Our approach is constructed to improve on a couple of aspects; preprocessing with an emphasis on humor sense detection, using embeddings from state-of-the-art language model(Elmo), and ensembling the results came up with using machine learning model Na ̈ıve Bayes(NB) with a deep learning pre-trained models. Elmo-NB participation has scored (0.5642) on the competition leader board, where results were measured by Root Mean Squared Error (RMSE).