Research

Currently, my research interest spans the broader spectrum of Generative AI. I gravitate towards problems at opposite ends of the spectrum - either those with clear practical applications or purely exploratory endeavors that spark curiosity.

Publications

• Hybrid Graphs for Table-and-Text based Question Answering using LLMs
  Ankush Agarwal, Ganesh S, Chaitanya Devaguptapu
  In Proceedings of the 2025 Conference of the North American Chapter of the Association for Computational Linguistics (NAACL 2025 - Main Track)
  arXiv

• Towards a Training Free Approach for 3D Scene Editing
  Vivek Madhavaram, Shivangana Rawat, Chaitanya Devaguptapu, Charu Sharma, Manohar Kaul
  In Proceedings of IEEE Winter Conference on Applications of Computer Vision (WACV), Hawaii, USA. 2025
  arXiv

• HOLMES: Hyper-Relational Knowledge Graphs for Multi-hop Question Answering using LLMs
  Pranoy Panda, Ankush Agarwal, Chaitanya Devaguptapu, Manohar Kaul, Prathosh A P
  In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024 - Main Track)
  arXiv

• Synergizing Contrastive Learning and Optimal Transport for 3D Point Cloud Domain Adaptation
  Siddharth Katageri^†, Arkadipta De^†, Chaitanya Devaguptapu^†, VSSV Prasad, Charu Sharma, Manohar Kaul
  In Proceedings of IEEE Winter Conference on Applications of Computer Vision (WACV), Hawaii, USA. 2024
  Oral (Top ~7% of all the accepted papers)
  ^{† Equal Contribution}
  arXiv Code Project Page

• Partial Label Learning meets Active Learning: Enhancing Annotation Efficiency through Binary Questioning
  Shivangana Rawat, Chaitanya Devaguptapu, Vineeth N Balasubramanian
  Workshop on Artificial Intelligence & Human-Computer Interaction, ICML-2023
  Published PDF Poster Video

• On Adversarial Robustness: A Neural Architecture Search perspective
  Chaitanya Devaguptapu, Devansh Agarwal, Gaurav Mittal, Pulkit Gopalani, Vineeth N Balasubramanian Workshop on Adversarial Robustness in the Real World, IEEE/CVF International Conference on Computer Vision (ICCV’21) Also accepted at ICLR-21 workshops as Contributed talk and Spotlight
  arXiv Code Poster Workshop

• On Initial Pools for Deep Active Learning
  Akshay L Chandra^†, S.V. Desai^†, Chaitanya Devaguptapu^†, Vineeth N Balasubramanian
  Preregistration Workshop at NeurIPS 2020 (PMLR Volume 148)
  ^{† Equal Contribution}
  arXiv Code Poster Workshop

• Borrow From Anywhere: Pseudo Multi-Modal Object Detection in Thermal Imagery
  Chaitanya Devaguptapu, Ninad Akolekar, Manuj M Sharma, Vineeth N Balasubramanian
  Workshop on Perception beyond visible spectrum, IEEE/CVF International Conference on Computer Vision and Pattern recognition (CVPR’19)
  (Spotlight, Among top 6 of the 30 accepted papers)
  arXiv Code Poster Workshop

Pre-prints

• ∆-Patching: A Framework for Rapid Adaptation of Pre-trained Convolutional Networks without Base Performance Loss
  Chaitanya Devaguptapu, Samarth Sinha, K J Joseph, Vineeth N Balasubramanian, Animesh Garg
  Under Review
  arXiv

• Regularizing Self-Supervised Learning on Small Datasets via Semantic Graph Consistency
  Chaitanya Devaguptapu, Sumukh Aithal, Yamada Moyuru, Manohar Kaul
  Under Review
  

Miscellaneous

• Object Detection in Low-Resolution Thermal Imagery - Enhanced the performance of object detection in thermal images by increasing the resolution of 160 x 120 images using super resolution and various deep learning based image interpolation techniques. This was a joint project with DRDO, Government of India

• Indian patent - Chaitanya Devaguptapu, Ninad Akolekar, Manuj M Sharma, Vineeth N Balasubramanian, A Methodology for Transfer of Knowledge from Data-Rich Domains for Thermal Image Processing, Indian Patent Application No. 202011032663 (filed in Aug 2020)