Publication
ForgetIT Deliverable D3.4: Strategies and Components for Managed Forgetting - Final Release
Xiaofei Zhu; Claudia Niederée; Tuan Tran; Andrea Ceroni; Kaweh Djafari Naini; Nam Khanh Tran; Heiko Maus; Christian Jilek
Deliverable, ForgetIT Consortium, ForgetIT Deliverables, Vol. D3.4, 1/2016.
Abstract
In the previous deliverables of this Work package, we have reported our work on managed forgetting, including research on memory buoyancy as well as first work on the preservation value and the policy framework. In this deliverable, we deepen our work on preservation value, present our work on the policy-based preservation framework, revisit the dimension of preservation value and relate them to appraisal. Furthermore, we discuss methods and components for preservation value assessment in different scenarios, as well as applications which realise these methods and components. Specifically, we have conducted a number of studies on the problem of estimating the preservation value for images, the preservation value for text, as well as the preservation value for social media. Regarding the preservation value for images, we extended the work reported in [Kanhabua et al., 2015] by including orthogonal information into the selection model, performing an extensive evaluation, involving more users and gathering more photo collections, and so on. Regarding the preservation value for text, we investigated different settings. For example, we aim at estimating the preservation values of text related to an entity of interest, and investigate how much it contributes to a summary of an entity’s (life) situation profile. Regarding the preservation value in social media, we extended our work of public text by conducting a study that uses entities as a pivot to evaluate the past news event’s preservation value. Furthermore, we discuss applications which realise the managed forgetting methods and components, including Memory Buoyancy for decluttering semantic information spaces and preservation value calculation in the semantic desktop.
Projekte
ForgetIT - Concise Preservation by combining Managed Forgetting and Contextualized Remembering,
CoMem - Corporate Memory
CoMem - Corporate Memory