TY - GEN
T1 - Benchmarking Named Entity Recognition Approaches for Extracting Research Infrastructure Information from Text
AU - Cheirmpos, Georgios
AU - Tabatabaei, Seyed Amin
AU - Kanoulas, Evangelos
AU - Tsatsaronis, Georgios
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
PY - 2024
Y1 - 2024
N2 - Named entity recognition (NER) is an important component of many information extraction and linking pipelines. The task is especially challenging in a low-resource scenario, where there is very limited amount of high quality annotated data. In this paper we benchmark machine learning approaches for NER that may be very effective in such cases, and compare their performance in a novel application; information extraction of research infrastructure from scientific manuscripts. We explore approaches such as incorporating Contrastive Learning (CL), as well as Conditional Random Fields (CRF) weights in BERT-based architectures and demonstrate experimentally that such combinations are very efficient in few-shot learning set-ups, verifying similar findings that have been reported in other areas of NLP, as well as Computer Vision. More specifically, we show that the usage of CRF weights in BERT-based architectures achieves noteworthy improvements in the overall NER task by approximately 12%, and that in few-shot setups the effectiveness of CRF weights is much higher in smaller training sets.
AB - Named entity recognition (NER) is an important component of many information extraction and linking pipelines. The task is especially challenging in a low-resource scenario, where there is very limited amount of high quality annotated data. In this paper we benchmark machine learning approaches for NER that may be very effective in such cases, and compare their performance in a novel application; information extraction of research infrastructure from scientific manuscripts. We explore approaches such as incorporating Contrastive Learning (CL), as well as Conditional Random Fields (CRF) weights in BERT-based architectures and demonstrate experimentally that such combinations are very efficient in few-shot learning set-ups, verifying similar findings that have been reported in other areas of NLP, as well as Computer Vision. More specifically, we show that the usage of CRF weights in BERT-based architectures achieves noteworthy improvements in the overall NER task by approximately 12%, and that in few-shot setups the effectiveness of CRF weights is much higher in smaller training sets.
KW - Contrastive Learning
KW - Few-Shot Learning
KW - Named Entity Recognition
KW - Natural Language Processing
UR - http://www.scopus.com/inward/record.url?scp=85187687854&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-53969-5_11
DO - 10.1007/978-3-031-53969-5_11
M3 - Contribución a la conferencia
AN - SCOPUS:85187687854
SN - 9783031539688
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 131
EP - 141
BT - Machine Learning, Optimization, and Data Science - 9th International Conference, LOD 2023
A2 - Nicosia, Giuseppe
A2 - Ojha, Varun
A2 - La Malfa, Emanuele
A2 - La Malfa, Gabriele
A2 - Pardalos, Panos M.
A2 - Umeton, Renato
PB - Springer Science and Business Media Deutschland GmbH
T2 - 9th International Conference on Machine Learning, Optimization, and Data Science, LOD 2023
Y2 - 22 September 2023 through 26 September 2023
ER -