BackGround Eraser Mod ATG
In a proof-of-concept study, education and artificial intelligence researchers have used machine-learning models to estimate how long museum audiences will engage with an exhibit given by an individual. Improving user engagement with informal learning tools opens the door to discovery for a host of new work.
“Education is an important part of most museums of the mission,” says study study co-author Jonathan Rowe and a research scientist at North Carolina State University’s Center for Educational Informatics (CEI). “The amount of time people spend in the exhibition is used as a proxy for engagement and helps us assess the quality of learning experiences in a museum setting. It’s not like school – you can’t make visitors to take exams.
“If we can determine how much time people will spend at the exhibition, or when a exhibit starts to lose its focus, we can use that information to develop and implement adaptive exhibits that visitors feel Responds to user behavior for living, ”says Andrew, Emerson, the study’s first author and a Ph.D. Students in NC State.
“We can also give relevant data to museum staff about what is working and what people are responding to,” says Rowe. “It can help them allocate personnel or other resources to shape the museum experience, depending on which visitors are on the floor at any given time.”
To determine how machine-learning programs might be able to predict a user’s interaction time, the researchers closely monitored 85 museum visitors as they engaged with an interactive exhibition on environmental science. Specifically, the researchers collected data on the facial expressions, posture of the study participants, where they looked on the display screen and which parts of the screen they touched.
The data were fed into five different machine-learning models to ascertain that the most accurate predictions were made by combining data and models.
“We found that a particular machine-learning method called ‘random forest’ worked quite well, even using only posture and facial expression data,” says Emerson.
Researchers also found that the model worked better at interacting with people’s performance, as it gave them more data to work with. For example, a prediction made after a few minutes would be more accurate than a prediction made after 30 seconds. For reference, user interaction with the display lasted 12 minutes.
“We are excited about this,” says Rowe, as it paves the way for the study of new perspectives on how visitors learn in museums. “Ultimately, we want to use technology to make learning more effective and more engaging.”
“Education is an important part of most museums of the mission,” says study study co-author Jonathan Rowe and a research scientist at North Carolina State University’s Center for Educational Informatics (CEI). “The amount of time people spend in the exhibition is used as a proxy for engagement and helps us assess the quality of learning experiences in a museum setting. It’s not like school – you can’t make visitors to take exams.
“If we can determine how much time people will spend at the exhibition, or when a exhibit starts to lose its focus, we can use that information to develop and implement adaptive exhibits that visitors feel Responds to user behavior for living, ”says Andrew, Emerson, the study’s first author and a Ph.D. Students in NC State.
“We can also give relevant data to museum staff about what is working and what people are responding to,” says Rowe. “It can help them allocate personnel or other resources to shape the museum experience, depending on which visitors are on the floor at any given time.”
To determine how machine-learning programs might be able to predict a user’s interaction time, the researchers closely monitored 85 museum visitors as they engaged with an interactive exhibition on environmental science. Specifically, the researchers collected data on the facial expressions, posture of the study participants, where they looked on the display screen and which parts of the screen they touched.
The data were fed into five different machine-learning models to ascertain that the most accurate predictions were made by combining data and models.
“We found that a particular machine-learning method called ‘random forest’ worked quite well, even using only posture and facial expression data,” says Emerson.
Researchers also found that the model worked better at interacting with people’s performance, as it gave them more data to work with. For example, a prediction made after a few minutes would be more accurate than a prediction made after 30 seconds.