Multimodal data

Here are overall classification and regression problems results on multimodal datasets across state-of-the-art AutoML frameworks using AutoML Multimodal Benchmark test suite:

	framework	AutoGluon	FEDOT
Dataset name	Metric name
prod	accuracy	0.895	0.897
salary	accuracy	0.528	0.402
airbnb	accuracy	0.466	0.434
channel	accuracy	0.550	0.542
wine	accuracy	0.842	0.784
imdb	auc	0.874	0.872
fake	auc	0.968	0.958
kick	auc	0.775	0.752
jigsaw	auc	0.915	0.998
qaa	r2	0.383	0.504
qaq	r2	0.426	0.554
book	r2	0.600	0.994
jc	r2	0.612	0.619
cloth	r2	0.654	0.733
ae	r2	0.979	0.974
pop	r2	0.020	0.020
house	r2	0.943	0.928
mercari	r2	0.569	0.520

The results are obtained using sever based on Xeon Cascadelake (2900MHz) with 12 cores and 24GB memory for experiments with the local infrastructure.