From the title, I wanna know if there's such a thing as RNN-LSTM. I know that they are their own thing but I've yet to know if there's such a "combination".

For context, I was reading a research paper and there was this passage (source below): RNN-LSTM (Hochreiter and Schmidhuber 1997): The Recurrent Neural Network using LSTM cell. We combine an LSTM layer and a dense layer to perform multivariate time series forecasting.

Source: Cheng, J., Huang, K., & Zheng, Z. (2020). Towards Better Forecasting by Fusing Near and Distant Future Visions. Proceedings of the AAAI Conference on Artificial Intelligence, 34(04), 3593–3600. https://doi.org/10.1609/aaai.v34i04.5766


1 Answer 1


RNN is a general term for Neural network architecture having recurrent connections. This architecture is built from a single entity called "cell". These so called cells are used repeatedly and connected in recursive fashion.

LSTM is one of these type of cells. If you build a RNN using LSTM cells, it becomes LSTM RNN. So basically as long as you use LSTM as a basic entity for building a RNN, it will be called LSTM RNN.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.