Scholar
Darshana Saravanan
Google Scholar ID: SgbmMLUAAAAJ
IIIT Hyderabad
Citations & Impact
Citations
8
H-index
1
i10-index
0
Publications
3
Co-authors
5
list available
Publications
Resume (English only)
Academic Achievements
- Published three papers:
- 1. Investigating Mechanisms for In-Context Vision Language Binding
- 2. VELOCITI: Benchmarking Video-Language Compositional Reasoning with Strict Entailment
- 3. Pseudo-labelling meets Label Smoothing for Noisy Partial Label Learning, and received the Best Paper Award for one of them.
Research Experience
- Worked as a Software Development Intern at Fidelity Investments.
Education
- Master's student at the Center for Visual Information Technology (CVIT), IIIT Hyderabad, advised by Prof. Vineet Gandhi; B.Tech. (Hons) in Information Technology from SASTRA Deemed University.
Background
- Research interests include multimodal models and interpretability. Currently focusing on understanding compositionality in Vision-Language Models.
