Samira Zare

Dr. Samira Zare

orcid

Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
5 December 2023 in IEEE Access
Verified
Journal Article
Open access
SZSamira Zare
Samira Zare
HNHien Van Nguyen
Hien Van Nguyen
Despite the remarkable progress of self-supervised learning (SSL), how self-supervised representations generalize to out-of-distribution data remains little understood. In this paper, we study the effects of distribution shifts on self-supervised representations. Our findings indicate that self-supervised representation learning is more robust than traditional supervised learning (52.8% versus 17.1% on the CMNIST dataset, 63.6% versus 60.6% on the Waterbirds dataset). However, self-supervised representations still suffer significantly from domain shifts, especially when spurious correlations are present. Motivated by this limitation, we propose a risk-extrapolated information NCE (ReinformNCE) to facilitate self-supervised learning algorithms to learn more stable representations. Our approach integrates the infoNCE loss function and a robust optimization approach that extrapolates the risks of training domains. Extensive experiments show that ReinformNCE helps to extract domain-invariant self-supervised representations and it substantially improves the robustness of the self-supervised representations (68.2% versus 52.8% on the CMNIST dataset, 77.9% versus 63.6% on the Waterbirds dataset). To the best of our knowledge, this is the first work demonstrating the feasibility of learning domain-invariant representations based on robust optimization theory and without supervised information.

16 September 2022
Verified
Book Chapter
Closed access
PDF file
SZSamira Zare
Samira Zare
HNHien Van Nguyen
Hien Van Nguyen

3 March 2025 in Kidney360
Verified
Journal Article
Open access
SZSamira Zare
Samira Zare
HVHuy Q. Vo
Huy Q. Vo
Nicola AltiniNicola Altini
Nicola Altini
VBVitoantonio Bevilacqua
Vitoantonio Bevilacqua
MRMichele Rossini
Michele Rossini
HNHien Nguyen
Hien Nguyen
Key Points: Permutation-invariant cascaded attentional set operator (PICASO) is a versatile set operator that uses Transformers to dynamically aggregate histopathologic features from a set of glomerular crops. For detecting active crescent in patients with IgA nephropathy on internal and external validation sets, PICASO achieved an area under the receiver-operating characteristic curve of 0.99 and 0.96, respectively. In the case-level classification of antibody-mediated rejection in kidney transplants, PICASO performed well, with an area under the receiver-operating characteristic curves of 0.97. Background: The advent of digital nephropathology offers the potential to integrate deep learning algorithms into the diagnostic workflow. We introduce permutation-invariant cascaded attentional set operator (PICASO), a novel permutation-invariant set operator to dynamically aggregate histopathologic features from instances. We applied PICASO to two nephropathology scenarios: detecting active crescent lesions in sets of glomerular crops with IgA nephropathy (IgAN) and case-level classification for antibody-mediated rejection (AMR) in kidney transplant. Methods: PICASO is a Transformer-based set operator that aggregates features from sets of instances to make predictions. It uses initial histopathologic vectors as a static memory component and continuously updates them on the basis of input embeddings. For active crescent detection in patients with IgAN, we obtained 6206 periodic acid–Schiff–stained glomerular crops (5792 no active crescent, 414 active crescent) from three different health institutes. For the AMR classification, we have 1655 periodic acid–Schiff–stained glomerular crops (769 AMR and 886 non-AMR images) from 89 biopsies. The performance of PICASO as a set operator was compared with other set operators, such as DeepSet, Set Transformer, DeepSet++, and Set Transformer++, using metrics including area under the receiver-operating characteristic curve (AUROC), area under the precision-recall curves, recall, and accuracy. Results: PICASO achieved superior performance in detecting active crescent in patients with IgAN, with an AUROC of 0.99 (95% confidence interval [CI], 0.98 to 0.99) on internal validation and 0.96 (95% CI, 0.95 to 0.98) on external validation, significantly outperforming other set operators (P < 0.001). It also attained the highest AUROC of 0.97 (95% CI, 0.90 to 1.0, P = 0.02) for case-level AMR classification. The area under the precision-recall curve, recall, and accuracy scores were also higher when using PICASO, and it significantly outperformed baselines (P < 0.001). Conclusions: PICASO can potentially advance nephropathology by improving performance through dynamic feature aggregation.

Verified
Conference Paper
Open access
PDF file
SZSamira Zare
Samira Zare
HNHien Van Nguyen
Hien Van Nguyen
Standard training via empirical risk minimization may result in making predictions that overly rely on spurious correlations. This can degrade the generalization to out-of-distribution settings where these correlations no longer hold. Invariant learning has been shown to be a promising approach for identifying predictors that ignore spurious correlations. However, an important limitation of this approach is that it assumes access to different “environments” (also known as domains), which may not always be available. This paper proposes a simple yet effective strategy for discovering maximally informative environments from a single dataset. Our frustratingly easy environment discovery (FEED) approach trains a biased reference classifier using a generalized cross-entropy loss function and partitions the dataset based on its performance. These environments can be used with various invariant learning algorithms, including Invariant Risk Minimization, Risk Extrapolation, and Group Distributionally Robust Optimization. The results indicate that FEED can discover environments with a higher group sufficiency gap compared to the state-of-the-art environment inference baseline and leads to improved test accuracy on CMNIST, Waterbirds, and CelebA datasets.

2 April 2024 in IEEE Transactions on Artificial Intelligence
Verified
Journal Article
Closed access
YKYuhan Kang
Yuhan Kang
SZSamira Zare
Samira Zare
ALAlex Tong Lin
Alex Tong Lin
ZHZhu Han
Zhu Han
SOStanley Osher
Stanley Osher
HNHien Nguyen
Hien Nguyen
Data augmentation is a critical component in building modern deep-learning systems. In this paper, we propose MFG Augment , a novel data augmentation method based on the Mean-Field-Game (MFG) theory, that can synthesize a sequence of data between every two images or features. The central idea is to consider every image as a distribution over its pixel or feature space. Using Mean-field Game theory, we can generate a time-continuous “path” from one distribution to another so that the points along the “path” are augmented images or features. Empirically, the experiment results on MNIST, CIFAR-10, and ImageNet demonstrate that the proposed technology has better generalization ability and higher classification accuracy as compared to several benchmark methods. More importantly, our MFG Augment improves the test accuracy significantly when the dataset size is small. MFG Augment consistently shows better affinity and diversity scores, two important empirical metrics for evaluating the generalization of data augmentation techniques.

25 May 2023 in arXiv
Verified
Preprint
Open access
PDF file
AAAKASH AWASTHI
AKASH AWASTHI
SLSon Ly
Son Ly
JNJaer Nizam
Jaer Nizam
SZSamira Zare
Samira Zare
VMVideet Mehta
Videet Mehta
HNHien Van Nguyen
Hien Van Nguyen
The definition of anomaly detection is the identification of an unexpected event. Real-time detection of extreme events such as wildfires, cyclones, or floods using satellite data has become crucial for disaster management. Although several earth-observing satellites provide information about disasters, satellites in the geostationary orbit provide data at intervals as frequent as every minute, effectively creating a video from space. There are many techniques that have been proposed to identify anomalies in surveillance videos; however, the available datasets do not have dynamic behavior, so we discuss an anomaly framework that can work on very high-frequency datasets to find very fast-moving anomalies. In this work, we present a diffusion model which does not need any motion component to capture the fast-moving anomalies and outperforms the other baseline methods.