Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes
A Bayesian Approach to Recurrence in Neural Networks
Philip N. Garner, Sibo Tong
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 |
|---|---|---|---|
| use unit-wise Bayesian recurrent unit instead of unidirectional GRU | Uni-GRU · TIMIT | phone error rate: p<0.001 | Improved |
| use layer-wise BRU instead of bidirectional GRU (no explicit backward pass) | Bi-GRU · TIMIT | phone error rate: p = 0.230 | No clear effect |
| make LBRU explicitly bidirectional (Bi-LBRU) versus Bi-GRU and Bi-LSTM | Bi-GRU · TIMIT | phone error rate: p=0.43 | No clear effect |
| use UBRU or LBRU instead of unidirectional GRU | Uni-GRU · WSJ eval92 | word error rate: p=0.19 | No clear effect |
| use LBRU instead of bidirectional GRU | Bi-GRU · WSJ eval92 | word error rate: p=0.21 | No clear effect |
| use layer-wise instead of unit-wise BRU | UBRU · AMI-IHM | word error rate: p<0.001 | Improved |
| make LBRU explicitly bidirectional versus Bi-GRU and Bi-LSTM | Bi-GRU · AMI-IHM | word error rate: p<0.001 | Improved |
| use Bi-LSTM (extra gate) instead of Bi-GRU | Bi-GRU · AMI-IHM | word error rate | No clear effect |
Closest papers by meaning
- Linear approximations to HMM filtering (NeurIPS 2026 · similarity 0.88)
- Stable Recurrent Models (ICLR 2019 · similarity 0.88)
- Expressive probabilistic sampling in recurrent neural networks (NeurIPS 2023 · similarity 0.87)
- Recurrent neural networks: vanishing and exploding gradients are not the end of the story (NeurIPS 2024 · similarity 0.86)
- Inverse Approximation Theory for Nonlinear Recurrent Neural Networks (ICLR 2024 · similarity 0.86)
- Learning dynamics in linear recurrent neural networks (ICML 2025 · 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.
