Sequence to Sequence Learning with Neural Networks

Seq2Seq is the clean predecessor to attention and Transformers. It is a good beginner paper because the encoder-decoder contract is easy to visualize and still appears in translation, summarization, and agents.

Reading focus: How one network encodes an input sequence and another decodes an output. Why fixed-size sequence representations created pressure for attention. How teacher forcing and beam search fit early neural translation.

NeurIPS 2014. Sutskever et al.. 35 min read, very easy difficulty.