Papers by Kfir Bar
Motivational Interviewing Transcripts Annotated with Global Scores (2024.lrec-main)
Copied to clipboard
| Challenge: | Motivational interviewing (MI) is a counseling approach that aims to increase intrinsic motivation and commitment to change. |
| Approach: | They propose to annotate MI therapy sessions written in English from public sources . they explore the potential use of the dataset for training MI language models . |
| Outcome: | The proposed dataset includes 242 MI demonstration transcripts annotated with therapist behavioral codes and global scores and client language EAsy Rating (CLEAR) tags for client speech. |
Improving Few-Shot Domain Transfer for Named Entity Disambiguation with Pattern Exploitation (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Named entity disambiguation is a critical subtask of entity linking . a model can be trained on a domain, but it needs to be adapted to the domain . |
| Approach: | They propose to reformulate named entity disambiguation as a masked language modeling problem. |
| Outcome: | The proposed model improves on a mental health news dataset without sacrifices in accuracy. |
Temporal Relation Classification using Boolean Question Answering (2023.findings-acl)
Copied to clipboard
| Challenge: | a new approach for temporal relation classification (TRC) is proposed . a boolean question answering model is used to classify temporal relations between two events . |
| Approach: | They propose an efficient approach for temporal relation classification using a boolean question answering model based on TRC annotation guidelines. |
| Outcome: | The proposed model outperforms state-of-the-art models by 2.4% on questions designed by human annotation experts. |
DiaSet: An Annotated Dataset of Arabic Conversations (2024.lrec-main)
Copied to clipboard
Abraham Israeli, Aviv Naaman, Guy Maduel, Rawaa Makhoul, Dana Qaraeen, Amir Ejmail, Dina Lisnanskey, Julian Jubran, Shai Fine, Kfir Bar
| Challenge: | DiaSet is a dataset of dialectical Arabic speech manually transcribed and annotated for two downstream tasks. |
| Approach: | They propose to manually transcribe and annotate Arabic speech for sentiment analysis and named entity recognition. |
| Outcome: | The proposed dataset encapsulates the Palestine dialect, predominantly spoken in Palestine, Israel, and Jordan. |
ID10M-JAM: Stress-Testing Idiom Identification Under Challenging Context (2026.findings-acl)
Copied to clipboard
| Challenge: | Large language models (LLMs) achieve strong performance on idiom identification benchmarks, yet their robustness to misleading contextual signals remains largely untested. |
| Approach: | They propose an adversarial extension of the ID10M dataset that jams idiom understanding by injecting coherent but conflicting context before each target sentence. |
| Outcome: | The proposed benchmark exposes systematic vulnerabilities in LLMs’ contextual reasoning, pushing idiom identification to its breaking point. |
Layer Duplication in LLMs (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Duplicating multihead self-attention layers once or twice yields significant performance boosts. |
| Approach: | They investigate the effect of duplicating multihead self-attention layers in large language models . |
| Outcome: | The proposed method improves performance across a range of language tasks with or without fine-tuning. |
Easy as PIE? Identifying Multi-Word Expressions with LLMs (2025.emnlp-main)
Copied to clipboard
| Challenge: | Multiword expressions (MWEs) are a semantically non-compositional subclass of multiword expression . authors show that prompt-based LLMs can perform competitively with supervised models . |
| Approach: | They propose a prompt-based approach to identify idiomatic expressions in running text . they find prompt-driven LLMs can perform competitively with supervised models . |
| Outcome: | The proposed approach can perform well with supervised models on annotated data. |
Beyond Pairwise: Global Zero-shot Temporal Graph Generation (2025.emnlp-main)
Copied to clipboard
| Challenge: | Temporal relation extraction (TRE) is a fundamental task in natural language processing (NLP) that involves identifying the temporal relationships between events in a document. |
| Approach: | They propose a method that generates a document’s complete temporal graph in a single step, followed by temporal constraint optimization to refine predictions and enforce temporal consistency across relations. |
| Outcome: | The proposed method outperforms existing zero-shot approaches and offers a competitive alternative to supervised TRE models. |
JRC-Names-Retrieval: A Standardized Benchmark for Name Search (2024.lrec-main)
Copied to clipboard
| Challenge: | a lack of research on the ability to search through databases of personal and organization name is hindering this area . specialized indexing methods which understand the structure of names are essential to efficient performance. |
| Approach: | They propose a neural solution which provides a 12% performance gain over baselines . they propose specialized indexing methods which understand the structure of names . |
| Outcome: | The proposed solution shows up to 12% performance gain over baselines . the proposed solution is compared against a similar dataset from a different dataset . |
Code-Switching and Back-Transliteration Using a Bilingual Model (2024.findings-eacl)
Copied to clipboard
| Challenge: | a wealth of Judeo-Arabic texts is available on the internet, but most speakers of Arabic are unfamiliar with the Hebrew script. |
| Approach: | They propose two machine-learning models for transliterating Judeo-Arabic into Arabic . they also propose a model for identifying non-Arabian words, predominantly Hebrew and Aramaic . |
| Outcome: | The proposed models outperform previous solutions for transliterating Judeo-Arabic into Arabic and identifying non-Arabian words, mainly Hebrew and Aramaic. |
How Much Pretraining Does Structured Data Need? (2026.eacl-long)
Copied to clipboard
| Challenge: | Large language models are increasingly adopted for handling structured data, despite pretraining on unstructured text. |
| Approach: | They propose to re-initialize subsets of layers with random weights before fine-tuning on structured datasets. |
| Outcome: | The proposed models are compared to unstructured datasets and show that they perform well over structured data. |