Semi-supervised Textual Analysis and Historical Research Helping Each Other: Some Thoughts and Observations

Nanni, Federico; Kümper, Hiram; Ponzetto, Simone Paolo

doi:10.3366/ijhac.2016.0160

Cited by 7 publications

(3 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we present and address the entity-aspect linking task in particular to meet the demands of the elds of Digital Humanities and Computational Social Science [15,34]. ese communities o en employ entity-linking as a semantically explicit alternative to Latent Dirichlet Allocations (LDA) [3,4] for corpus exploration and topic-based document selection [19,31], as latent topics detected by LDA are often di cult to interpret and evaluate [7,24,45]. To satisfy the need of ne-grained interpretable topics, we have developed a system that adopts and provides explicit representations of pre-de ned entity-aspects harvested from Wikipedia.…”

Section: Advantagesmentioning

confidence: 99%

Entity-Aspect Linking

Nanni

Ponzetto²,

Dietz

2018

Proceedings of the 18th ACM/IEEE on Joint Conference on Digital Libraries

Self Cite

View full text Add to dashboard Cite

e availability of entity linking technologies provides a novel way to organize, categorize, and analyze large textual collections in digital libraries. However, in many situations a link to an entity o ers only relatively coarse-grained semantic information. is is problematic especially when the entity is related to several different events, topics, roles, and -more generally -when it has di erent aspects. In this work, we introduce and address the task of entity-aspect linking: given a mention of an entity in a contextual passage, we re ne the entity link with respect to the aspect of the entity it refers to. We show that a combination of di erent features and aspect representations in a learning-to-rank se ing correctly predicts the entity-aspect in 70% of the cases. Additionally, we demonstrate signi cant and consistent improvements using entityaspect linking on three entity prediction and categorization tasks relevant for the digital library community.

show abstract

Section: Advantagesmentioning

confidence: 99%

Entity-Aspect Linking

Nanni

Ponzetto²,

Dietz

2018

Proceedings of the 18th ACM/IEEE on Joint Conference on Digital Libraries

Self Cite

View full text Add to dashboard Cite

show abstract

“…First of all, the absence of a solid knowledge of data analysis has serious consequences for the humanities scholar who intends to use text mining methodologies, since it can limit both his/her capacity to engineer/re-adjust features and to adapt the chosen computational technique (as also remarked in [57]). Moreover, his/her understanding of quantitative results will be always partial, compared to the one exhibited by other researchers from other disciplines (such as computational linguistics [37], natural language processing [58] or information retrieval [36]) that are currently also experimenting with text mining methods to solve humanities tasks.…”

Section: A Generation Of Humanists-machine Learning Experienced Usersmentioning

confidence: 99%

From Close to Distant and Back: How to Read with the Help of Machines

Bonfiglioli¹,

Nanni

2016

IFIP Advances in Information and Communication Technology

Self Cite

View full text Add to dashboard Cite

In recent years a common trend characterised by the adoption of text mining methods for the study of digital sources emerged in digital humanities, often in opposition to traditional hermeneutic approaches. In our paper, we intend to show how text mining methods will always need a strong support from the humanist. On the one hand we remark how humanities research involving computational techniques should be thought of as a three steps process: from close reading (identification of a specific case study, initial feature selection) to distant reading (text mining analysis) to close reading again (evaluation of the results, interpretation, use of the results). Moreover, we highlight how failing to understand the importance of all the three steps is a major cause for the mistrust in text mining techniques developed around the humanities. On the other hand we observe that text mining techniques could be a very promising tool for the humanities and that researchers should not renounce to such approaches, but should instead experiment with advanced methods such as the ones belonging to the family of deep learning. In this sense we remark that, especially in the field of digital humanities, exploiting complementarity between computational methods and humans will be the most advantageous research direction.

show abstract

“…As a direct consequence of this fact, digital humanities scholars are currently stuck in a situation where they adopt topic models because they have a strong need for the potential benefits offered by such a method, especially now that large collections of primary sources are available for the first time in digital format. However, at the same time, scholars cannot derive new humanities knowledge from adopting topic models, given the current limitations of the results obtained (Schmidt 2012b;Nanni, Kümper, and Ponzetto 2016). Specific Contribution.…”

Section: Introductionmentioning

confidence: 99%

Entities as Topic Labels: Combining Entity Linking and Labeled LDA to Improve Topic Interpretability and Evaluability

Lauscher¹,

Nanni²,

Fabo³

et al. 2016

ijcol

Self Cite

View full text Add to dashboard Cite

Digital humanities scholars strongly need a corpus exploration method that provides topics easier to interpret than standard LDA topic models. To move towards this goal, here we propose a combination of two techniques, called Entity Linking and Labeled LDA. Our method identifies in an ontology a series of descriptive labels for each document in a corpus. Then it generates a specific topic for each label. Having a direct relation between topics and labels makes interpretation easier; using an ontology as background knowledge limits label ambiguity. As our topics are described with a limited number of clear-cut labels, they promote interpretability and support the quantitative evaluation of the obtained results. We illustrate the potential of the approach by applying it to three datasets, namely the transcription of speeches from the European Parliament fifth mandate, the Enron Corpus and the Hillary Clinton Email Dataset. While some of these resources have already been adopted by the natural language processing community, they still hold a large potential for humanities scholars, part of which could be exploited in studies that will adopt the fine-grained exploration method presented in this paper.

show abstract

Semi-supervised Textual Analysis and Historical Research Helping Each Other: Some Thoughts and Observations

Cited by 7 publications

References 10 publications

Entity-Aspect Linking

Entity-Aspect Linking

From Close to Distant and Back: How to Read with the Help of Machines

Entities as Topic Labels: Combining Entity Linking and Labeled LDA to Improve Topic Interpretability and Evaluability

Contact Info

Product

Resources

About