Attention for Inference Compilation
This work addresses a bottleneck in amortized inference for probabilistic programming, offering an incremental improvement for researchers in machine learning and statistics.
The paper tackles the problem of modeling long-range dependencies in inference compilation for probabilistic programs by introducing an attention mechanism, resulting in improved performance with better matching of proposal distributions to the true posterior.
We present a new approach to automatic amortized inference in universal probabilistic programs which improves performance compared to current methods. Our approach is a variation of inference compilation (IC) which leverages deep neural networks to approximate a posterior distribution over latent variables in a probabilistic program. A challenge with existing IC network architectures is that they can fail to model long-range dependencies between latent variables. To address this, we introduce an attention mechanism that attends to the most salient variables previously sampled in the execution of a probabilistic program. We demonstrate that the addition of attention allows the proposal distributions to better match the true posterior, enhancing inference about latent variables in simulators.