Transcript Accessing Multimodal Meeting Data: Systems, Problems and

Why Memories for Life?
Steve Whittaker, University of Sheffield
Funded by EU Marie Curie Grant.
Memoir: Digital tools for sharing and managing
memories
7/17/2015
1
4 meanings of digital memory (DM)
 Record of everything I ever experienced
 Collection of my personal digital stuff – pictures,
documents, music
 Collection of stuff to do with me (includes
information generated by others)
 Collective record of the past
17 July, 2015
Steve Whittaker, Sheffield University
2
Who’s interested?
 Neuroscientists – brain models of human memory –
function and mechanism
 Cognitive Scientists – behavioural models of
memory function and mechanism
 Information/Computer Scientists – representation &
retrieval of personal information
 Writers/Social & Cultural Historians – representation
of personal and collective pasts
 HCI – function and design of memory prostheses
17 July, 2015
Steve Whittaker, Sheffield University
3
Applications: What could DM
do for us?
 Individual recall/reflection
Absent-mindedness aids – where are my keys?
Where’s that document?
Learning from prior experience –
technopsychoanalysis
 Sharing experience with others
Identity and relationships - digital albums for
family/friends
Community ties and histories
 Prospective memory - reminding
17 July, 2015
Steve Whittaker, Sheffield University
4
Most important benefit?
 Sharing experiences with others??
 Social sharing among family and friends (finesses
community problems of critical mass)
Digital pictures too passive, need to:
Share with others
Add narrative
Add event structures
17 July, 2015
Steve Whittaker, Sheffield University
5
Will technology support user goals?
 Various failures
Long-term personal archives often unsuccessful
(email, files, web bookmarks, paper)
2485 emails, 2200 e-documents, 113 web
bookmarks, 62kg paper
Organisational memory, meeting retrieval
Many proof of concept prototypes – little
evidence of utility
 ->Successful access to long-term information not a trivial problem
17 July, 2015
Steve Whittaker, Sheffield University
6
Evidence from various disciplines?
 Psychology, Sociology
Experiential sharing important social function, sense
of identity
Problems with prospective memory
Cueing processes and representation of the past
 Neuroscientists/Cognitive Scientists – unclear how their
models can inform design
 Information/Computer Scientists – access and retrieval
of complex personal data – abstraction is crucial
 Writers/Social/Cultural Historians – representation of
personal and collective pasts can inform design
 HCI – function and design of effective DMs
17 July, 2015
Steve Whittaker, Sheffield University
7
Setting a research agenda
 Beyond current model of ‘saving everything’
 Support situations where human memory is poor
 Work in synergy with human memory rather than
replacing it
Digital information can trigger human memories
We remember certain things well unaided
When do people use DM rather than their own memory?
 Support forgetting
 Exploit other disciplines’ accounts of function of memory
– history/arts
 Apply what we know about representation and retrieval
from CS/IS
17 July, 2015
Steve Whittaker, Sheffield University
8