Zobrazeno 1 - 10
of 1 388
pro vyhledávání: '"A. Pêcher"'
Autor:
Cegin, Jan, Pecher, Branislav, Simko, Jakub, Srba, Ivan, Bielikova, Maria, Brusilovsky, Peter
The generative large language models (LLMs) are increasingly used for data augmentation tasks, where text samples are paraphrased (or generated anew) and then used for classifier fine-tuning. Existing works on augmentation leverage the few-shot scena
Externí odkaz:
http://arxiv.org/abs/2410.10756
While fine-tuning of pre-trained language models generally helps to overcome the lack of labelled training samples, it also displays model performance instability. This instability mainly originates from randomness in initialisation or data shuffling
Externí odkaz:
http://arxiv.org/abs/2406.12471
While learning with limited labelled data can improve performance when the labels are lacking, it is also sensitive to the effects of uncontrolled randomness introduced by so-called randomness factors (e.g., varying order of data). We propose a metho
Externí odkaz:
http://arxiv.org/abs/2402.12817
When solving NLP tasks with limited labelled data, researchers can either use a general large language model without further update, or use a small number of labelled examples to tune a specialised smaller model. In this work, we address the research
Externí odkaz:
http://arxiv.org/abs/2402.12819
In few-shot learning, such as meta-learning, few-shot fine-tuning or in-context learning, the limited number of samples used to train a model have a significant impact on the overall success. Although a large number of sample selection strategies exi
Externí odkaz:
http://arxiv.org/abs/2402.03038
Autor:
Cegin, Jan, Pecher, Branislav, Simko, Jakub, Srba, Ivan, Bielikova, Maria, Brusilovsky, Peter
The latest generative large language models (LLMs) have found their application in data augmentation tasks, where small numbers of text samples are LLM-paraphrased and then used to fine-tune downstream models. However, more research is needed to asse
Externí odkaz:
http://arxiv.org/abs/2401.06643
Learning with limited labelled data, such as prompting, in-context learning, fine-tuning, meta-learning or few-shot learning, aims to effectively train a model using only a small amount of labelled samples. However, these approaches have been observe
Externí odkaz:
http://arxiv.org/abs/2312.01082
Publikováno v:
Proceedings of the 17th International Workshop on Semantic Evaluation (SemEval-2023)
This paper presents the best-performing solution to the SemEval 2023 Task 3 on the subtask 3 dedicated to persuasion techniques detection. Due to a high multilingual character of the input data and a large number of 23 predicted labels (causing a lac
Externí odkaz:
http://arxiv.org/abs/2304.11924
Autor:
Srba, Ivan, Moro, Robert, Tomlein, Matus, Pecher, Branislav, Simko, Jakub, Stefancova, Elena, Kompan, Michal, Hrckova, Andrea, Podrouzek, Juraj, Gavornik, Adrian, Bielikova, Maria
Publikováno v:
ACM Transactions on Recommender Systems. 1, 1, Article 6 (March 2023), 33 pages
In this paper, we present results of an auditing study performed over YouTube aimed at investigating how fast a user can get into a misinformation filter bubble, but also what it takes to "burst the bubble", i.e., revert the bubble enclosure. We empl
Externí odkaz:
http://arxiv.org/abs/2210.10085
Autor:
Srba, Ivan, Pecher, Branislav, Tomlein, Matus, Moro, Robert, Stefancova, Elena, Simko, Jakub, Bielikova, Maria
Publikováno v:
ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2022)
False information has a significant negative influence on individuals as well as on the whole society. Especially in the current COVID-19 era, we witness an unprecedented growth of medical misinformation. To help tackle this problem with machine lear
Externí odkaz:
http://arxiv.org/abs/2204.12294