Papers by Iacopo Ghinassi
Recent Trends in Linear Text Segmentation: A Survey (2024.findings-emnlp)
Copied to clipboard
| Challenge: | Linear text segmentation is the task of automatically tagging text documents with topic shifts . the task is based on coherence modeling and/or local cues to identify topic boundaries . |
| Approach: | They provide an overview of current advances in linear text segmentation . they highlight limitations of available resources and of the task itself . |
| Outcome: | The proposed task is based on the most recent literature and under-explored research directions. |
Language Pivoting from Parallel Corpora for Word Sense Disambiguation of Historical Languages: A Case Study on Latin (2024.lrec-main)
Copied to clipboard
| Challenge: | Word Sense Disambiguation (WSD) is an important task in NLP . most of the work on this task has been done on contemporary English or other modern languages, leaving challenges posed by low-resource languages and diachronic change open. |
| Approach: | They propose to use existing bilingual corpora instead of native English datasets to generate a Latin WSD model. |
| Outcome: | The proposed approach achieves state-of-the-art on a standard benchmark for Latin WSD. |
Efficient Solutions For An Intriguing Failure of LLMs: Long Context Window Does Not Mean LLMs Can Analyze Long Sequences Flawlessly (2025.coling-main)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) have demonstrated remarkable capabilities in comprehending and analyzing lengthy sequential inputs. |
| Approach: | They propose to implement ad-hoc solutions that enhance LLMs’ performance on long input sequences by up to 50% while reducing API cost and latency by up . to address this limitation, they propose to use three datasets and two tasks to analyze news categorization and sentence analysis to evaluate their models. |
| Outcome: | The proposed solutions significantly improve LLMs’ performance on long input sequences by up to 50% while reducing API cost and latency by up . to 93% and 50%, respectively. |
When Cohesion Lies in the Embedding Space: Embedding-Based Reference-Free Metrics for Topic Segmentation (2024.lrec-main)
Copied to clipboard
| Challenge: | Recent advances in topic segmentation have led to a surge in interest in reference-free metrics, designed to score a hypothesised segmentation of a document without the need to refer to any expert annotation. |
| Approach: | They propose a common framework for reference-free topic segmentation metrics and a new method for the embedding space. |
| Outcome: | The proposed framework outperforms existing metrics based on human annotations while allowing for conversational data to outperformed other metrics. |