Open Intent Detection
Open Intent Detection is a crucial task in the field of natural language processing, aiming to identify known n types of intents and detect unclassified open intents. This task enhances the intelligence level of human-computer interaction systems by accurately capturing users' real needs, improves user experience, and has significant application value in scenarios such as intelligent customer service and virtual assistants.
ATIS (25% known)
ATIS (50% known)
BANKING-77 (50% known)
BANKING-77 (75% known)
Metric learning + Adaptive Decision Boundary
Banking_CG
BANKING77 (25%known)
ADB
OOS(25%known)
OOS(50%known)
OOS(75%known)
OOS_CG
SNIPS (25% known)
LMCL
SNIPS (50% known)
SNIPS (75% known)
StackOverFlow(25%known)
StackOverFlow(50%known)
StackOverFlow(75%known)
ADB
StackOverflow_CG
DA-ADB