Sarah Rastegar

Hello, I am a Ph.D. candidate at the University of Amsterdam (UvA), privileged to be under the supervision of Prof. Cees Snoek. Additionally, I am fortunate to receive insightful guidance from Prof. Yuki Asano.
Before embarking on my Ph.D., I pursued my Master's degree at Sharif University of Technology, Tehran, Iran, where I was honored to work under the supervision of Dr. Mahdieh Soleymani Baghshah.

Email  /  Google Scholar  /  Github  /  Linkedin  /  Twitter   

Best Papers Top Venues  /  Open World Papers

In this paper, we introduce a novel concept called self-expertise, which enhances the model's ability to recognize subtle differences and uncover unknown categories. Our approach combines unsupervised and supervised self-expertise strategies to refine the model's discernment and generalization. Our supervised technique differs from traditional methods by utilizing more abstract positive and negative samples, aiding in the formation of clusters that can generalize to novel categories. Meanwhile, our unsupervised strategy encourages the model to sharpen its category distinctions by considering within-category examples as `hard' negatives. Supported by theoretical insights, our empirical results showcase that our method outperforms existing state-of-the-art techniques in Generalized Category Discovery across several fine-grained datasets.

What exactly delineates a category? In this paper, we conceptualize a category through the lens of optimization, viewing it as an optimal solution to a well-defined problem. Harnessing this unique conceptualization, we propose a novel, efficient and self-supervised method capable of discovering previously unknown categories at test time. A salient feature of our approach is the assignment of minimum length category codes to individual data instances, which encapsulates the implicit category hierarchy prevalent in real-world datasets

We aim to recognize actions under an appearance distribution-shift between a source training-domain and target test-domain. To enable such video domain generalization, our key idea is to intervene on the action to remove the confounding effect of the domain-background on the class label using causal inference. Towards this, we propose to learn a causally debiased model on a source domain that in- tervenes on the action through three possible Do-operators which separate the action and background. To better align the source and target distributions we also introduce a test-time action intervention.

An ideal model for multimodal data can reason about missing modalities using the available ones, and usually provides more information when multiple modalities are being considered. All the previous deep mod- els contain separate modality-specific networks and find a shared representation on top of those networks. Therefore, they only consider high level interactions between modal- ities to find a joint representation for them. In this paper, we propose a multimodal deep learning framework (MDL- CW) that exploits the cross weights between representation of modalities, and try to gradually learn interactions of the modalities in a deep network manner from low to high level interactions.

Template from Jon Barron.