Papers by Oren Tsur
With Measured Words: Simple Sentence Selection for Black-Box Optimization of Sentence Compression Algorithms (2021.eacl-main)
Copied to clipboard
| Challenge: | Sentence Compression is the task of generating a shorter, yet grammatical, version of a given sentence, preserving the essence of the original sentence. |
| Approach: | They propose a Black-Box Optimizer for Compression to find the best candidates for compression . they use a black-box compression algorithm to predict how well each sentence could be compressed . |
| Outcome: | The proposed algorithm improves both accuracy and Rouge-F1-score on three datasets. |
A Deeper (Autoregressive) Approach to Non-Convergent Discourse Parsing (2023.emnlp-main)
Copied to clipboard
| Challenge: | Existing frameworks for dialogic discourse parsing are not suitable for contentious discussions . authors propose a model for non-convergent discourse paring that does not require label collocation . |
| Approach: | They propose a multi-label scheme for contentious dialog parsing that uses multiple labels . they propose combining embeddings of the utterance, context and the labels through GRN layers . |
| Outcome: | The proposed model achieves comparable results with SOTA without label collocation and without training a unique architecture/model for each label. |
Acquired TASTE: Multimodal Stance Detection with Textual and Structural Embeddings (2025.coling-main)
Copied to clipboard
| Challenge: | Prior work has demonstrated the importance of the conversational context in stance detection. |
| Approach: | They propose a multimodal architecture for stance detection that fuses transformer-based content embedding with unsupervised structural embeddment. |
| Outcome: | The proposed model outperforms strong baselines on common benchmarks and outperformed existing models on common frameworks. |
How to Do Things without Words: Modeling Semantic Drift of Emoji (2022.findings-emnlp)
Copied to clipboard
| Challenge: | Emoji have become a significant part of our informal textual communication. |
| Approach: | They propose to model and analyze the semantic drift of emoji and explore the relations between graphical changes and semantic changes. |
| Outcome: | The proposed model and analysis examines the relationship between graphical changes and semantic drift. |
Open-Mindedness and Style Coordination in Argumentative Discussions (2021.eacl-main)
Copied to clipboard
| Challenge: | Previous research has shown that linguistic accommodation correlates with gaps in the power and status of the speakers and the way it promotes approval and discussion efficiency. |
| Approach: | They propose a novel perspective on linguistic accommodation, exploring its correlation with the open-mindedness of a speaker, rather than to her social status. |
| Outcome: | The proposed approach improves the open-mindedness of a speaker and lowers discussion efficiency. |
Detecting Suicide Risk in Online Counseling Services: A Study in a Low-Resource Language (2022.coling-1)
Copied to clipboard
| Challenge: | Existing domain-specific models for detecting suicide are lacking in low-resource languages. |
| Approach: | They propose a model that combines pre-trained language models with a fixed set of suicidal cues and a two-stage fine-tuning process to detect SI. |
| Outcome: | The proposed model outperforms baseline models even early on in the conversation and performs well across genders and age groups. |