Translution: Unifying Self-attention and Convolution for Adaptive and Relative Modeling
This work addresses a foundational challenge in ML/AI by proposing a novel operation to combine strengths of self-attention and convolution, though it is incremental as it builds on existing methods.
The paper tackled the problem of unifying adaptive identification from self-attention and relative encoding from convolution, resulting in Translution and its lightweight variant α-Translution, which achieved superior accuracy compared to self-attention on computer vision and NLP tasks.
When modeling a given type of data, we consider it to involve two key aspects: 1) identifying relevant elements (e.g., image pixels or textual words) to a central element, as in a convolutional receptive field, or to a query element, as in self-attention, and 2) encoding these tokens effectively. Self-attention can adaptively identify these elements but relies on absolute positional embedding for structural representation learning. In contrast, convolution encodes elements in a relative manner, yet their fixed kernel size limits their ability to adaptively select the relevant elements. In this paper, we introduce Translution, an operation that unifies the adaptive identification capability of self-attention and the relative encoding advantage of convolution. However, this integration leads to a substantial increase in the number of parameters, exceeding most currently available computational resources. Therefore, we propose a lightweight variant of Translution, named α-Translution. Experiments on computer vision and natural language processing tasks show that Translution (including α-Translution) achieves superior accuracy compared to self-attention. The code is available at https://github.com/hehefan/Translution.