Nemesis: Neural Mean Teacher Learning-Based Emotion-Centric Speaker

Image captioning is the multi-modal task of automatically describing a digital image based on its contents and their semantic relationship. This research area has gained increasing popularity over past few years; however, most previous studies have been focused purely objective content-based descriptions scenes. In this study, efforts made to generate more engaging captions by leveraging human-like emotional responses. To achieve task, mean teacher learning-based method applied recently introduced ArtEmis dataset. first large-scale dataset for emotion-centric captioning, containing 455K 80K artworks from WikiArt. includes self-distillation relationship between memory-augmented language models with meshed connectivity. These are trained in cross-entropy phase then fine-tuned self-critical sequence training phase. According various popular natural processing metrics, such as BLEU, METEOR, ROUGE-L, CIDEr, our proposed model obtained new state art ArtEmis.