This diagram represents our vision of building upon the platform of lifelogging towards creating persistent personalised lifebraries. The realisation of this vision requires truly cross-disciplinary serendipitous collaboration including input from technologists, psychologists, sociologists, philosophers, ethnographers, artists, and ethics experts. We now provide some sample use-case scenarios highlighting the need for this cross-fertilisation of ideas. Use Case 1 – Archiving & Preserving Actively Created Personal Digital Material Luisa Gandarias at the age of 60 is still in the amazement of technological revolutions over the last 30 years. The memories of early days where the primary form of communication was used to postal service’s the technological revolutions have made it possible to share information across the globe with only a few clicks of a mouse. As a technology hobbyist, she has certainly kept up with what was offered by such revolutions and now she is surrounded by most of the high tech gadgets ranging from smart mobiles to the use of Bluetooth. In addition, her interest in recording her activities on day-to-day basis has resulted in her using wearable devices. Luisa Gandarias is an environmental consultant about to be retired from the company where she has been working during the last 28 years. She has combined her professional activity with the photography, which has been her main hobby during the last 30 years. As she is contemplating of retiring from a professional environment, she wants to share her legacy with her friends, family and most importantly her social peers who have cheered on her achievements and encouraged on her adventurous trips. Her dedicated work ethics and passion for photography has resulted in a lot of content commonly categorized as “User Generated Content”. Her need for assembling an automatic lifebrary, consisting of a collection of her lifeworks highlighting the source (or point of origin) of the digital media comes from her wish to share her experiences with multi-view perspective of her life, i.e. family sharing personal experiences, memories, trips etc and social peers having similar interest in environment sharing her adventurous journeys through different terrains. Firstly in this scenario the guidance of librarians and information and library science experts is essential in helping technologists become aware of the methods that humans use to construct a catalogue of Luisa’s life. Secondly the guidance of ethnographers is required to help construct an idea of the relevant social networks that may be important and significant to Luisa. Carrying out a semantic analysis of Luisa’s online social network interactions may prove a successful path in attempting to automate this scenario. Thirdly the role of artists in presenting Luisa’s personal lifebrary in novel fashions which bring her digital preservation and legacy to life is a worthwhile persuit. Use Case 2 – Personal Lifebraries for those with dementia Given the demographic shift towards older populations in the EU and that dementia become increasingly prevelant as people get older, it is worthwhile to consider this scenario as a potential benefit that may be offered in moving from lifelogging to lifebraries. Mary was healthy throughout her whole life, but unfortunately in her early sixties she has developed a memory impairment. This has been very distressing for Mary as she now very much struggles to remember events that happened as recently as the previous day. She has been embarrassed to talk to others for fear of forgetting seemingly trivial pieces of information, e.g. who she met at church 3 days ago, etc. However of late she has begun to wear the Vicon Revue, a visual lifelogging device. She now puts it on at the start of every day after letting it charge overnight when sleeping. Every evening she uploads images to her PC, and our software automatically processes this data. As Mary's Vicon Revue takes approximately 2,000 images per day, it would overwhelm herself and her husband John (who is in effect her carer) to look through all these images. Therefore we automatically summarise this data by exploiting characteristics of how the human memory operates. An example of this would be segmenting sequences of the images into distinct events/activities, e.g. when she had breakfast in the morning, when she went for a walk in the afternoon, when she was talking to Ann next door, etc. This helps herself and John more easily review what Mary has done in any given day. Indeed they particularly enjoy reviewing their Grandson's 4th birthday that took place last Friday. Naturally there were some routine/mundane events in the morning first of all, so again in exploiting an understanding of the human memory system it is important for our software to these mundane events appear smaller, while the main event of the birthday gathering is magnified so as to catch John and Mary's attention. By playing through the images of this event Mary can remember that she was talking to her friend Jack about how quickly her Grandson is growing. For certain events John can prompt Mary on what she was doing, with the SenseCam images being very beneficial. Indeed Mary wonders what other times recently she was talking to Jack, so our system provides her with the ability to retrieve other similar events to the one of talking to Jack. She is then provided with a list of other potentially relevant events, which she quite enjoys looking at as they trigger some memories. This use case represents a number of accomplished goals: 1. Gathering and storage of passively recorded lifelog data, and considering when it is ethical to record such information 2. An understanding of how the human mind works (e.g. information stored as events, associating similar events, distinctive events more strongly remembered, retrieval of events leads to a stronger encoding, etc.) so as to summarise one’s lifelog information by presenting the more memorable events 3. Development of information retrieval models (extraction of relevant image features, good fusion models, BM25/etc. retrieval models) to summarise one’s lifelog and offer multi-faceted options to retrieve such relevant data 4. Effective presentation of the most important and relevant content to Mary. This requires the guidance of dementia clinicians which then guides computing researchers develop effective “personal lifebraries” Use Case 3 – Mining a community of lifelog data and creating the appropriate lifebrary: Consider the following scenario as a motivation for the role of lifelogging. Jim decides to go on a one day holiday to Dublin city centre and he wears his lifelogging device all day long. That evening he wishes to review the lifelog images of his day. Firstly it will be important for his viewing application to automatically summarise his content in the appropriate manner and identifying those events that are more likely to be remembered as important (quite similar to that described in the previous scenario, and also understanding how self-driven goal events are strongly encoded in the human mind). While in the city centre he started talking to his friend Mark (also wearing his miniature wearable lifelogging capture device) and this was a very important event due to its uniqueness as Aiden had not talked to Mark for quite a while, especially in Dublin city. That evening Aiden wonders when he was last talking to Mark, and he uses his personal lifelog application to find previous occasions similar to this. He can see that he had last talked to John 1 month ago, while it was almost 3 months ago since he talked to Mark in O'Connell St. However, he also nostalgically notices that he was talking to his friend Mary in O'Connell St. 2 weeks ago, as this event is visually quite similar (and thus somewhat associated) to that of talking to Mark earlier today. While reviewing the event of talking to his friend Mary, he notices that he was at a football match later that evening, which is a vital memory hook. After browsing through some past memories Aiden then looks at his favourite picture from his trip today; an image of Trinity College. In this instance it is of interest to the community that a shared lifebrary is created of this network of individuals (Jim, Mark, and Mary in this case). This network may be a close-knit community, a family, or it could also be an entire region or nation. By firstly utilising the guidance of ethics experts, we can create an environment where people are aware when it is acceptable to capture lifelogged material, and also in how to share such material. Secondly in collaboration with memory experts, technologists strive to store events of personal significance, and through the guidance of ethnographers and sociologists, models can be conceived to capture and preserve material that is of significance to the relevant social groups. In addition to the drivers of the previous use case, in this scenario we show the role of these additional facets in addressing the challenge of moving from lifelogs to lifebraries: 1. A social networking analysis to find out what information other people find most interesting about an individual e.g. work colleagues are interested in one’s work achievements, training partners are only interested in sporting accomplishments, family most interested in attendance of family occasions, etc. … Here the guidance of ethnographers is essential 2. Storage and processing of one’s lifelog material so that we can then create a “lifebrary” of an individual, which then should fit into a community lifebrary. As different people can have different views of this lifebrary, we require the expertise of librarians and library information scientists to decide how decisions are made on different views. They guide our algorithms and selection methods 3. Final presentation of a personalised community lifebrary is very important. This is where our work will be guided by artists to disseminate lifebrary information of people and communities in novel and meaningful fashions.