Summary of Were Rnns All We Needed?, by Leo Feng et al.
Were RNNs All We Needed?
by Leo Feng, Frederick Tung, Mohamed Osama Ahmed, Yoshua Bengio, Hossein Hajimirsadeghi
First submitted to arxiv on: 2 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The introduction of Transformers in 2017 revolutionized deep learning, but its scalability limitations have sparked interest in novel recurrent models. This paper revisits sequence modelling from a historical perspective, focusing on Recurrent Neural Networks (RNNs), which dominated the field before the rise of Transformers. We examine LSTMs and GRUs, simplifying them to derive minimal versions (minLSTMs and minGRUs) that use fewer parameters, are fully parallelizable during training, and achieve surprisingly competitive performance on various tasks, rivalling recent models including Transformers. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Transformers changed the game in 2017, but they have limitations. This paper looks back at old models called RNNs that were popular before Transformers came along. It takes a close look at two special kinds of RNNs: LSTMs and GRUs. By making these models simpler, we can make new versions (minLSTMs and minGRUs) that use less information, work faster during training, and do surprisingly well on different tasks. |
Keywords
» Artificial intelligence » Deep learning
