no code implementations • WS 2020 • Kang Xue, Victoria Yaneva, Christopher Runyon, Peter Baldwin
The results indicate that, for our sample, transfer learning can improve the prediction of item difficulty when response time is used as an auxiliary task but not the other way around.