11 - Interactive intent modeling: Information discovery beyond search -
Ruotsalo et al. (2015)
https://dl.acm.org/doi/pdf/10.1145/2656334
Er wordt in dit artikel een alternatief gepresenteerd voor zoekmachines zoals die van Google, die
niet goed zijn in exploratory search. Het alternatief combineert interactive intent modeling en
visuele gebruikersinterfaces om gebruikers stapsgewijs naar relevante informatie te sturen.
• Hoe werkt het voorgestelde systeem?
• Welke principes liggen daar ten grondslag aan?
• Wat zijn de grote voordelen van het systeem?
Key insights:
- current search engines offer limited assistance in complex search tasks; users are distracted
by having to focus their cognitive effort on finding navigation cues rather than on learning and
selecting relevant information
- interactive intent modeling enhances human information exploration through computational
modeling (visualized for interaction), helping users search and explore via user interfaces that
are highly functional but not cluttered or distracting
- interactive intent modeling can improve task-level information-seeking performance a lot
- combining intent modeling and visual user interfaces can help users discover novel information
and improve their information-exploration performance
- search engines are not good at supporting complex information-exploration and discovery
tasks that go beyond simple keyword queries
- vocabulary mismatch problem (VMP): refers to human communication behavior in which the
humans writing the documents to be retrieved and the humans searching for them are likely to
use very different vocabularies to encode and decode their intended meaning
- interactive intent modeling: addresses the VMP by giving users potential intents to explore,
visualizing them as directions in the information space around the user’s present position, and
allowing interaction to improve estimates of the user’s search intents, based on:
→ visualization: visualizing the current search intent and directions in the information
space
→ adaptation: interactive adaptation of the intent model, balancing exploration of the
information space and exploitation of user feedback
→ the intent model must be able to rigorously handle uncertainty due to limited, possibly
suboptimal, user feedback
→ the model is only manageable for the user if the system is able to predict a sufficient subset
of the potentially relevant information
- the user is able to provide feedback for the intent model → system improves intent estimates
on subsequent iterations, retrieve and rank data, and update the visualization of directions in
the information space
- the model estimates relevance related to potential search intent and uncertainty related to
these estimates based on user feedback
Ruotsalo et al. (2015)
https://dl.acm.org/doi/pdf/10.1145/2656334
Er wordt in dit artikel een alternatief gepresenteerd voor zoekmachines zoals die van Google, die
niet goed zijn in exploratory search. Het alternatief combineert interactive intent modeling en
visuele gebruikersinterfaces om gebruikers stapsgewijs naar relevante informatie te sturen.
• Hoe werkt het voorgestelde systeem?
• Welke principes liggen daar ten grondslag aan?
• Wat zijn de grote voordelen van het systeem?
Key insights:
- current search engines offer limited assistance in complex search tasks; users are distracted
by having to focus their cognitive effort on finding navigation cues rather than on learning and
selecting relevant information
- interactive intent modeling enhances human information exploration through computational
modeling (visualized for interaction), helping users search and explore via user interfaces that
are highly functional but not cluttered or distracting
- interactive intent modeling can improve task-level information-seeking performance a lot
- combining intent modeling and visual user interfaces can help users discover novel information
and improve their information-exploration performance
- search engines are not good at supporting complex information-exploration and discovery
tasks that go beyond simple keyword queries
- vocabulary mismatch problem (VMP): refers to human communication behavior in which the
humans writing the documents to be retrieved and the humans searching for them are likely to
use very different vocabularies to encode and decode their intended meaning
- interactive intent modeling: addresses the VMP by giving users potential intents to explore,
visualizing them as directions in the information space around the user’s present position, and
allowing interaction to improve estimates of the user’s search intents, based on:
→ visualization: visualizing the current search intent and directions in the information
space
→ adaptation: interactive adaptation of the intent model, balancing exploration of the
information space and exploitation of user feedback
→ the intent model must be able to rigorously handle uncertainty due to limited, possibly
suboptimal, user feedback
→ the model is only manageable for the user if the system is able to predict a sufficient subset
of the potentially relevant information
- the user is able to provide feedback for the intent model → system improves intent estimates
on subsequent iterations, retrieve and rank data, and update the visualization of directions in
the information space
- the model estimates relevance related to potential search intent and uncertainty related to
these estimates based on user feedback
