Unbiased Online Recurrent Optimization
This addresses the issue of biased gradient estimates in online recurrent learning, which can cause divergence, making it significant for researchers and practitioners in machine learning, though it is an incremental improvement over prior methods like NoBackTrack.
The paper tackles the problem of online learning for recurrent networks by introducing the Unbiased Online Recurrent Optimization (UORO) algorithm, which provides unbiased gradient estimates and converges on synthetic tasks where truncated Backpropagation Through Time (BPTT) diverges, such as when parameters have conflicting short-term and long-term influences.
The novel Unbiased Online Recurrent Optimization (UORO) algorithm allows for online learning of general recurrent computational graphs such as recurrent network models. It works in a streaming fashion and avoids backtracking through past activations and inputs. UORO is computationally as costly as Truncated Backpropagation Through Time (truncated BPTT), a widespread algorithm for online learning of recurrent networks. UORO is a modification of NoBackTrack that bypasses the need for model sparsity and makes implementation easy in current deep learning frameworks, even for complex models. Like NoBackTrack, UORO provides unbiased gradient estimates; unbiasedness is the core hypothesis in stochastic gradient descent theory, without which convergence to a local optimum is not guaranteed. On the contrary, truncated BPTT does not provide this property, leading to possible divergence. On synthetic tasks where truncated BPTT is shown to diverge, UORO converges. For instance, when a parameter has a positive short-term but negative long-term influence, truncated BPTT diverges unless the truncation span is very significantly longer than the intrinsic temporal range of the interactions, while UORO performs well thanks to the unbiasedness of its gradients.