mlsanityData to 14 Oct 2026

Rethinking Recurrent Neural Networks for Time Series Forecasting: A Reinforced Recurrent Encoder with Prediction-Oriented Proximal Policy Optimization

arXiv 2601.03683 ↗First posted 7 Jan 2026cs.LG cs.NEcited by 0

On the map: Deep reinforcement learning methods → Deep RL algorithms and representations

Abstract (arXiv)

Time series forecasting plays a crucial role in contemporary engineering information systems for supporting decision-making across various industries, where Recurrent Neural Networks (RNNs) have been widely adopted due to their capability in modeling sequential data. Conventional RNN-based predictors adopt an encoder-only strategy with sliding historical windows as inputs to forecast future values. However, this approach treats all time steps and hidden states equally without considering their distinct contributions to forecasting, leading to suboptimal performance. To address this limitation, we propose a novel Reinforced Recurrent Encoder with Prediction-oriented Proximal Policy Optimization, RRE-PPO4Pred, which significantly improves time series modeling capacity and forecasting accuracy of the RNN models. The core innovations of this method are: (1) A novel Reinforced Recurrent Encoder (RRE) framework that enhances RNNs by formulating their internal adaptation as a Markov Decision Process, creating a unified decision environment capable of learning input feature selection, hidden skip connection, and output target selection; (2) An improved Prediction-oriented Proximal Policy Optimization algorithm, termed PPO4Pred, which is equipped with a Transformer-based agent for temporal reasoning and develops a dynamic transition sampling strategy to enhance sampling efficiency; (3) A co-evolutionary optimization paradigm to facilitate the learning of the RNN predictor and the policy agent, providing adaptive and interactive time series modeling. Comprehensive evaluations on five real-world datasets indicate that our method consistently outperforms existing baselines, and attains accuracy better than state-of-the-art Transformer models, thus providing an advanced time series predictor in engineering informatics.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Deep RL algorithms and representations, one of 15 directions in Deep reinforcement learning methods. Builds on 29 of its 60 references that are in the map. First posted on arXiv 7 Jan 2026.

Builds on

29 of 60 references.

References · 29

  1. 2017Proximal Policy Optimization Algorithmscited by 15,215
  2. 2014Learning Phrase Representations using RNN Encoder-Decoder for Statistical Machine Translationcited by 4,057
  3. 2014Sequence to Sequence Learning with Neural Networkscited by 2,986
  4. 2022A Time Series is Worth 64 Words: Long-term Forecasting with Transformerscited by 2,039
  5. 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 2,026
  6. 2019Federated Learning: Challenges, Methods, and Future Directionscited by 1,538
  7. 2017Get To The Point: Summarization with Pointer-Generator Networkscited by 1,292
  8. 2023iTransformer: Inverted Transformers Are Effective for Time Series Forecastingcited by 931
  9. 2019N-BEATS: Neural basis expansion analysis for interpretable time series forecastingcited by 753
  10. 2015Sequence Level Training with Recurrent Neural Networkscited by 576
  11. 2018Fundamentals of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) Networkcited by 420
  12. 2019Recurrent Neural Networks for Time Series Forecasting: Current Status and Future Directionscited by 172
17 more
  1. 2018Independently Recurrent Neural Network (IndRNN): Building A Longer and Deeper RNNcited by 143
  2. 2016Phased LSTM: Accelerating Recurrent Network Training for Long or Event-based Sequencescited by 99
  3. 2017Dilated Recurrent Neural Networkscited by 83
  4. 2014A Clockwork RNNcited by 72
  5. 2016Zoneout: Regularizing RNNs by Randomly Preserving Hidden Activationscited by 54
  6. 2017Skip RNN: Learning to Skip State Updates in Recurrent Neural Networkscited by 49
  7. 2016Minimal Gated Unit for Recurrent Neural Networkscited by 40
  8. 2023A Survey on Efficient Training of Transformerscited by 39
  9. 2024Unlocking the Power of LSTM for Long Term Time Series Forecastingcited by 29
  10. 2020AutoFS: Automated Feature Selection via Diversity-aware Interactive Reinforcement Learningcited by 24
  11. 2017Learning to Skim Textcited by 18
  12. 2016Architectural Complexity Measures of Recurrent Neural Networkscited by 17
  13. 2019Neural Speed Reading with Structural-Jump-LSTMcited by 17
  14. 2024xLSTMTime : Long-term Time Series Forecasting With xLSTMcited by 14
  15. 2018Long Short-Term Memory with Dynamic Skip Connectionscited by 4
  16. 2019Leap-LSTM: Enhancing Long Short-Term Memory for Text Categorizationcited by 2
  17. 2023Feature Selection with Annealing for Forecasting Financial Time Seriescited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2021Deep Probabilistic Time Series Forecasting using Augmented Recurrent Input for Dynamic Systems
  2. 2025IRNN: Innovation-driven Recurrent Neural Network for Time-Series Data Modeling and Prediction
  3. 2021Stochastic Recurrent Neural Network for Multistep Time Series Forecasting
  4. 2024A Policy Gradient-Based Sequence-to-Sequence Method for Time Series Prediction
  5. 2021Learning Interpretable Deep State Space Model for Probabilistic Time Series Forecasting
  6. 2025LangTime: A Language-Guided Unified Model for Time Series Forecasting with Proximal Policy Optimization
  7. 2025Small but Mighty: Enhancing Time Series Forecasting with Lightweight LLMs
  8. 2024Recurrent Interpolants for Probabilistic Time Series Prediction