Overcoming Non-monotonicity in Transducer-based Streaming Generation
This addresses a specific bottleneck in industrial streaming applications like translation, but appears incremental as it builds on existing Transducer architectures.
The paper tackled the problem of non-monotonic alignments in Transducer-based streaming generation, such as in simultaneous translation, by integrating a learnable monotonic attention mechanism with the forward-backward algorithm to infer alignments without enumerating exponential possibilities, resulting in a robust solution for complex tasks.
Streaming generation models are utilized across fields, with the Transducer architecture being popular in industrial applications. However, its input-synchronous decoding mechanism presents challenges in tasks requiring non-monotonic alignments, such as simultaneous translation. In this research, we address this issue by integrating Transducer's decoding with the history of input stream via a learnable monotonic attention. Our approach leverages the forward-backward algorithm to infer the posterior probability of alignments between the predictor states and input timestamps, which is then used to estimate the monotonic context representations, thereby avoiding the need to enumerate the exponentially large alignment space during training. Extensive experiments show that our MonoAttn-Transducer effectively handles non-monotonic alignments in streaming scenarios, offering a robust solution for complex generation tasks.