Learning Associative Inference Using Fast Weight Memory
This addresses the challenge of rapid associative learning in AI systems, particularly for reasoning and adaptation in novel contexts, though it appears incremental as it builds on existing LSTM architectures.
The paper tackles the problem of enabling neural networks to perform associative inference by introducing a Fast Weight Memory (FWM) module integrated with an LSTM, which learns compositional state representations. The model achieves excellent performance on tasks like compositional language reasoning, meta-reinforcement learning for POMDPs, and small-scale word-level language modeling.
Humans can quickly associate stimuli to solve problems in novel contexts. Our novel neural network model learns state representations of facts that can be composed to perform such associative inference. To this end, we augment the LSTM model with an associative memory, dubbed Fast Weight Memory (FWM). Through differentiable operations at every step of a given input sequence, the LSTM updates and maintains compositional associations stored in the rapidly changing FWM weights. Our model is trained end-to-end by gradient descent and yields excellent performance on compositional language reasoning problems, meta-reinforcement-learning for POMDPs, and small-scale word-level language modelling.