Person Re-identification Method Based on Multi-view and Attention Mechanism

Liu, Xianglan; Pan, Xuguang; Li, Runqing; Song, Bo; Rong, Leilei; Xu, Yan

doi:10.1109/prml56267.2022.9882201

Cited by 1 publication

(1 citation statement)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The weights are computed based on the feature's interactions with other features and the sample's proximity to the training data. These algorithms are not the only ones that are available and other approaches like Anchors [17], Counterfactuals [18], and Attention [19][20][21] mechanisms can also be used to explain DL models. The main idea behind these algorithms is to linearize the model locally around the sample space to provide an explanation, which can be more interpretable and understandable than the complex deep learning model itself.…”

Section: Introductionmentioning

confidence: 99%

SpecXAI -- Spectral interpretability of Deep Learning Models

Druc¹,

Wooldridge²,

Krishnamurthy³

et al. 2023

Preprint

View full text Add to dashboard Cite

Deep learning is becoming increasingly adopted in business and industry due to its ability to transform large quantities of data into high-performing models. These models, however, are generally regarded as black boxes, which, in spite of their performance, could prevent their use. In this context, the field of eXplainable AI attempts to develop techniques that temper the impenetrable nature of the models and promote a level of understanding of their behavior.Here we present our contribution to XAI methods in the form of a framework that we term SpecXAI, which is based on the spectral characterization of the entire network. We show how this framework can be used to not only understand the network but also manipulate it into a linear interpretable symbolic representation.

show abstract