Skip to main content

Research Repository

Advanced Search

Dr SHREYANK NARAYANA GOWDA's Outputs (3)

ZeroDiff: Solidified Visual-semantic Correlation in Zero-Shot Learning (2025)
Presentation / Conference Contribution
Ye, Z., Gowda, S. N., Chen, S., Huang, X., Xu, H., Fahad Khan, S., Jin, Y., Huang, K., & Jin, X. (2025, April). ZeroDiff: Solidified Visual-semantic Correlation in Zero-Shot Learning. Presented at ICLR 2025: The Thirteenth International Conference on Learning Representations, Singapore

Zero-shot Learning (ZSL) aims to enable classifiers to identify unseen classes. This is typically achieved by generating visual features for unseen classes based on learned visual-semantic correlations from seen classes. However, most current generat... Read More about ZeroDiff: Solidified Visual-semantic Correlation in Zero-Shot Learning.

Bridging the Projection Gap: Overcoming Projection Bias Through Parameterized Distance Learning (2024)
Presentation / Conference Contribution
Zhang, C., Jin, M., Yu, Q., Xue, H., Gowda, S. N., & Jin, X. (2024, December). Bridging the Projection Gap: Overcoming Projection Bias Through Parameterized Distance Learning. Presented at 17th Asian Conference on Computer Vision (ACCV 2024), Hanoi, Vietnam

Generalized zero-shot learning (GZSL) aims to recognize samples from both seen and unseen classes using only seen class samples for training. However, GZSL methods are prone to bias towards seen classes during inference due to the projection function... Read More about Bridging the Projection Gap: Overcoming Projection Bias Through Parameterized Distance Learning.

Continual Learning Improves Zero-Shot Action Recognition (2024)
Presentation / Conference Contribution
Gowda, S. N., Moltisanti, D., & Sevilla-Lara, L. (2024, December). Continual Learning Improves Zero-Shot Action Recognition. Presented at Computer Vision – ACCV 2024 17th Asian Conference on Computer Vision, Hanoi, Vietnam

Zero-shot action recognition requires a strong ability to generalize from pre-training and seen classes to novel unseen classes. Similarly, continual learning aims to develop models that can generalize effectively and learn new tasks without forgetti... Read More about Continual Learning Improves Zero-Shot Action Recognition.