Estimating PageRank deviations in crawled graphs

Holzmann, Helge; Anand, Avishek; Khosla, Megha

doi:10.1007/s41109-019-0201-9

Cited by 5 publications

(4 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Although scientists have looked into graph properties of the web in general, both in static (Albert et al, 1999;Broder et al, 2000;Adamic et al, 2000;Suel and Yuan, 2001;Boldi and Vigna, 2004) and evolving graphs (Huberman and Adamic, 1999;Leskovec et al, 2005Leskovec et al, , 2007, we found that certain traits of web archives lead to new kinds of questions. For instance, as we show in Holzmann et al (2018Holzmann et al ( , 2019, the inherent incompleteness of archives can affect rankings produced by graph algorithms on web archive graphs.…”

Section: Open Challengesmentioning

confidence: 83%

The Past Web

2021

View full text Add to dashboard Cite

Table Of Content 0.1 Dedication 0.2 Foreword 0.3 Preface Part 1 The era of information abundance and memory scarcity Chapter 1.0 Part introduction Chapter 1.1 The problem of web ephemera Chapter 1.2 Web archives preserve our digital collective memory Part 2 Collecting before it vanishes Chapter 2.0 Part introduction Chapter 2.1 National web archiving in Australia -representing the comprehensive Chapter 2.2 Web Archiving Singapore: The Realities of National Web Archiving Chapter 2.3 Archiving the social media -the Twitter case Chapter 2.4 Creating Event-Centric Collections from Web ArchivesPart 3 Access methods to analyse the Past web Chapter 3.0 Part introduction Chapter 3.1 Full-text and URL search Chapter 3.2 A holistic view on Web archives Chapter 3.3 Interoperability for Accessing Versions of Web Resources with the Memento Protocol Chapter 3.4 Linking Twitter archives with TV archives Chapter 3.5 Image analytics in web archives Part 4 Researching the past Web Chapter 4.0 Part introduction Chapter 4.1 Digital archaeology in the web of links: reconstructing a late-90s web sphere Chapter 4.2 Quantitative approaches to the Danish web archive Chapter 4.3 Critical Web Archive Research Chapter 4.4 Exploring Online Diasporas: London's French and Latin American Communities in the UK Web Archive Chapter 4.5 Platform and app histories: Assessing source availability in web archives and app repositories Part 5 Web archives as infrastructures to develop innovative tools Chapter 5.0 Part introduction Chapter 5.1 The need for infrastructures for the study of web-archived material Chapter 5.2 Automatic generation of timelines for past events Chapter 5.3 Political opinions of the past Web Chapter 5.4 Framing web archives with browsers contemporary to a website's creation Chapter 5.5 Big Data analytics over past web data Part 6 The Past Web: a look into the future.This book is dedicated to Vitalino Gomes who taught me that the value of a Man is in his Integrity.

show abstract

Section: Open Challengesmentioning

confidence: 83%

The Past Web

2021

View full text Add to dashboard Cite

show abstract

“…Our approach, therefore, allows us to attribute any decision to a small subset of the node neighborhood, hence increasing interpretability. We believe that our work can be extended to ML tasks in multiple domains like Web tasks [12,13], rankings [31,30] and tabular data [8].…”

Section: Discussionmentioning

confidence: 99%

Learnt Sparsification for Interpretable Graph Neural Networks

Rathee¹,

Zhang²,

Funke³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Graph neural networks (GNNs) have achieved great success on various tasks and fields that require relational modeling. GNNs aggregate node features using the graph structure as inductive biases resulting in flexible and powerful models. However, GNNs remain hard to interpret as the interplay between node features and graph structure is only implicitly learned. In this paper, we propose a novel method called KEdge for explicitly sparsifying the underlying graph by removing unnecessary neighbors. Our key idea is based on a tractable method for sparsification using the Hard Kumaraswamy distribution that can be used in conjugation with any GNN model. KEdge learns edge masks in a modular fashion trained with any GNN allowing for gradient-based optimization in an end-to-end fashion. We demonstrate through extensive experiments that our model KEdge can prune a large proportion of the edges with only a minor effect on the test accuracy. Specifically, in the PubMed dataset, KEdge learns to drop more than 80% of the edges with an accuracy drop of merely ≈ 2% showing that graph structure has only a small contribution in comparison to node features. Finally, we also show that KEdge effectively counters the over-smoothing phenomena in deep GNNs by maintaining good task performance with increasing GNN layers.

show abstract

“…In [14], the authors focused on the problem of the deviations in PageRank values caused by restricted crawling. Some further variation of traditional PageRank is proposed in [24] replacing the original transition matrix is replaced with one whose entries are based on the number of a node's N-step neighbours.…”

Section: Background and Related Work 21 Pagerank -A Measure Of Import...mentioning

confidence: 99%

A Comparative Analysis of Retrievability and PageRank Measures

Sinha,

Mall,

Roy

et al. 2023

Proceedings of the 15th Annual Meeting of the Forum for Information Retrieval Evaluation

View full text Add to dashboard Cite

The accessibility of documents within a collection holds a pivotal role in Information Retrieval, signifying the ease of locating specific content in a collection of documents. This accessibility can be achieved via two distinct avenues. The first is through some retrieval model using a keyword or other feature-based search, and the other is where a document can be navigated using links associated with them, if available. Metrics such as PageRank, Hub, and Authority illuminate the pathways through which documents can be discovered within the network of content while the concept of Retrievability is used to quantify the ease with which a document can be found by a retrieval model. In this paper, we compare these two perspectives, PageRank and retrievability, as they quantify the importance and discoverability of content in a corpus. Through empirical experimentation on benchmark datasets, we demonstrate a subtle similarity between retrievability and PageRank particularly distinguishable for larger datasets. CCS CONCEPTS• Information systems → Retrieval effectiveness.

show abstract

Estimating PageRank deviations in crawled graphs

Cited by 5 publications

References 34 publications

The Past Web

The Past Web

Learnt Sparsification for Interpretable Graph Neural Networks

A Comparative Analysis of Retrievability and PageRank Measures

Contact Info

Product

Resources

About