Numerous methods have been developed to monitor the spread of negativity in modern years by eliminating vulgar, offensive, and fierce comments from social media platforms. However, there are relatively lesser amounts of study that converges on embracing positivity, reinforcing supportive and reassuring content in online forums. Consequently, we propose creating an English-Kannda Hope speech dataset, KanHope and comparing several experiments to provide benchmarking for the dataset. The dataset consists of 6,176 user-generated comments in code mixed Kannada crawled from YouTube and manually labelled as bearing hope speech or not-hope speech. In addition, we introduce DC-BERT4HOPE, a dual-channel model that uses the English translation of KanHopeEDI for additional training to promote hope speech detection. The approach achieves a weighted F1-score of 0.756, bettering other models. Henceforth, KanHope aims to instigate research in Kannada while broadly promoting researchers to take a pragmatic approach towards online content that encourages, positive, and supportive.

@article{hande-etal-kanhope, title = "Hope Speech detection in under-resourced Kannada language", author = "Hande, Adeep and Priyadharshini, Ruba and Sampath, Anbukkarasi and Thamburaj, Kingston Pal and Chandran, Prabakaran and Chakravarthi, Bharathi Raja ", journal={SN Computer Science}, publisher={Springer} }