Keyphrase Extraction via an Attentive Model

Passon, Marco; Comuzzo, Massimo; Serra, Giuseppe; Tasso, Carlo

doi:10.1007/978-3-030-11226-4_24

Cited by 6 publications

(2 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Text mining techniques play a crucial role in Digital Libraries to automatically extract information used to serve user's needs better. Serra [26] proposed an approach to keyphrase extraction via an Attentive Model, a neural network designed to focus on the most relevant parts of data. Carducci [7] presented a system combining standard and semantic learning for automatically annotating bibliographic records.…”

Section: Open Science Publishing and Scientific Workflowsmentioning

confidence: 99%

Digital Libraries

et al. 2020

View full text Add to dashboard Cite

The Italian Research Conference on Digital Libraries (IRCDL) is the annual Italian forum to discuss research topics on Digital Libraries and related technical, practical, and social issues. Along the years, IRCDL touched several aspects underlying the ?Digital Library" domain and promptly adapted to the evolution of the field. Today, the ?Digital Library" field includes theory and practices reflecting the evolution of the role of libraries in the scholarly communication domain, and also embracing scholarly communication and open science. The theme of IRCDL 2019 was ?Digital Libraries: Supporting Open Science". Three main reasons motivated this theme: (i) science is increasingly becoming digital, meaning that research is performed using data services and digital tools; (ii) the results of the research are no longer just traditional scientific publications; (iii) the outcomes of science are increasingly encompassing datasets, software, and experiments. As digital artifacts, such products can be shared and re-used together with the article, thus enabling comprehensive research assessment and various degrees of reproducibility of science. Positive consequences of this shift towards Open Science are: accelerating science, optimizing the cost of research, fraud detection, and fullyfledged scientific reward. Digital Libraries are central in the evolution of research outputs by targeting findability, preservation, interlinking, and re-use of research products and by integrating the components of the scholarly communication process. The conference has been organized in Pisa, and the proceedings are published in the Springer CCIS series Vol. 988 [22]. Pre-print versions, research datasets, and research software relative to the accepted contributions are accessible via Zenodo.org.

show abstract

Section: Open Science Publishing and Scientific Workflowsmentioning

confidence: 99%

Digital Libraries

et al. 2020

View full text Add to dashboard Cite

show abstract

“…Second, we introduce a novel spatio-temporal attention mechanism with the aim to select relevant information from different areas of the frames of the input video, and from their evolution over time. Attention mechanisms have been largely exploited in a variety of different implementations and in many different domains of Deep Learning such as Natural Language Processing [9] and Computer Vision [10]. The intuition behind Attention in Computer Vision is to mimic the human visual process.…”

Section: Introductionmentioning

confidence: 99%

Video-Based Convolutional Attention for Person Re-Identification

Zamprogno

Passon

Martinel

et al. 2019

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

In this paper we consider the problem of video-based person re-identification, which is the task of associating videos of the same person captured by different and non-overlapping cameras. We propose a Siamese framework in which video frames of the person to re-identify and of the candidate one are processed by two identical networks which produce a similarity score. We introduce an attention mechanisms to capture the relevant information both at frame level (spatial information) and at video level (temporal information given by the importance of a specific frame within the sequence). One of the novelties of our approach is given by a joint concurrent processing of both frame and video levels, providing in such a way a very simple architecture. Despite this fact, out approach achieves better performance than the state-of-the-art on the challenging iLIDS-VID dataset.

show abstract

Digital Libraries: Supporting Open Science

2019

Communications in Computer and Information Science

View full text Add to dashboard Cite

Keyphrase Extraction via an Attentive Model

Cited by 6 publications

References 19 publications

Digital Libraries

Digital Libraries

Video-Based Convolutional Attention for Person Re-Identification

Digital Libraries: Supporting Open Science

Contact Info

Product

Resources

About