| Challenge: | Neural network-based language models (LMs) have been shown to learn relevant properties of language without being explicitly trained for them. |
| Approach: | They extend their previous work to analyze whether language models capture anaphoric relations and pronoun-antecedent relations in English. |
| Outcome: | The Transformer outperforms the LSTM in all analyses. |
Similar Papers
Predicting Reference: What do Language Models Learn about Discourse Models? (2020.emnlp-main)
Copied to clipboard
| Challenge: | a growing literature that probes neural language models to assess their latent acquisition of grammatical knowledge has not investigated their acquisition of discourse modeling ability. |
| Approach: | They draw on a psycholinguistic literature that has established how different contexts affect referential biases concerning who is likely to be referred to next. |
| Outcome: | The proposed models do not resemble human language users, the authors show . their models capture the linguistic knowledge required to perform discourse modeling . |
Targeted Syntactic Evaluation of Language Models (D18-1)
Copied to clipboard
| Challenge: | Recent advances have led to an explosion of neural network-based LM architectures. |
| Approach: | They propose to supplement perplexity with a metric that assesses whether a language model can predict the grammatical sentence more accurately than an ungrammatically-based model. |
| Outcome: | The proposed model performed poorly on many of the constructions. |
Visual Referring Expression Recognition: What Do Systems Actually Learn? (N18-2)
Copied to clipboard
| Challenge: | Existing systems for referring expression recognition ignore linguistic structure, instead relying on shallow correlations introduced by unintended biases in the data selection and annotation process. |
| Approach: | They propose to use a system trained on the input image without the input referring expression to achieve a precision of 71.2% in top-2 predictions. |
| Outcome: | The proposed model can achieve 71.2% accuracy on the input image without the input referring expression and 84.2% on the object category given the input. |
Transparency Helps Reveal When Language Models Learn Meaning (2023.tacl-1)
Copied to clipboard
| Challenge: | Existing language models are trained to optimize unsupervised objectives on text . despite their centrality, current models do not represent natural language semantics well . |
| Approach: | They show that autoregressive and masked language models learn to emulate semantic relations between expressions when context-dependent . they argue that a learner that has access to all Java code can never learn execution . |
| Outcome: | a new study shows that language models fail to represent natural language semantics well . the authors show that the model learning fails when denotations are changed to be context-dependent . |
Probing Relational Knowledge in Language Models via Word Analogies (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Existing studies have focused on probing relational knowledge by filling the blanks in pre-defined prompts such as “The capital of France is —” but these are affected by the co-occurrence of target relation words and entities in the pre-training corpus. |
| Approach: | They extend probing methodologies by using analogical proportions as a proxy to probe relational knowledge in transformer-based PLMs without directly presenting the desired relation. |
| Outcome: | The proposed methods are extremely accurate at (1) and (2), but have room for improvement for (3). |
Probing Contextual Language Models for Common Ground with Visual Representations (2021.naacl-main)
Copied to clipboard
| Challenge: | Contextual language models have attracted great interest in probing what is encoded in their representations. |
| Approach: | They propose a probing model that evaluates how effective are text-only representations in distinguishing between matching and non-matching visual representations. |
| Outcome: | The proposed model outperforms text-only language models in instance retrieval, but underperform humans. |
A Closer Look at Linguistic Knowledge in Masked Language Models: The Case of Relative Clauses in American English (2020.coling-main)
Copied to clipboard
Marius Mosbach, Stefania Degaetano-Ortlieb, Marie-Pauline Krielke, Badr M. Abdullah, Dietrich Klakow
| Challenge: | Despite the high performance of transformer-based language models, we still lack understanding of the kind of linguistic knowledge they learn and rely on. |
| Approach: | They evaluate three transformer-based language models and test their grammatical and semantic knowledge by sentence-level probing, diagnostic cases, and masked prediction tasks. |
| Outcome: | The models capture grammatical and semantic knowledge, but they lack model-specific weaknesses especially on semantic knowledge. |
A Study on Accessing Linguistic Information in Pre-Trained Language Models by Using Prompts (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods to access linguistic information in pre-trained multilingual language models are difficult to use. |
| Approach: | They propose prompting and formulate linguistic tasks to test the LM's access to explicit grammar principles and find out what type of information can be obtained . |
| Outcome: | The proposed method can provide access to linguistic features in pre-trained models, but some are harder to capture . |
More room for language: Investigating the effect of retrieval on language models (2024.naacl-short)
Copied to clipboard
| Challenge: | Retrieval-augmented language models are a promising alternative to standard pretraining, but little attention has been put into understanding what this type of training scheme does to the underlying language model when analyzed as a standalone -separated from the overall retrieval pipeline. |
| Approach: | They propose an ‘ideal retrieval’ methodology to study these models in a fully controllable setting and propose a retrieval augmentation methodology to examine their effects. |
| Outcome: | The proposed model saves substantially less world knowledge in their weights, but is worse at comprehending global context. |
Probing Multilingual Cognate Prediction Models (2022.findings-acl)
Copied to clipboard
| Challenge: | linguistic interpretations of cognate prediction have been based on external analysis (accuracy, raw results, errors). |
| Approach: | They propose to use character-based machine translation models to store linguistic and diachronic information but not in previously assumed ways. |
| Outcome: | The proposed model stores linguistic and diachronic information but does not achieve it in previously assumed ways. |