[Retracted] Defending Privacy Inference Attacks to Federated Learning for Intelligent IoT with Parameter Compression

Table 1

Comparison between Gaussian noise and parameter compression.

Defense methods		round	round	Decrease (%)
Baseline		0.9591	0.9507	0.84
Gaussian noise		0.9560	0.9440	1.20
		0.9570	0.9428	1.42
		0.9334	0.8035	14.1
Parameter compression		0.9585	0.9406	1.79
		0.9593	0.9456	1.37
		0.9565	0.9557	0.08