arXiv:1812.01404

Deep Attention-guided Hashing

Zhan Yang, Osolo Ian Raymond, Wuqing Sun, Jun Long

제출 2018-12-04

원문 초록

With the rapid growth of multimedia data (e.g., image, audio and video etc.) on the web, learning-based hashing techniques such as Deep Supervised Hashing (DSH) have proven to be very efficient for large-scale multimedia search. The recent successes seen in Learning-based hashing methods are largely due to the success of deep learning-based hashing methods. However, there are some limitations to previous learning-based hashing methods (e.g., the learned hash codes containing repetitive and highly correlated information). In this paper, we propose a novel learning-based hashing method, named Deep Attention-guided Hashing (DAgH). DAgH is implemented using two stream frameworks. The core idea is to use guided hash codes which are generated by the hashing network of the first stream framework (called first hashing network) to guide the training of the hashing network of the second stream framework (called second hashing network). Specifically, in the first network, it leverages an attention network and hashing network to generate the attention-guided hash codes from the original images. The loss function we propose contains two components: the semantic loss and the attention loss. The attention loss is used to punish the attention network to obtain the salient region from pairs of images; in the second network, these attention-guided hash codes are used to guide the training of the second hashing network (i.e., these codes are treated as supervised labels to train the second network). By doing this, DAgH can make full use of the most critical information contained in images to guide the second hashing network in order to learn efficient hash codes in a true end-to-end fashion. Results from our experiments demonstrate that DAgH can generate high quality hash codes and it outperforms current state-of-the-art methods on three benchmark datasets, CIFAR-10, NUS-WIDE, and ImageNet.

함께 살펴볼 논문

공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.

arXiv:2408.17221v3cs.LG공통 태그 Deep Attention공식 코드 · Papers with Code

Lightning Self-Attention의 기하학: 식별 가능성과 차원

Geometry of Lightning Self-Attention: Identifiability and Dimension

정규화가 없는 셀프 어텐션 네트워크가 정의하는 함수 공간을 고려하고, 그 기하학을 이론적으로 분석한다. 이러한 네트워크는 다항식이므로 대수 기하학의 도구를 활용한다. 특히, 임의의 계층 수에 대해 매개화의 일반 섬유에 대한 기술을 제공함으로써 깊은 어텐션의 식별 가능성을 연구하고, 그 결과 함수 공간의 차원을 계산한다. 또한 단일 계층 모델에 대해서는 특이점과 경계점을 특성화한다. 마지막으로, 본 결과를 정규화된 셀프 어텐션 네트워크로 확장하는 추측을 정식화하고, 단일 계층에 대해서는 이를 증명하며, 깊은 경우에는 수치적으로 검증한다.

토론

이 논문과 연결된 포럼 글이 아직 없습니다.