Papers by Haim Dubossarsky
Analyzing Semantic Change through Lexical Replacements (2024.acl-long)
Copied to clipboard
| Challenge: | Modern language models can contextualize words based on their surrounding contexts, but semantic change can compromise this capability. |
| Approach: | They propose a replacement schema where a target word is replaced with lexical replacements of varying relatedness . they leverage the replacement schema as a basis for a novel interpretable model for semantic change . |
| Outcome: | The proposed model is the first to evaluate LLaMa for semantic change detection . it shows that lexical replacements can detect unexpected contexts . |
Logical Reasoning with Span-Level Predictions for Interpretable and Robust NLI Models (2022.emnlp-main)
Copied to clipboard
| Challenge: | Current models learn from annotation artefacts and dataset biases, but it is unclear to what extent they are learning the task of NLI. |
| Approach: | They propose a logical reasoning framework that allows models to learn from annotation artefacts and dataset biases. |
| Outcome: | The proposed model outperforms humans on in-distribution test sets without using span labels . the model is more robust in a reduced data setting, and out-of-disturbance performance is improved . |
Strengthening the WiC: New Polysemy Dataset in Hindi and Lack of Cross Lingual Transfer (2024.lrec-main)
Copied to clipboard
| Challenge: | a new study addresses the problem of natural language processing in low-resource languages such as Hindi . the paper focuses on Word Sense Disambiguation, a fundamental NLP task that deals with polysemous words. |
| Approach: | They propose a Hindi WSD dataset that allows training and testing of contextualized models. |
| Outcome: | The proposed dataset enables training and testing of contextualized models in Hindi . the results show that the proposed dataset can handle polysemy tasks in low-resource languages . |
(Chat)GPT v BERT Dawn of Justice for Semantic Change Detection (2024.findings-eacl)
Copied to clipboard
| Challenge: | In the universe of Natural Language Processing, Transformer-based language models like BERT and (Chat)GPT have emerged as lexical superheroes with great power to solve open research problems. |
| Approach: | They propose to use (Chat)GPT to solve two diachronic extensions of the Word-in-Context task: TempoWiC and HistoWic. |
| Outcome: | The proposed technology performs significantly worse than the foundational GPT version of (Chat)GPT for studying semantic change. |
Time-Out: Temporal Referencing for Robust Modeling of Lexical Semantic Change (P19-1)
Copied to clipboard
| Challenge: | State-of-the-art lexical semantic change detection models suffer from noise stemming from vector space alignment. |
| Approach: | They propose a method to simulate lexical semantic change and control for possible biases by avoiding alignment. |
| Outcome: | The proposed method outperforms state-of-the-art models on a synthetic task and a manual testset. |
Computational modeling of semantic change (2024.eacl-tutorials)
Copied to clipboard
| Challenge: | Languages change constantly over time, influenced by social, technological, cultural and political factors that affect how people express themselves. |
| Approach: | They propose to categorise the types of change, the causes and the mechanisms underlying the different types of changes using large diachronic corpora and evaluation benchmarks. |
| Outcome: | In historical linguistics, tools and methods have been developed to analyse the process . they include categorisations of types of change, causes and mechanisms . but traditional methods, while informative, are often based on small, carefully curated samples. |
DWUG: A large Resource of Diachronic Word Usage Graphs in Four Languages (2021.emnlp-main)
Copied to clipboard
| Challenge: | Existing methods for graded contextual word meaning annotation have not been implemented yet. |
| Approach: | They propose a multi-round incremental annotation process and a clustering algorithm to group usages into senses to create a large-scale dataset. |
| Outcome: | The proposed method is the largest resource of graded contextualized, diachronic word meaning annotation in four different languages, based on 100,000 human semantic proximity judgments. |
Definition Generation for Word Meaning Modeling: Monolingual, Multilingual, and Cross-Lingual Perspectives (2025.emnlp-main)
Copied to clipboard
| Challenge: | Recent advances in text generation have opened up new opportunities for word meaning modeling. |
| Approach: | They extend definition generation task beyond English to a suite of 22 languages . they use Llama-based models to evaluate models in monolingual, multilingual, cross-lingual settings . |
| Outcome: | The proposed model outperforms pretrained models in monolingual, multilingual, and cross-lingual settings. |
LSC-Eval: A General Framework to Evaluate Methods for Assessing Dimensions of Lexical Semantic Change Using LLM-Generated Synthetic Data (2025.findings-acl)
Copied to clipboard
| Challenge: | Existing methods for measuring Lexical Semantic Change are lacking historical benchmarks. |
| Approach: | They propose a three-stage general-purpose evaluation framework that simulates theory-driven LSC using In-Context Learning and a lexical database. |
| Outcome: | The proposed framework evaluates the sensitivity of computational methods to synthetic change and their suitability for detecting change in specific dimensions and domains. |
Atomic Inference for NLI with Generated Facts as Atoms (2024.emnlp-main)
Copied to clipboard
| Challenge: | Existing models that can provide accurate explanations are not interpretable, i.e. they do not reflect the inner workings of the model. |
| Approach: | They propose to use LLM-generated facts as atoms to make interpretable models that can be used to make accurate predictions for each component part of an input. |
| Outcome: | The proposed method outperforms existing methods on natural language understanding tasks with a multi-stage fact generation process and a training regime that incorporates the facts. |
Toward Sentiment Aware Semantic Change Analysis (2024.eacl-srw)
Copied to clipboard
| Challenge: | Current approaches to analyze semantic change are lagging behind . current methods only detect semantic change as a binary classification or graded change scores . |
| Approach: | They propose to augment models of semantic change with sentiment information . they demonstrate that existing models extract reliable sentiment information from historical corpora . |
| Outcome: | The proposed approach shows mixed results on the English SemEval of Lexical Semantic Change and its associated historical corpora. |
Avoiding the Hypothesis-Only Bias in Natural Language Inference via Ensemble Adversarial Training (2020.emnlp-main)
Copied to clipboard
| Challenge: | Neural models pick up on annotation artefacts and spurious correlations, resulting in learning sentences that suffer from the same biases. |
| Approach: | They propose to tackle this problem by using adversarial training to reduce the bias in sentence representations by using an ensemble of adversaries. |
| Outcome: | The proposed approach produces more robust models outperforming previous de-biasing efforts when generalised to 12 other NLI datasets. |
Coming to Your Senses: on Controls and Evaluation Sets in Polysemy Research (D18-1)
Copied to clipboard
| Challenge: | a prior art claim that sense-specific vectors provide an advantage over normal vectors is unfounded in two ways. |
| Approach: | They claim that sense-specific vectors provide an advantage over normal vectors due to the polysemy that they presumably represent. |
| Outcome: | The proposed results show that ground-truth polysemy degrades performance in word similarity tasks and that random assignment of words to senses improves performance. |
The Secret is in the Spectra: Predicting Cross-lingual Task Performance with Spectral Similarity Measures (2020.emnlp-main)
Copied to clipboard
| Challenge: | Existing studies have suggested that bilingual lexicon induction is influenced by the (dis)similarity of the languages at hand. |
| Approach: | They propose to measure the isomorphism of monolingual embedding spaces based on their spectra and introduce isometric measures to measure their similarity. |
| Outcome: | The proposed measures outperform standard isomorphism measures while being more tractable and easier to interpret. |
Multilinguality Does not Make Sense: Investigating Factors Behind Zero-Shot Cross-Lingual Transfer in Sense-Aware Tasks (2025.emnlp-main)
Copied to clipboard
| Challenge: | Cross-lingual transfer allows models to perform tasks in languages unseen during training and is often assumed to benefit from increased multilinguality. |
| Approach: | They challenge this assumption by analyzing polysemy disambiguation and lexical semantic change in 28 languages and using confounding factors to account for perceived advantages. |
| Outcome: | The proposed models and benchmarks are compared across 28 languages and show that multilingual training is neither necessary nor beneficial for effective transfer. |