Research map / Architectures and efficiency / RNNs and normalization
On Multiplicative Integration with Recurrent Neural Networks
Yuhuai Wu, Saizheng Zhang, Ying Zhang, Yoshua Bengio, Ruslan Salakhutdinov
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Replace additive integration with Multiplicative Integration | RNN · Penn-Treebank | BPC: 1.42 -> 1.39 | Improved |
| Replace additive integration with Multiplicative Integration | LSTM · text8 | BPC: $1.51$ -> 1.44 | Improved |
| Replace additive integration with Multiplicative Integration | LSTM · HutterWikipedia | BPC: 1.47 -> 1.44 | Improved |
| Replace additive integration with Multiplicative Integration | LSTM+CTC+WFST · WSJ eval92 | CER: 6.5 -> 6.0 | Improved |
| Replace additive integration with Multiplicative Integration | LSTM+CTC+WFST · WSJ eval92 | WER: 8.7 -> 8.2 | Improved |
| Apply MI in the encoder GRU | Skip-Thought · SICK semantic relatedness | MSE: 0.2946 -> 0.2679 | Improved |
| Apply MI in the encoder GRU | Skip-Thought · Microsoft Research Paraphrase Corpus | Accuracy: 74.0 -> 74.0 | No clear effect |
| Apply MI in the encoder GRU | Skip-Thought · Microsoft Research Paraphrase Corpus | F1: 81.9 -> 82.1 | Improved |
| Apply MI in the encoder GRU | Skip-Thought · MR sentiment | Accuracy: 75.9 -> 77.9 | Improved |
| Apply MI in the encoder GRU | Skip-Thought · CR product reviews | Accuracy: 80.1 -> 82.3 | Improved |
| Apply MI in the encoder GRU | Skip-Thought · SUBJ classification | Accuracy: 93.0 -> 93.3 | Improved |
| Apply MI in the encoder GRU | Skip-Thought · MPQA polarity | Accuracy: 87.0 -> 88.1 | Improved |
Closest papers by meaning
- Multiplicative Interactions and Where to Find Them (ICLR 2020 · similarity 0.87)
- Generalized Tensor Models for Recurrent Neural Networks (ICLR 2019 · similarity 0.87)
- Architectural Complexity Measures of Recurrent Neural Networks (NeurIPS 2016 · similarity 0.86)
- Framing RNN as a kernel method: A neural ODE approach (NeurIPS 2021 · similarity 0.86)
- RNNs of RNNs: Recursive Construction of Stable Assemblies of Recurrent Neural Networks (NeurIPS 2022 · similarity 0.86)
- Recurrent neural networks: vanishing and exploding gradients are not the end of the story (NeurIPS 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
