Papers by Guillaume Wisniewski

15 papers
Beyond Surprisal: A Dual Metric Framework for Lexical Skill Acquisition in LLMs (2025.coling-main)

Copied to clipboard

Challenge: Existing learning curves capture when and how a model learns to use words correctly, but they neglect the equally important skill of avoiding incorrect usage.
Approach: They propose a new metric which measures a model's capacity to refrain from using words in unexpected or unexpected contexts.
Outcome: The proposed metric measures the model's ability to refrain from using words in unexpected or unexpected contexts.
What Do Neural Speech Models Know About Phonology? Evidence from Structured Phoneme Confusions (2026.findings-acl)

Copied to clipboard

Challenge: acoustic and phonological models of speech recognition are often limited to the phoneme level . a recent study has shown that phoneme confusions are strongly structured in phonology space .
Approach: They adopt a featural representation of phonemes grounded in phonological theory which models speech sounds as structured bundles of distinctive articulatory and acoustic properties.
Outcome: The proposed model allows us to analyse phoneme confusions at a finer granularity and to investigate whether certain phonological features are more vulnerable than others.
Establishing degrees of closeness between audio recordings along different dimensions using large-scale cross-lingual models (2024.findings-eacl)

Copied to clipboard

Challenge: Existing methods to analyze speech representations using pretraining data are difficult to achieve for endangered languages.
Approach: They propose an unsupervised method to examine the level of abstraction in vector representations of speech from a pretrained model to determine their level of abstractness.
Outcome: The proposed method is fully unsupervised and could be used in comparative studies on under-documented languages.
Phonetic Normalization for Machine Translation of User Generated Content (D19-55)

Copied to clipboard

Challenge: a method to correct noisy User Generated Content (UGC) in French is proposed . it leverages on the existence of UGC specific noise due to the misuse of words with similar pronunciations.
Approach: They propose a phonetizer-based method to correct noisy User Generated Content (UGC) they use phonetic similarity to generate IPA pronunciations of words .
Outcome: The proposed method improves translation quality of noisy User Generated Content (UGC) in french.
Using Artificial French Data to Understand the Emergence of Gender Bias in Transformer Language Models (2023.emnlp-main)

Copied to clipboard

Challenge: Existing studies have demonstrated the ability of neural language models to learn linguistic properties without direct supervision.
Approach: They propose to use an artificial corpus generated by a PCFG to control the gender distribution in training data and determine under which conditions a model correctly captures gender information.
Outcome: The proposed approach allows to control the gender distribution in training data and determine under which conditions a model correctly captures gender information or appears gender-biased.
Exploiting Dynamic Oracles to Train Projective Dependency Parsers on Non-Projective Trees (N18-2)

Copied to clipboard

Challenge: Several strategies have been proposed to overcome the projectivity constraint by introducing transition-based dependency parsers that can build non-projective dependencies.
Approach: They propose a modification of dynamic oracles to allow use of non-projective data . their method consistently outperforms traditional projectivization and pseudo-projectivisation approaches .
Outcome: The proposed method outperforms projectivization and pseudo-projectivisation methods on 73 treebanks and achieves significant gains for non-projective languages.
Automatically Selecting the Best Dependency Annotation Design with Dynamic Oracles (N18-2)

Copied to clipboard

Challenge: Multiple annotation conventions have been proposed for representing dependency structures.
Approach: They propose to consider a set of syntactic references encoding alternative syntak representations to train a parser with a dynamic oracle.
Outcome: The proposed approach can predict the best syntactic representation among all possible references.
Errator: a Tool to Help Detect Annotation Errors in the Universal Dependencies Project (L18-1)

Copied to clipboard

Challenge: UD project aims to develop cross-linguistically consistent treebank annotations for a wide array of languages.
Approach: They introduce tools that implement the annotation variation principle to help annotators find and correct errors in UD treebanks.
Outcome: The proposed tools can be used to correct errors in UD treebank annotations.
Quantifying training challenges of dependency parsers (C18-1)

Copied to clipboard

Challenge: a new metric is introduced to evaluate the difficulty to learn a given class of dependencies . a series of systematic computations using that metric have revealed interesting properties of the 3 considered parsing algorithms .
Approach: They introduce a new metric to evaluate the difficulty to learn a given class of dependencies . they use it to characterize the information conveyed by cross-lingual parsers .
Outcome: The proposed metric reveals the kind of dependencies that require high effort during training . it also shows that cross-lingual parsers can provide better quality information .
Are Transformers a Modern Version of ELIZA? Observations on French Object Verb Agreement (2021.emnlp-main)

Copied to clipboard

Challenge: Recent studies have shown that unsupervised sentence representations of neural networks encode syntactic information by observing that neural language models are able to predict the agreement between a verb and its subject.
Approach: They propose to take an alternative look at these results by studying whether neural networks are able to build an abstract sentence representation rather than capture surface statistical regularities.
Outcome: The proposed model can achieve high accuracy on the long-range French object-verb agreement, indicating a possible flaw in the model's syntactic ability.
Automated Paraphrase Lattice Creation for HyTER Machine Translation Evaluation (N18-2)

Copied to clipboard

Challenge: Existing machine translation evaluation metrics use synonyms and paraphrases to reward meaning-equivalent but lexically divergent translations.
Approach: They propose a machine translation evaluation metric which exploits reference translations enriched with meaning equivalent expressions.
Outcome: The proposed metric achieves medium performance on large and noisier datasets . it is compared with the existing HyTER evaluation metric .
How Bad are PoS Tagger in Cross-Corpora Settings? Evaluating Annotation Divergence in the UD Project. (N19-1)

Copied to clipboard

Challenge: Using annotation variation principles, Part-of-Speech tagging performance degrades when applied to test sentences that depart from training data.
Approach: They propose to use the annotation variation principle to identify inconsistencies between annotations . they also evaluate their impact on prediction performance .
Outcome: The proposed method can detect errors in gold standard annotations and improve prediction performance.
Are Neural Networks Extracting Linguistic Properties or Memorizing Training Data? An Observation with a Multilingual Probe for Predicting Tense (2021.eacl-main)

Copied to clipboard

Challenge: a recent study has shown that neural networks can learn from linguistic representations without supervision . many studies have tried to identify which linguistic properties are encoded in the embeddings .
Approach: They evaluate the ability of Bert embeddings to represent tense information . they use a multilingual linguistic probe to predict the morphology of a word .
Outcome: The proposed model can predict tenses in French and Chinese, but the results drop sharply for Chinese.
Assessing the Capacity of Transformer to Abstract Syntactic Representations: A Contrastive Analysis Based on Long-distance Agreement (2023.tacl-1)

Copied to clipboard

Challenge: Existing studies have shown that transformers are able to predict subject-verb agreement, demonstrating their ability to uncover an abstract representation of the sentence in an unsupervised way.
Approach: They propose to compare how transformers handle subject-verb and object-past participle agreements in French using probing and counterfactual analysis methods.
Outcome: The proposed model handles subject-verb and object-past participle agreements in a way consistent with their modeling in theoretical linguistics.
How Distributed are Distributed Representations? An Observation on the Locality of Syntactic Information in Verb Agreement Tasks (2022.acl-short)

Copied to clipboard

Challenge: Using probing, causal analysis and feature selection, we find that syntactic information is encoded locally in the transformers representations consistent with the French grammar.
Approach: They address the question of the localization of syntactic information encoded in transformers representations by probing, causal analysis and feature selection methods.
Outcome: The proposed representations are consistent with the object-past participle agreement in French and are consistent in both languages.

What is GenGO?

GenGO is an NLP powered publication search system. It currenctly indexes 30k+ papers from ACL Anthology, and implements multi-aspect summarization, semantic search, and more!

Information

About
Limitations