Researchers at Vanderbuilt University have conducted experiments which explore ways of filtering information.
Researchers at Vanderbilt University have suggested a system of value assignment and selective forgetting is one way to cope with information overload.
The aspects of human memory considered for their research include time-based decay (memories disappearing over time), and interference (multiple memories vying for the same information whether they’re correct or incorrect). To test out their theory, they created algorithms for robots to use different methods of gathering and retaining information.
The best performing software, ActSimple, assigns value to gathered data based on how frequently it’s used and its relation to other data pieces. As a robot roamed through a virtual environment seeking Wi-Fi signals, it assigned values from 1-100 considering strength of signal and different levels of noise. Then it would try to create from memory a map of the signals. Other algorithms used to create these maps either considered all the data gathered (including errors) or disregarded the oldest data points. On average, ActSimple created the most accurate maps.
We’ve already seen the inverse application of value assignments in programs such as BumpTop, where the most frequently used and opened documents become bigger in relation to the other documents on the desktop.
[ via Technology Review ]