Learning Longer Memory in Recurrent Neural Networks
arXiv:1412.7753
Abstract
Recurrent neural network is a powerful model that learns temporal patterns in sequential data. For a long time, it was believed that recurrent networks are difficult to train using simple optimizers, such as stochastic gradient descent, due to the so-called vanishing gradient problem. In this paper, we show that learning longer term patterns in real data, such as in natural language, is perfectly possible using gradient descent. This is achieved by using a slight structural modification of the simple recurrent neural network architecture. We encourage some of the hidden units to change their state slowly by making part of the recurrent weight matrix close to identity, thus forming kind of a longer term memory. We evaluate our model in language modeling experiments, where we obtain similar performance to the much more complex Long Short Term Memory (LSTM) networks (Hochreiter & Schmidhuber, 1997).
References in corpus (2)
Cited by in corpus (14)
- A Simple Way to Initialize Recurrent Networks of Rectified Linear Units
- Inferring Algorithmic Patterns with Stack-Augmented Recurrent Nets
- Compositional Vector Space Models for Knowledge Base Completion
- DR-RNN: A deep residual recurrent neural network for model reduction
- Gated Recurrent Neural Tensor Network
- A New Unbiased and Efficient Class of LSH-Based Samplers and Estimators for Partition Function Computation in Log-Linear Models
- A recurrent neural network without chaos
- Unsupervised Learning from Continuous Video in a Scalable Predictive Recurrent Network
- A Linear Dynamical System Model for Text
- A Simple LSTM model for Transition-based Dependency Parsing
- Simplified Long Short-term Memory Recurrent Neural Networks: part III
- An Attention Mechanism for Answer Selection Using a Combined Global and Local View
- Learning long-term dependencies for action recognition with a biologically-inspired deep network
- Self-organized Hierarchical Softmax