Cluster-Former: Clustering-based Sparse Transformer for Long-Range Dependency Encoding
This addresses the bottleneck of handling long-range dependencies in tasks like question answering, offering a novel method for efficient encoding.
The paper tackles the problem of Transformer's quadratic complexity with long sequences by proposing Cluster-Former, a clustering-based sparse Transformer that encodes local and global context iteratively, achieving state-of-the-art performance on several major QA benchmarks.
Transformer has become ubiquitous in the deep learning field. One of the key ingredients that destined its success is the self-attention mechanism, which allows fully-connected contextual encoding over input tokens. However, despite its effectiveness in modeling short sequences, self-attention suffers when handling inputs with extreme long-range dependencies, as its complexity grows quadratically with respect to the sequence length. Therefore, long sequences are often encoded by Transformer in chunks using a sliding window. In this paper, we propose Cluster-Former, a novel clustering-based sparse Transformer to perform attention across chunked sequences. The proposed framework is pivoted on two unique types of Transformer layer: Sliding-Window Layer and Cluster-Former Layer, which encode local sequence information and global context jointly and iteratively. This new design allows information integration beyond local windows, which is especially beneficial for question answering (QA) tasks that rely on long-range dependencies. Experiments show that Cluster-Former achieves state-of-the-art performance on several major QA benchmarks.