基本信息
浏览量:166
职业迁徙
个人简介
The research carried out during the Ph.D. and during the PostDoc focused on the development
of novel methodologies to introduce semantics in content-based representation employed by
intelligent and adaptive platforms.
During the Ph.D., this led to the development of a framework, called eVSM (enhanced Vector Space Model), that exploits the typical strengths of Vector Space Model (VSM) and extends
it through a Quantum Negation operator as well as through Distributional Semantics Models
(DSM) to semantically represented the information about items and users. DSM implement
non-supervised approaches to represent terms (and documents as well) in large vector spaces
according to their co-occurrences in large corpora of data. The main advantage that follows
the adoption of DSM is that the semantics conveyed by terms is learned in an incremental way,
without any training.
In the experimental sessions the effectiveness of eVSM is evaluated in both online and
online settings, and it emerged that eVSM overcomes several state-of-the-art models in terms of
goodness of recommendations and accuracy of the proposed ranking. Specifically, it outperforms
in a significant way LSI, the classical VSM and a Bayes text classifier in the task of providing
users with recommendations about movies. The effectiveness of the approach was confirmed during the stage carried out in Philips Research, where the above described techniques were
used to recommend personalized TV shows.
sed to recommend personalized TV shows.
During the PostDoc the research has evolved by following two research lines: on one side, eVSM model was extended by modeling contextual information as well. This led to the definition
of a context-aware recommendation framework called Contextual eVSM which was evaluated,
with good results, against several state-of-the-art baselines. Next, the research focused on the
analysis of semantics-aware representation different from those based on DSM.
of novel methodologies to introduce semantics in content-based representation employed by
intelligent and adaptive platforms.
During the Ph.D., this led to the development of a framework, called eVSM (enhanced Vector Space Model), that exploits the typical strengths of Vector Space Model (VSM) and extends
it through a Quantum Negation operator as well as through Distributional Semantics Models
(DSM) to semantically represented the information about items and users. DSM implement
non-supervised approaches to represent terms (and documents as well) in large vector spaces
according to their co-occurrences in large corpora of data. The main advantage that follows
the adoption of DSM is that the semantics conveyed by terms is learned in an incremental way,
without any training.
In the experimental sessions the effectiveness of eVSM is evaluated in both online and
online settings, and it emerged that eVSM overcomes several state-of-the-art models in terms of
goodness of recommendations and accuracy of the proposed ranking. Specifically, it outperforms
in a significant way LSI, the classical VSM and a Bayes text classifier in the task of providing
users with recommendations about movies. The effectiveness of the approach was confirmed during the stage carried out in Philips Research, where the above described techniques were
used to recommend personalized TV shows.
sed to recommend personalized TV shows.
During the PostDoc the research has evolved by following two research lines: on one side, eVSM model was extended by modeling contextual information as well. This led to the definition
of a context-aware recommendation framework called Contextual eVSM which was evaluated,
with good results, against several state-of-the-art baselines. Next, the research focused on the
analysis of semantics-aware representation different from those based on DSM.
研究兴趣
论文共 154 篇作者统计合作学者相似作者
按年份排序按引用量排序主题筛选期刊级别筛选合作者筛选合作机构筛选
时间
引用量
主题
期刊级别
合作者
合作机构
USER MODELING AND USER-ADAPTED INTERACTIONpp.1-34, (2023)
引用0浏览0引用
0
0
Pierluigi Amodio,Lucia Siciliani,Linda Antonella Antonucci, Cristiano Tamborrino,Pierpaolo Basile,Paolo Taurisano,Marco de Gemmis,Felice Iavernaro,Pasquale Lops, Francesca Mazzia,Cataldo Musto,Marco Polignano,
Ital-IApp.341-346, (2023)
引用0浏览0EI引用
0
0
Information Systems (2023): 102273
SSRN Electronic Journalpp.1-32, (2023)
PROCEEDINGS OF THE 17TH ACM CONFERENCE ON RECOMMENDER SYSTEMS, RECSYS 2023pp.856-862, (2023)
PROCEEDINGS OF THE 17TH ACM CONFERENCE ON RECOMMENDER SYSTEMS, RECSYS 2023pp.1259-1262, (2023)
PROCEEDINGS OF THE 17TH ACM CONFERENCE ON RECOMMENDER SYSTEMS, RECSYS 2023pp.554-564, (2023)
加载更多
作者统计
合作学者
合作机构
D-Core
- 合作者
- 学生
- 导师
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn