Table 2.

Examples of pre-trained DNN models and DNN architectures for quantifying naturalistic behavior

Modality	Application	Citation
Facial Expression	Emotion Labelling	(Cheong et al., 2021)
		(Wasi et al., 2023)
		(Xue et al., 2022)
		(Wen et al., 2023)
		(Vo et al., 2020)
		(Wang, Peng, et al., 2020)
	Action Unit Annotation	(Cheong et al., 2023)
		(Savchenko, 2022)
		(Luo et al., 2022)
		(Pumarola et al., 2018a)
Body Pose	Pose Estimation	(Kocabas et al., 2021)
		(Xu, Zhang, et al., 2022)
		(Liu et al., 2022)
		(Cai et al., 2020)
		(Sun et al., 2019)
		(Xiao et al., 2018)
Speech	Speech Separation	(Lutati et al., 2023)
		(Lutati et al., 2022b)
		(Subakan et al., 2021)
		(Luo et al., 2020)
		(Luo & Mesgarani, 2019)
	Speech-to-Text	(Bain et al., 2023)
		(Gállego et al., 2021)
		(Conneau et al., 2021)
		(Wang, Tang, et al., 2020)
		(Park et al., 2019)
Text	Text Embedding	(He et al., 2021)
		(Gao et al., 2021)
		(Lan et al., 2020)
		(Reimers & Gurevych, 2019)
		(Cer et al., 2018)
	Sentiment Analysis	(Heinsen, 2022)
		(Sun et al., 2020)
		(Heinsen, 2020)
		(Peters et al., 2018)