of Adapters in Transformers (2020, neutral): “This is in line with recent work showing that some linguistic tasks heavily rely on information from the first layers (Vulić et al., 2020).”
- The Interplay of Variant, Size, and Task Type in Arabic Pre-trained Language Models (2021, background): “…why pre-trained language models perform well on downstream tasks either in monolingual (Hewitt and Manning, 2019; Jawahar et al., 2019; Liu et al., 2019a; Tenney et al., 2019a,b) or multi-lingual settings (Wu and Dredze, 2019; Chi et al., 2020; Kulmizev et al., 2020; Vuli´c et al., 2020).”
- Can Language Models Encode Perceptual Structure Without Grounding? A Case Study in Color (2021, methodology): “Representation Extraction We follow Bom-masani et al. (2020) and Vuli´c et al. (2020)”
- Self-Alignment Pretraining for Biomedical Entity Representations (2020, methodology): “6We tried multiple strategies including first-token, meanpooling, [CLS] and also NOSPEC (recommended by Vulić et al. 2020) but found no consistent best strategy (optimal strategy varies on different *BERTs). informative training examples (i.e. hard positive/negative pairs) within a mini-batch for…”
- Emergent Structures and Training Dynamics in Large Language Models (2022, background): “Pires et al. (2019) and Vuli´c et al. (2020) further showed – as expected – that the similarity depends on the typo-347 logical distance between the languages.”
- All Bark and No Bite: Rogue Dimensions in Transformer Language Models Obscure Representational Quality (2021, background): “…the representational similarity of word embedding spaces corresponds to human semantic similarity/relatedness judgments (Bom-masani et al., 2020; Vuli´c et al., 2020; Chronis and Erk, 2020; A. Rodriguez and Merlo, 2020), human brain activation patterns/cross-model similarity (Abnar et al.,…”
- How Good is Your Tokenizer? On the Monolingual Performance of Multilingual Language Models (2020, background): “Vulić et al. (2020) probed mBERT and monolingual BERT models across six typologically diverse languages (German, English, Chinese, Russian, Finnish, Turkish) for lexical semantics.”