A survey of transfer learning

Journal of Big Data

Table 2 Homogeneous transfer learning approaches surveyed in “Homogeneous transfer learning” section listing different characteristics of each approach

Approach	Transfer category	Source data	Target data	Multiple sources	Generic solution	Negative transfer
CP-MDA [14]	Parameter	Labeled	Limited labels	✓	✓
2SW-MDA [14]	Instance	Labeled	Unlabeled	✓	✓
FAM [22]	Asymmetric feature	Labeled	Limited labels	✓	✓
DTMKL [27]	Asymmetric feature	Labeled	Unlabeled		✓
JDA [69]	Asymmetric feature	Labeled	Unlabeled		✓
ARTL [68]	Asymmetric feature	Labeled	Unlabeled		✓
TCA [87]	Symmetric feature	Labeled	Unlabeled		✓
SFA [83]	Symmetric feature	Labeled	Limited labels	✓	✓
SDA [41]	Symmetric feature	Labeled	Unlabeled		✓
GFK [42]	Symmetric feature	Labeled	Unlabeled		✓	✓
DCP [106]	Symmetric feature	Labeled	Unlabeled		✓
TCNN [81]	Symmetric feature	Labeled	Limited labels		✓
MMKT [114]	Parameter	Labeled	Limited labels	✓	✓	✓
DSM [28]	Parameter	Labeled	Unlabeled	✓		✓
MsTrAdaBoost [138]	Instance	Labeled	Limited labels	✓	✓	✓
TaskTrAdaBoost [138]	Parameter	Labeled	Limited labels	✓	✓	✓
RAP [62]	Relational	Labeled	Unlabeled
SSFE [132]	Hybrid (instance and feature)	Labeled	Limited labels