With the escalating use of the Internet worldwide and substantially increasing impact produced by the availability of ambiguous information, the challenge to quickly identify fake news in digital media in various languages becomes more acute. In this work, we have worked on a poor resource language i.e. Urdu for detecting fake news. The latent features of the news articles are extracted from Bidirectional GRU, followed by the concatenation of average and max-pooling layers. Finally, the class

more »

... bel is predicted from the softmax layer. This work is a part of UrduFake, a shared task of FIRE-2020. An average f1 of 80.78% is achieved on the test data. The developed system achieved the fourth position in the competition.