Turning a Blind Eye: Explicit Removal of Biases and Variation from Deep Neural Network Embeddings

Alvi, Mohsan; Zisserman, Andrew; Nellåker, Christoffer

doi:10.1007/978-3-030-11009-3_34

Cited by 175 publications

(232 citation statements)

References 26 publications

Supporting

Mentioning

217

Contrasting

Order By: Relevance

“…These approaches rely on an oracle for a subset of test queries. Rather than relying on an oracle, Alvi et al [1] proposed joint learning and unlearning method to remove bias from neural network embedding. To unlearn the bias, the authors applied confusion loss, which is computed by calculating the cross-entropy between classifier output and a uniform distribution.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Learning Not to Learn: Training Deep Neural Networks With Biased Data

Kim

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

290

307

View full text Add to dashboard Cite

We propose a novel regularization algorithm to train deep neural networks, in which data at training time is severely biased. Since a neural network efficiently learns data distribution, a network is likely to learn the bias information to categorize input data. It leads to poor performance at test time, if the bias is, in fact, irrelevant to the categorization. In this paper, we formulate a regularization loss based on mutual information between feature embedding and bias. Based on the idea of minimizing this mutual information, we propose an iterative algorithm to unlearn the bias information. We employ an additional network to predict the bias distribution and train the network adversarially against the feature embedding network. At the end of learning, the bias prediction network is not able to predict the bias not because it is poorly trained, but because the feature embedding network successfully unlearns the bias information. We also demonstrate quantitative and qualitative experimental results which show that our algorithm effectively removes the bias information from feature embedding.

show abstract

Section: Related Workmentioning

confidence: 99%

“…As mentioned by Alvi et al in the paper [1], the unsupervised domain adaptation (UDA) problem is closely related to the biased data problem. The UDA problem involves generalizing the network embedding over different domains [8,23,21].…”

Section: Related Workmentioning

confidence: 99%

Learning Not to Learn: Training Deep Neural Networks With Biased Data

Kim

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

290

307

View full text Add to dashboard Cite

show abstract

“…A disentanglement loss (DL) is used to encourage explicit separation of representations-for this we use the confusion loss implemented by [29] (inspired by [30]). This loss is used to assess the amount of spurious variation information left in either feature representation and then remove it (for the identity representation, content information is a spurious variation and vice versa).…”

Section: Loss Functionsmentioning

confidence: 99%

Disentangled Speech Embeddings Using Cross-Modal Self-Supervision

Nagrani

Chung

Albanie

et al. 2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

The objective of this paper is to learn representations of speaker identity without access to manually annotated data. To do so, we develop a self-supervised learning objective that exploits the natural cross-modal synchrony between faces and audio in video. The key idea behind our approach is to tease apart-without annotation-the representations of linguistic content and speaker identity. We construct a two-stream architecture which: (1) shares low-level features common to both representations; and (2) provides a natural mechanism for explicitly disentangling these factors, offering the potential for greater generalisation to novel combinations of content and identity and ultimately producing speaker identity representations that are more robust.We train our method on a large-scale audio-visual dataset of talking heads 'in the wild', and demonstrate its efficacy by evaluating the learned speaker representations for standard speaker recognition performance.

show abstract

“…But how can age prediction performances be enhanced in this case? With this objective in mind, the analysis of bias in age perception has recently emerged [2], [9]. Can we better understand age perception and their biases so that the findings can be used to regress a better real age estimation?…”

Section: Analysis Of Bias In Age Estimationmentioning

confidence: 99%

“…However, an end-to-end approach for bias removal was not considered. According to Alvi et al [9], training an age predictor on a dataset that is not balanced for gender can lead to gender biased predictions. They presented an algorithm to remove biases from the feature representation, as well as to ensure that the network is blind to a known bias in the dataset.…”

Section: Analysis Of Bias In Age Estimationmentioning

confidence: 99%

On the effect of age perception biases for real age regression

Jacques

Özçinar

Marjanović

et al. 2019

2019 14th IEEE International Conference on Automatic Face &Amp; Gesture Recognition (FG 2019)

View full text Add to dashboard Cite

Automatic age estimation from facial images represents an important task in computer vision. This paper analyses the effect of gender, age, ethnic, makeup and expression attributes of faces as sources of bias to improve deep apparent age prediction. Following recent works where it is shown that apparent age labels benefit real age estimation, rather than direct real to real age regression, our main contribution is the integration, in an end-to-end architecture, of face attributes for apparent age prediction with an additional loss for real age regression. Experimental results on the APPA-REAL dataset indicate the proposed network successfully take advantage of the adopted attributes to improve both apparent and real age estimation. Our model outperformed a state-of-the-art architecture proposed to separately address apparent and real age regression. Finally, we present preliminary results and discussion of a proof of concept application using the proposed model to regress the apparent age of an individual based on the gender of an external observer.

show abstract

Turning a Blind Eye: Explicit Removal of Biases and Variation from Deep Neural Network Embeddings

Cited by 175 publications

References 26 publications

Learning Not to Learn: Training Deep Neural Networks With Biased Data

Learning Not to Learn: Training Deep Neural Networks With Biased Data

Disentangled Speech Embeddings Using Cross-Modal Self-Supervision

On the effect of age perception biases for real age regression

Contact Info

Product

Resources

About