More Like this

1. Channel Normalization in Convolutional Neural Networks avoids Vanishing Gradients

   Dai, Z.; Heckel, R. (April 2019, International Conference on Machine Learning, Workshop Deep Phenomena)

   Normalization layers are widely used in deep neural networks to stabilize training. In this paper, we consider the training of convolutional neural networks with gradient descent on a single training example. This optimization problem arises in recent approaches for solving inverse problems such as the deep image prior or the deep decoder. We show that for this setup, channel normalization, which centers and normalizes each channel individually, avoids vanishing gradients, whereas without normalization, gradients vanish which prevents efficient optimization. This effect prevails in deep single-channel linear convolutional networks, and we show that without channel normalization, gradient descent takes at least exponentially many steps to come close to an optimum. Contrary, with channel normalization, the gradients remain bounded, thus avoiding exploding gradients. 

   more » « less
2. Orthogonal Gated Recurrent Unit With Neumann-Cayley Transformation

   https://doi.org/10.1162/neco_a_01710  

   Zadorozhnyy, Vasily; Mucllari, Edison; Pospisil, Cole; Nguyen, Duc; Ye, Qiang (November 2024, Neural Computation)

   In recent years, using orthogonal matrices has been shown to be a promising approach to improving recurrent neural networks (RNNs) with training, stability, and convergence, particularly to control gradients. While gated recurrent unit (GRU) and long short-term memory (LSTM) architectures address the vanishing gradient problem by using a variety of gates and memory cells, they are still prone to the exploding gradient problem. In this work, we analyze the gradients in GRU and propose the use of orthogonal matrices to prevent exploding gradient problems and enhance long-term memory. We study where to use orthogonal matrices and propose a Neumann series–based scaled Cayley transformation for training orthogonal matrices in GRU, which we call Neumann-Cayley orthogonal GRU (NC-GRU). We present detailed experiments of our model on several synthetic and real-world tasks, which show that NC-GRU significantly outperforms GRU and several other RNNs. 

   more » « less
3. SBO-RNN: Reformulating Recurrent Neural Networks via Stochastic Bilevel Optimization

   Zhang, Z.; Yue, Y.; Wu, G.; Li, Y.; Zhang, H. (December 2021, Thirty-fifth Conference on Neural Information Processing Systems)

   In this paper we consider the training stability of recurrent neural networks (RNNs) and propose a family of RNNs, namely SBO-RNN, that can be formulated using stochastic bilevel optimization (SBO). With the help of stochastic gradient descent (SGD), we manage to convert the SBO problem into an RNN where the feedforward and backpropagation solve the lower and upper-level optimization for learning hidden states and their hyperparameters, respectively. We prove that under mild conditions there is no vanishing or exploding gradient in training SBO-RNN. Empirically we demonstrate our approach with superior performance on several benchmark datasets, with fewer parameters, less training data, and much faster convergence. Code is available at https://zhang-vislab.github.io. 

   more » « less
4. Gated recurrent units viewed through the lens of continuous time dynamical systems

   https://doi.org/10.3389/fncom.2021.678158  

   Jordan, Ian; Sokol, Piotr; Park, Il Memming (January 2021, Frontiers in computational neuroscience)

   null (Ed.)

   In recent years, the efficacy of using artificial recurrent neural networks to model cortical dynamics has been a topic of interest. Gated recurrent units (GRUs) are specialized memory elements for building these recurrent neural networks. Despite their incredible success in natural language, speech, video processing, and extracting dynamics underlying neural data, little is understood about the specific dynamics representable in a GRU network, and how these dynamics play a part in performance and generalization. As a result, it is both difficult to know a priori how successful a GRU network will perform on a given task, and also their capacity to mimic the underlying behavior of their biological counterparts. Using a continuous time analysis, we gain intuition on the inner workings of GRU networks. We restrict our presentation to low dimensions, allowing for a comprehensive visualization. We found a surprisingly rich repertoire of dynamical features that includes stable limit cycles (nonlinear oscillations), multi-stable dynamics with various topologies, and homoclinic bifurcations. At the same time GRU networks are limited in their inability to produce continuous attractors, which are hypothesized to exist in biological neural networks. We contextualize the usefulness of different kinds of observed dynamics and support our claims experimentally. 

   more » « less
5. Eigenvalue Normalized Recurrent Neural Networks for Short Term Memory

   https://doi.org/10.1609/aaai.v34i04.5831  

   Helfrich, Kyle; Ye, Qiang (June 2020, Proceedings of the AAAI Conference on Artificial Intelligence)

   Several variants of recurrent neural networks (RNNs) with orthogonal or unitary recurrent matrices have recently been developed to mitigate the vanishing/exploding gradient problem and to model long-term dependencies of sequences. However, with the eigenvalues of the recurrent matrix on the unit circle, the recurrent state retains all input information which may unnecessarily consume model capacity. In this paper, we address this issue by proposing an architecture that expands upon an orthogonal/unitary RNN with a state that is generated by a recurrent matrix with eigenvalues in the unit disc. Any input to this state dissipates in time and is replaced with new inputs, simulating short-term memory. A gradient descent algorithm is derived for learning such a recurrent matrix. The resulting method, called the Eigenvalue Normalized RNN (ENRNN), is shown to be highly competitive in several experiments. 

   more » « less