arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2302.11208cs.CVcs.LG
收藏

KS-DETR:检测Transformer注意学习中的知识共享

KS-DETR: Knowledge Sharing in Attention Learning for Detection Transformer

Toyota Technological Institute(丰田工业大学)

Kaikai Zhao, Norimichi Ukita

摘要:

Scaled dot-product attention对查询和键的缩放点积应用softmax函数来计算权重,然后将权重和值相乘。在这项工作中,我们研究了如何改善缩放点积注意力的学习,以提高DETR的准确性。我们的方法基于以下观察:使用地面真实前景-背景掩码(GT Fg-Bg掩码)作为权重/值学习中的额外线索,可以学习更好的权重/值;使用更好的权重/值,可以学习更好的值/权重。我们提出了一个三重注意力模块,其中第一个注意力是一个普通的缩放点积注意力,第二/第三个注意力生成高质量的权重/值(在GT Fg-Bg Mask的帮助下),并与第一个注意力共享值/权重,以提高值/权重的质量。第二和第三注意力在推理过程中被移除。我们称我们的方法知识共享DETR(KS-DETR),这是一个扩展的知识蒸馏(KD)的方式,改进的权重和值的教师(第二次和第三次注意)直接共享,而不是模仿,由学生(第一次注意),使更有效的知识转移从教师到学生。各种DETR类方法的实验表明,在MS COCO基准的基线方法的一致改进。代码可在https://github.com/edocanonymous/KS-DETR上获得。

英文摘要:

Scaled dot-product attention applies a softmax function on the scaled dot-product of queries and keys to calculate weights and then multiplies the weights and values. In this work, we study how to improve the learning of scaled dot-product attention to improve the accuracy of DETR. Our method is based on the following observations: using ground truth foreground-background mask (GT Fg-Bg Mask) as additional cues in the weights/values learning enables learning much better weights/values; with better weights/values, better values/weights can be learned. We propose a triple-attention module in which the first attention is a plain scaled dot-product attention, the second/third attention generates high-quality weights/values (with the assistance of GT Fg-Bg Mask) and shares the values/weights with the first attention to improve the quality of values/weights. The second and third attentions are removed during inference. We call our method knowledge-sharing DETR (KS-DETR), which is an extension of knowledge distillation (KD) in the way that the improved weights and values of the teachers (the second and third attentions) are directly shared, instead of mimicked, by the student (the first attention) to enable more efficient knowledge transfer from the teachers to the student. Experiments on various DETR-like methods show consistent improvements over the baseline methods on the MS COCO benchmark. Code is available at https://github.com/edocanonymous/KS-DETR.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑