A Neural Multi-sequence Alignment TeCHnique (NeuMATCH)
This addresses a challenging alignment problem for applications in multimedia and AI, with incremental improvements over existing methods.
The paper tackles the problem of aligning heterogeneous sequential data like video to text, proposing an end-to-end neural architecture that outperforms state-of-the-art baselines in experiments on semi-synthetic and real datasets.
The alignment of heterogeneous sequential data (video to text) is an important and challenging problem. Standard techniques for this task, including Dynamic Time Warping (DTW) and Conditional Random Fields (CRFs), suffer from inherent drawbacks. Mainly, the Markov assumption implies that, given the immediate past, future alignment decisions are independent of further history. The separation between similarity computation and alignment decision also prevents end-to-end training. In this paper, we propose an end-to-end neural architecture where alignment actions are implemented as moving data between stacks of Long Short-term Memory (LSTM) blocks. This flexible architecture supports a large variety of alignment tasks, including one-to-one, one-to-many, skipping unmatched elements, and (with extensions) non-monotonic alignment. Extensive experiments on semi-synthetic and real datasets show that our algorithm outperforms state-of-the-art baselines.