Vita Lanfranchi

1
User Interfaces and Interaction for CI

• Vita Lanfranchi
• University of Sheffield
• vita_at_dcs.shef.ac.uk

2
Outline

• Project aims
• Personal Intelligence vs. Collective Intelligence
• Requirements
• Interaction models
• Upload
• Access
• Interfaces mock ups

3
Aims

• WeKnowIt focuses its research effort on
  developing methodologies, technologies and tools
  to exploit Collective Intelligence
• The Personal Intelligence layer will enable users
  to efficiently
• upload and retrieve intelligence
• using a range of devices
• from mobile phones, to PDAs and personal
  computers
• Once content is submitted, a series of processing
  and analysis procedures take place
• Knowledge is created
• User and context modelling paradigms will be
  employed to enable personalised access to the
  content and knowledge, available from the
  proposed applications

4
Personal Intelligence

• Personalised upload to and retrieval from
  Collective Intelligence
• emerging from the analysis of multimedia content
• supporting access to different type of users and
  organisations
• using multiple devices.

5
User studies

• User Studies were performed
• Questionnaires
• Interviews
• to better understand WeKnowIt use cases
• Emergency Response
• Consumer Social Study
• This lead to the definition of
• Requirements
• Scenarios
• Interaction models
• Mock-ups

6
Functional Requirements

• Interaction
• Support for multiple users and multiple
  visualisations
• Support for Multimodal Interaction
• Support for Information Visualisation for Data
  Analysis
• Personal Intelligence Management
• Support for Personal Intelligence Upload
• Tagging/Annotation functionality
• Feedback
• Support for Personal Intelligence Access
• Search
• Browse
• Workflow Management
• Automatic Services
• System Suggestions
• User and context profiling

7
Non-functional Requirements

• Trust
• Privacy
• Accessibility
• Reliability
• Ease of use
• Documentation/Help
• Resilience
• Personalisation
• Latency
• Self-declaring design

8
User-centric interaction model

• A large number of entities are interoperating
• the actual events
• the mechanisms of capturing event information
  (e.g. image and video cameras, geolocation
  possibilities, automatic recognition of voice
  annotations, etc.)
• the content types
• the types of users
• the devices
• the network

9
Multimodal Interaction for CI

• WeKnowIt applications should have interfaces
  accessible by different devices.
• ability to present the content and
  functionalities using different modalities.
• The modalities of presenting the content will
  depend on different variables
• Device used
• Context of interaction (i.e. during an emergency
  or on a beach during a holiday)
• User preferences

10
Personal Intelligence Upload

• Personal Intelligence Upload does not simply
  refer to the physical upload of information on a
  server
• but to the upload of knowledge into a Knowledge
  Base.
• an upload action is made whenever the user
  interacts with the Knowledge Base to add new
  content
• either in the form of new data or metadata

11
Upload Interaction Model

• Upload/Create new raw data
• Pictures, videos, text
• This raw data may be already enriched with
  metadata like EXIF data or geolocation
  coordinates
• Annotate new/already existing data
• Tag new/already existing data
• user must be able to tag information, edit other
  people tags and remove tags and classifications.
• Users are still free to upload content without
  annotation
• The tagging process should be performed through
  different modalities

12
Upload Mock-ups
13
Personal Intelligence Access

• Functionalities to access a heterogeneous and
  large-scale Knowledge Base
• The access can be performed
• either using push or pull modalities
• accordingly to the user preferences
• and the context of interaction.
• The system should provide support for different
  tasks such as
• browsing
• and searching

14
Access Interaction model

• The user interaction is primarily effected by the
  ability of the user to express their information
  needs
• The user may require help in formulating that
  need into a query
• Where the user does not have such a specific goal
  the system can provide the ability to browse the
  Information Space
• Once the user is satisfied with the query they
  can use it to search the system KB, and receive
  the results matching the query
• If the results do not satisfy the Information
  Need the user may try to reformulate the query
• again the system should aid in this process of
  query refinement

15
Access mock-ups