A causal perspective on dataset bias in machine learning for medical imaging

C Jones, DC Castro, F De Sousa Ribeiro… - Nature Machine …, 2024 - nature.com
As machine learning methods gain prominence within clinical decision-making, the need to
address fairness concerns becomes increasingly urgent. Despite considerable work …

Fairness in deep learning: A survey on vision and language research

O Parraga, MD More, CM Oliveira, NS Gavenski… - ACM Computing …, 2023 - dl.acm.org
Despite being responsible for state-of-the-art results in several computer vision and natural
language processing tasks, neural networks have faced harsh criticism due to some of their …

On the opportunities and risks of foundation models

R Bommasani, DA Hudson, E Adeli, R Altman… - arXiv preprint arXiv …, 2021 - arxiv.org
AI is undergoing a paradigm shift with the rise of models (eg, BERT, DALL-E, GPT-3) that are
trained on broad data at scale and are adaptable to a wide range of downstream tasks. We …

Conceptual 12m: Pushing web-scale image-text pre-training to recognize long-tail visual concepts

S Changpinyo, P Sharma, N Ding… - Proceedings of the …, 2021 - openaccess.thecvf.com
The availability of large-scale image captioning and visual question answering datasets has
contributed significantly to recent successes in vision-and-language pre-training. However …

Don't take the easy way out: Ensemble based methods for avoiding known dataset biases

C Clark, M Yatskar, L Zettlemoyer - arXiv preprint arXiv:1909.03683, 2019 - arxiv.org
State-of-the-art models often make use of superficial patterns in the data that do not
generalize well to out-of-domain or adversarial settings. For example, textual entailment …

Think locally, act globally: Federated learning with local and global representations

PP Liang, T Liu, L Ziyin, NB Allen, RP Auerbach… - arXiv preprint arXiv …, 2020 - arxiv.org
Federated learning is a method of training models on private data distributed over multiple
devices. To keep device data private, the global model is trained by only communicating …

Iti-gen: Inclusive text-to-image generation

C Zhang, X Chen, S Chai, CH Wu… - Proceedings of the …, 2023 - openaccess.thecvf.com
Text-to-image generative models often reflect the biases of the training data, leading to
unequal representations of underrepresented groups. This study investigates inclusive text …

Towards fairness in visual recognition: Effective strategies for bias mitigation

Z Wang, K Qinami, IC Karakozis… - Proceedings of the …, 2020 - openaccess.thecvf.com
Computer vision models learn to perform a task by capturing relevant statistics from training
data. It has been shown that models learn spurious age, gender, and race correlations when …

Harms of gender exclusivity and challenges in non-binary representation in language technologies

S Dev, M Monajatipoor, A Ovalle… - arXiv preprint arXiv …, 2021 - arxiv.org
Gender is widely discussed in the context of language tasks and when examining the
stereotypes propagated by language models. However, current discussions primarily treat …

Dall-eval: Probing the reasoning skills and social biases of text-to-image generation models

J Cho, A Zala, M Bansal - Proceedings of the IEEE/CVF …, 2023 - openaccess.thecvf.com
Recently, DALL-E, a multimodal transformer language model, and its variants including
diffusion models have shown high-quality text-to-image generation capabilities. However …