Papers by Yuval Pinter
Attention is not not Explanation (D19-1)
Copied to clipboard
| Challenge: | Attention mechanisms play a central role in NLP systems, especially within recurrent neural network (RNN) models. |
| Approach: | They propose to use a simple uniform-weights baseline, a variance calibration and a diagnostic framework to determine when/whether attention can be used as explanation in RNN models. |
| Outcome: | The proposed tests show that even reliable adversarial distributions don't perform well on the simple diagnostic, indicating that prior work does not disprove the usefulness of attention mechanisms for explainability. |
Learning to Faithfully Rationalize by Construction (2020.acl-main)
Copied to clipboard
| Challenge: | Neural models dominate NLP but it remains difficult to know why they make specific predictions for sequential text inputs. |
| Approach: | They propose a model to produce faithful rationales for neural text classification by defining independent snippet extraction and prediction modules. |
| Outcome: | The proposed model produces faithful explanations even when the model is complex and complex. |
Si O No, Que Penses? Catalonian Independence and Linguistic Identity on Social Media (N18-2)
Copied to clipboard
| Challenge: | This study examines the use of Catalan on Twitter in discourse related to the 2017 independence referendum. |
| Approach: | They use code-switching to determine the role of Catalan in political discourse . they corroborate prior findings that pro-independence tweets are more likely to include the local language than anti-independent tweets . |
| Outcome: | The proposed method corroborates previous findings that pro-independence tweets are more likely to include the local language than anti-independent tweets. |
Predicting Semantic Relations using Global Graph Properties (D18-1)
Copied to clipboard
| Challenge: | Semantic graphs encode the structural qualities of language as a representation of human knowledge. |
| Approach: | They propose a global-theoretic model that integrates global and local properties of semantic graphs to improve local prediction of relational relations between synsets. |
| Outcome: | The proposed model improves on the local task of predicting semantic relations between synsets, yielding state-of-the-art results on the WN18RR dataset. |
Universal NER: A Gold-Standard Multilingual Named Entity Recognition Benchmark (2024.naacl-long)
Copied to clipboard
Stephen Mayhew, Terra Blevins, Shuheng Liu, Marek Suppa, Hila Gonen, Joseph Marvin Imperial, Börje Karlsson, Peiqin Lin, Nikola Ljubešić, Lester James Miranda, Barbara Plank, Arij Riabi, Yuval Pinter
| Challenge: | In named entity recognition, the majority of annotation efforts are centered on English, and cross-lingual transfer performance remains brittle. |
| Approach: | They propose to develop gold-standard named entity recognition benchmarks in many languages using a cross-lingual consistent schema. |
| Outcome: | The proposed benchmarks will be released to the public in 2022 . they will provide baselines on in-language and cross-lingual learning settings. |
Will it Unblend? (2020.findings-emnlp)
Copied to clipboard
| Challenge: | Blends, such as “innoventor”, are one particularly challenging class of OOV terms, as they are formed by fusing together two or more bases that relate to the intended meaning in unpredictable manners and degrees. |
| Approach: | They propose to use a dataset of English OOV blends to quantify the difficulty of interpreting the meanings of blends by large-scale contextual language models such as BERT. |
| Outcome: | The proposed model outperforms character-level and context-free embeddings, although their results are still far from satisfactory. |
CIAug: Equipping Interpolative Augmentation with Curriculum Learning (2022.naacl-main)
Copied to clipboard
| Challenge: | Current methods for interpolative data augmentation select samples at random, which might make it difficult for the model to generalize better and converge faster. |
| Approach: | They propose a curriculum-based learning method that leverages the relative position of samples in hyperbolic embedding space as a complexity measure to gradually mix up increasingly difficult and diverse samples along training. |
| Outcome: | The proposed method achieves state-of-the-art results over existing methods on 10 benchmark datasets across 4 languages in text classification and named-entity recognition tasks. |
Restoring Hebrew Diacritics Without a Dictionary (2022.findings-naacl)
Copied to clipboard
| Challenge: | a number of modern Hebrew texts are written in a letter-only version of the Hebrew script, which omits the diacritics present in the full diacritized, or dotted variant. |
| Approach: | They propose a character-level LSTM that can accurately diacritize Hebrew script without human-curated resources. |
| Outcome: | The proposed model performs on par with complex curation-dependent systems across a diverse array of modern Hebrew sources. |
Probing Subphonemes in Morphology Models (2025.findings-acl)
Copied to clipboard
| Challenge: | morphological inflection models have achieved state-of-the-art performance, yet their ability to generalize across languages and morphology rules remains limited. |
| Approach: | They propose a language-agnostic probing method to investigate phonological feature encoding in transformers trained directly on phonemes and perform it across seven morphologically diverse languages. |
| Outcome: | The proposed method shows that phonological features which are local are captured well in phoneme embeddings, whereas long-distance dependencies like vowel harmony are better represented in the transformer’s encoder. |
Which Pieces Does Unigram Tokenization Really Need? (2026.findings-acl)
Copied to clipboard
| Challenge: | Despite its theoretical elegance, its implementation in practice is complex, limiting its adoption to SentencePiece. |
| Approach: | They propose a Unigram-based probabilistic alternative to the greedy heuristics of Byte-Pair Encoding that is based on C++. |
| Outcome: | The proposed algorithm is remarkably robust to hyperparameter choices and can be simplified to reduce computational costs. |
Emptying the Ocean with a Spoon: Should We Edit Models? (2023.findings-emnlp)
Copied to clipboard
| Challenge: | a recent study has questioned the use of direct model editing for factual corrections in LLMs. aaron s. de stefano, a sociologist, says that model editing is not a systematic remedy for factuality. |
| Approach: | They argue that direct model editing cannot be trusted as a remedy for LLM disadvantages . authors call for cautious promotion and application of model editing as part of LLM deployment process . |
| Outcome: | The proposed method is not trusted as a remedy for the disadvantages inherent to LLMs, the authors argue . they argue that it opens risks by reinforcing the notion that models can be trusted for factuality . |
Tokenization Matters: Navigating Data-Scarce Tokenization for Gender Inclusive Language Technologies (2024.findings-naacl)
Copied to clipboard
Anaelia Ovalle, Ninareh Mehrabi, Palash Goyal, Jwala Dhamala, Kai-Wei Chang, Richard Zemel, Aram Galstyan, Yuval Pinter, Rahul Gupta
| Challenge: | a recent study documented the harmful limitations of gender binary-centric large language models . data scarcity is a known culprit, but the precise mechanisms through which scarcity affects this behavior remain underexplored. |
| Approach: | They propose to use BPE tokenization to enforce consistent tokenization across gendered pronouns to improve neopronoun proficiency. |
| Outcome: | The proposed methods outperform finetuning with standard BPE, and improve neopronoun proficiency. |
1,729 vs. 1729: The Effect of Scripts and Formats on LLM Numeracy (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) have impressive proficiency in basic arithmetic, but little attention has been given to how they perform when numerical expressions deviate from the prevailing conventions present in their training corpora. |
| Approach: | They investigate numerical reasoning across a wide range of numeral scripts and formats . they show that LLM accuracy drops substantially when numerical inputs are rendered in underrepresented scripts or formats despite the underlying mathematical reasoning being identical . |
| Outcome: | The proposed methods can narrow the gap between LLMs and human models when they deviate from prevailing numerical conventions. |
Tokenization Is More Than Compression (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing tokenization approaches like Byte-Pair Encoding (BPE) have been suggested that their effectiveness stems from their ability to condense text into a relatively small number of tokens. |
| Approach: | They propose a tokenizer that segments a document’s text into the minimum number of tokens for a given vocabulary and propose fewer tokens to improve downstream performance. |
| Outcome: | The proposed tokenizers can initialize vocabulary construction and pre-tokenization, and the results show that fewer tokens lead to better performance. |
UniMorph 4.0: Universal Morphology (2022.lrec-1)
Copied to clipboard
Khuyagbaatar Batsuren, Omer Goldman, Salam Khalifa, Nizar Habash, Witold Kieraś, Gábor Bella, Brian Leonard, Garrett Nicolai, Kyle Gorman, Yustinus Ghanggo Ate, Maria Ryskina, Sabrina Mielke, Elena Budianskaya, Charbel El-Khaissi, Tiago Pimentel, Michael Gasser, William Abbott Lane, Mohit Raj, Matt Coler, Jaime Rafael Montoya Samame, Delio Siticonatzi Camaiteri, Esaú Zumaeta Rojas, Didier López Francis, Arturo Oncevay, Juan López Bautista, Gema Celeste Silva Villegas, Lucas Torroba Hennigen, Adam Ek, David Guriel, Peter Dirix, Jean-Philippe Bernardy, Andrey Scherbakov, Aziyana Bayyr-ool, Antonios Anastasopoulos, Roberto Zariquiey, Karina Sheifer, Sofya Ganieva, Hilaria Cruz, Ritván Karahóǧa, Stella Markantonatou, George Pavlidis, Matvey Plugaryov, Elena Klyachko, Ali Salehi, Candy Angulo, Jatayu Baxi, Andrew Krizhanovsky, Natalia Krizhanovskaya, Elizabeth Salesky, Clara Vania, Sardana Ivanova, Jennifer White, Rowan Hall Maudslay, Josef Valvoda, Ran Zmigrod, Paula Czarnowska, Irene Nikkarinen, Aelita Salchak, Brijesh Bhatt, Christopher Straughn, Zoey Liu, Jonathan North Washington, Yuval Pinter, Duygu Ataman, Marcin Wolinski, Totok Suhardijanto, Anna Yablonskaya, Niklas Stoehr, Hossep Dolatian, Zahroh Nuriah, Shyam Ratan, Francis M. Tyers, Edoardo M. Ponti, Grant Aiton, Aryaman Arora, Richard J. Hatcher, Ritesh Kumar, Jeremiah Young, Daria Rodionova, Anastasia Yemelina, Taras Andrushko, Igor Marchenko, Polina Mashkovtseva, Alexandra Serova, Emily Prud’hommeaux, Maria Nepomniashchaya, Fausto Giunchiglia, Eleanor Chodroff, Mans Hulden, Miikka Silfverberg, Arya D. McCarthy, David Yarowsky, Ryan Cotterell, Reut Tsarfaty, Ekaterina Vylomova
| Challenge: | The Universal Morphology project provides broad-coverage instantiated morphological inflection tables for hundreds of diverse languages. |
| Approach: | They propose a language-independent feature schema for rich morphological annotation and a type-level resource of annotated data in diverse languages realizing that schema. |
| Outcome: | The proposed schema has added 66 new languages, including 24 endangered languages. |
Don’t Touch My Diacritics (2025.naacl-short)
Copied to clipboard
| Challenge: | a recent paper examines the effects of preprocessing text with diacritics on model performance . we show that inconsistent encoding of diacritized characters and removing diacritical characters can have detrimental downstream effects . |
| Approach: | They propose to improve the handling of diacritized text by preserving diacritics and removing them altogether. |
| Outcome: | The proposed approach reduces the number of errors in the preprocessing process, the authors argue . they show that the proposed approach can reduce the number and complexity of errors . |
Splintering Nonconcatenative Languages for Better Tokenization (2025.findings-acl)
Copied to clipboard
| Challenge: | Subword tokenizers are not effective in nonconcatenative languages such as Hebrew and Arabic, where morphology is encoded in root-template patterns, or Malay and Georgian, where split affixes are common. |
| Approach: | They propose a pre-processing step which rearranges text into a linear form that better represents nonconcatenative morphologies. |
| Outcome: | The proposed method is based on measures evaluating token vocabularies in Hebrew, Arabic, and Malay, and on BERT-architecture models trained for Hebrew. |
BiVert: Bidirectional Vocabulary Evaluation Using Relations for Machine Translation (2024.lrec-main)
Copied to clipboard
| Challenge: | Neural machine translation (NMT) evaluation is crucial to determine the quality of translations. |
| Approach: | They propose a bidirectional semantic-based evaluation method to assess the sense distance of the translation from the source text. |
| Outcome: | The proposed method uses the multilingual encyclopedic dictionary BabelNet . it shows a strong correlation between the evaluation scores and human assessments . |
Incorporating Context into Subword Vocabularies (2023.eacl-main)
Copied to clipboard
| Challenge: | Current tokenizers are trained on word frequency statistics over a corpus without considering information about co-occurrence or context. |
| Approach: | They propose a tokenizer that bakes in contextualized signal at the vocabulary creation phase to tailor subwords for their downstream use. |
| Outcome: | The proposed tokenizer is able to keep token contexts cohesive while not incurring a large price in terms of encoding efficiency or domain robustness. |
UniMorph 3.0: Universal Morphology (2020.lrec-1)
Copied to clipboard
Arya D. McCarthy, Christo Kirov, Matteo Grella, Amrit Nidhi, Patrick Xia, Kyle Gorman, Ekaterina Vylomova, Sabrina J. Mielke, Garrett Nicolai, Miikka Silfverberg, Timofey Arkhangelskiy, Nataly Krizhanovsky, Andrew Krizhanovsky, Elena Klyachko, Alexey Sorokin, John Mansfield, Valts Ernštreits, Yuval Pinter, Cassandra L. Jacobs, Ryan Cotterell, Mans Hulden, David Yarowsky
| Challenge: | Explicit modeling of morphology has demonstrable benefits for language modeling, speech recognition, word embedding and keyword search. |
| Approach: | They propose a language-independent feature schema for rich morphological annotation and a type-level resource for annotated data in diverse languages. |
| Outcome: | The proposed schema has been improved to make it more complete and correct, and adds 66 new languages and parts of speech for 12 languages. |
NYTWIT: A Dataset of Novel Words in the New York Times (2020.coling-main)
Copied to clipboard
| Challenge: | Novel words, or Out-Of-Vocabulary words, are a pervasive problem in modern natural language processing. |
| Approach: | They present a dataset of over 2,500 novel English words published in the New York Times . they use uncontextual and contextual predictions to predict novelty class . |
| Outcome: | The proposed dataset includes over 2,500 novel English words published in the New York Times between November 2017 and March 2019 . baseline results show that there is room for improvement even for state-of-the-art NLP systems . |
Analyzing Cognitive Plausibility of Subword Tokenization (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing evaluations of subword tokenization focus on engineering criteria such as compression rate . a recent study evaluated subwords for their cognitive plausibility in languages with limited vocabulary size . |
| Approach: | They propose a new evaluation paradigm that focuses on the cognitive plausibility of subword tokenization. |
| Outcome: | The proposed tokenization algorithm yields less cognitively plausible tokenization behavior and worse coverage of derivational morphemes than previous evaluations. |