Zeynep Akata

Interpreting Adversarial Examples with Attributes

2019 · arXiv (Cornell University)

Deep computer vision systems being vulnerable to imperceptible and carefully crafted noise have raised questions regarding the robustness of their decisions. We take a step back and approach this problem from an orthogonal direction. We …

Learning Deep Representations of Fine-grained Visual Descriptions

2016 · arXiv (Cornell University)

State-of-the-art methods for zero-shot visual recognition formulate learning as a joint embedding problem of images and side information. In these formulations the current best complement to visual features are attributes: manually encoded vectors describing shared …

Fantastic Gains and Where to Find Them: On the Existence and Prospect of General Knowledge Transfer between Any Pretrained Model

2023 · arXiv (Cornell University)

Training deep networks requires various design decisions regarding for instance their architecture, data augmentation, or optimization. In this work, we find these training variations to result in networks learning unique feature sets from the data. …

COSMOS: Cross-Modality Self-Distillation for Vision Language Pre-training

2024 · arXiv (Cornell University)

Vision-Language Models (VLMs) trained with contrastive loss have achieved significant advancements in various vision and language tasks. However, the global nature of the contrastive loss makes VLMs focus predominantly on foreground objects, neglecting other crucial …