Lee, Wonjun Ham, Bumsub Kim, Suhyun 2025-09-02T03:00:20Z 2025-09-02T03:00:20Z 2025-08-26 2025-02 2159-5399 https://pubs.kist.re.kr/handle/201004/153098 In vision transformers, position embedding (PE) plays a crucial role in capturing the order of tokens. However, in vision transformer structures, there is a limitation in the expressiveness of PE due to the structure where position embedding is simply added to the token embedding. A layer-wise method that delivers PE to each layer and applies independent Layer Normalizations for token embedding and PE has been adopted to overcome this limitation. In this paper, we identify the conflicting result that occurs in a layer-wise structure when using the global average pooling (GAP) method instead of the class token. To overcome this problem, we propose MPVG, which maximizes the effectiveness of PE in a layer-wise structure with GAP. Specifically, we identify that PE counterbalances token embedding values at each layer in a layer-wise structure. Furthermore, we recognize that the counterbalancing role of PE is insufficient in the layer-wise structure, and we address this by maximizing the effectiveness of PE through MPVG. Through experiments, we demonstrate that PE performs a counterbalancing role and that maintaining this counterbalancing directionality significantly impacts vision transformers. As a result, the experimental results show that MPVG outperforms existing methods across vision transformers on various tasks. English ASSOC ADVANCEMENT ARTIFICIAL INTELLIGENCE Maximizing the Position Embedding for Vision Transformers with Global Average Pooling Conference 10.1609/aaai.v39i17.33997 1 39th AAAI Conference on Artificial Intelligence, pp.18154 - 18162 39th AAAI Conference on Artificial Intelligence 18154 18162 US Philadelphia, PA 2025-02-25 THIRTY-NINTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE, AAAI-25, VOL 39 NO 17 001477527200072