NTCIR-12 Lifelog Data Analytics

Li, Na; Gurrin, Cathal; Crane, Martin; Ruskin, Heather J.

doi:10.1145/2983576.2983583

Cited by 4 publications

(2 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We study the effectiveness of PARADE on the TREC DL Track and NTCIR WWW-3 Track. We report results in this section and refer the readers to the TREC and NTCIR task papers for details on the specific hyperparameters used [53,54].…”

Section: Results On the Trec DL Track And Ntcir Www-3 Trackmentioning

confidence: 99%

PARADE: Passage Representation Aggregation forDocument Reranking

Yates

MacAvaney

et al. 2023

ACM Trans. Inf. Syst.

View full text Add to dashboard Cite

Pre-trained transformer models, such as BERT and T5, have shown to be highly effective at ad-hoc passage and document ranking. Due to the inherent sequence length limits of these models, they need to process document passages one at a time rather than processing the entire document sequence at once. Although several approaches for aggregating passage-level signals into a document-level relevance score have been proposed, there has yet to be an extensive comparison of these techniques. In this work, we explore strategies for aggregating relevance signals from a document’s passages into a final ranking score. We find that passage representation aggregation techniques can significantly improve over score aggregation techniques proposed in prior work, such as taking the maximum passage score. We call this new approach PARADE. In particular, PARADE can significantly improve results on collections with broad information needs where relevance signals can be spread throughout the document (such as TREC Robust04 and GOV2). Meanwhile, less complex aggregation techniques may work better on collections with an information need that can often be pinpointed to a single passage (such as TREC DL and TREC Genomics). We also conduct efficiency analyses and highlight several strategies for improving transformer-based aggregation.

show abstract

Section: Results On the Trec DL Track And Ntcir Www-3 Trackmentioning

confidence: 99%

PARADE: Passage Representation Aggregation forDocument Reranking

Yates

MacAvaney

et al. 2023

ACM Trans. Inf. Syst.

View full text Add to dashboard Cite

show abstract

“…Over the past decade, one can see the increasing use of shared lifelog data as a source of evidence for epidemiological studies, notwithstanding the increasing concerns regarding personal data privacy from individuals, organisations and societies. A review of related literature highlights the application of lifelogging tools as a means of human memory understanding (Milton et al , 2011; Barnard et al , 2011), for supporting human memory (Berry et al , 2007; Harvey et al , 2016), for facilitating large-scale epidemiological studies in health-care (Signal et al , 2017), lifestyle monitoring (Wilson et al , 2018; Nguyen et al , 2016), diet/obesity analytics (Zhou et al , 2019), or behaviour analysis (Everson et al , 2019; Li et al , 2016a, 2016b). A typical feature of such activities is the use of wearable camera data, numbering thousands of images per day, captured by individuals in real-world settings.…”

Section: History and Backgroundmentioning

confidence: 99%

Considerations on privacy in the era of digitally logged lives

Gupta

Crane

Gurrin

2020

OIR

Self Cite

View full text Add to dashboard Cite

PurposeThe continuous advancements in wearable sensing technologies enable the easy collection and publishing of visual lifelog data. The widespread adaptation of visual lifelog technologies would have the potential to pose challenges for ensuring the personal privacy of subjects and bystanders in lifelog data. This paper presents preliminary findings from a study of lifeloggers with the aim of better understanding their concerns regarding privacy in lifelog data.Design/methodology/approachIn this study, we have collected a visual dataset of 64,837 images from 25 lifelogging participants over a period of two days each, and we conducted an interactive session (face to face conversation) with each participant in order to capture their concerns when sharing the lifelog data across three specified categories (i.e. Private (Only for Me), Semi-Private (Family/Friends) and Public).FindingsIn general, we found that participants tend to err on the side of conservative privacy settings and that there is a noticeable difference in what different participants are willing to share. In summary, we found that the categories of images that the participants wished to be kept private included personally identifiable information and professional information; categories of images that could be shared with family/friends include family moments or content related to daily routine lifestyle, and other visual lifelog data could potentially be made public).Originality/valueWe analysed the potential differences in the willingness of 25 participants to share data. In addition, reasons for being a volunteer to collect lifelog data and how the lifelogging device affected the lifestyle of the lifelogger are analysed. Based on the findings of this study, we propose a set of challenges for the anonymisation of lifelog data that should be solved when supporting lifelog data sharing.

show abstract