publications

2026

  1. Universal Hypothesis in Neural Networks
    Prakhar Kaushik, and  collaborators
    A concentrated effort to establish a universal, testable hypothesis for mechanistic interpretation of neural network components. , 2026
  2. Large-Scale 3D Part Dataset
    Prakhar Kaushik, and  collaborators
    A richly annotated, large-scale 3D part dataset for objects enabling part-level recognition, segmentation and part-aware pose estimation. , 2026
  3. Cognitive Visual Reasoning in VLMs and LLMs
    Prakhar Kaushik, and  collaborators
    Benchmarks and models to evaluate and improve compositional visual reasoning and multimodal chain-of-thought grounding. , 2026
  4. High-Quality 3D/4D Animal Asset Creation
    Prakhar Kaushik, and  collaborators
    Pipelines and tools for photorealistic, temporally-coherent 3D/4D animal asset creation from real world data. , 2026

2025

  1. Scaling 3D Compositional Models for Robust Classification and Pose Estimation
    Prakhar Kaushik*, Xiaoding Yuan*, Guofeng Zhang*, and 3 more authors
    The IEEE/CVF International Conference on Computer Vision (ICCV) , 2025
  2. gs2.png
    Gaussian Scenes: Pose-Free Sparse-View Scene Reconstruction using Depth-Enhanced Diffusion Priors
    Soumava Paul, Prakhar Kaushik, and Alan Yuille
    Transactions on Machine Learning Research (TMLR) , 2025
    HTML Website
  3. Progressive Prompt Detailing for Improved Alignment in Text-to-Image Generative Models
    Ketan Suhaas Saichandran, Xavier Thomas, Prakhar Kaushik, and 1 more author
    Oral Presentation. The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , 2025
    HTML
  4. elorax.jpg
    EigenLoRAx: Recycling Adapters to Find Principal Subspaces for Resource-Efficient Adaptation and Inference
    Prakhar Kaushik, Ankit Vaidya, Shravan Chaudhari, and 1 more author
    Oral Presentation. The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , 2025
    HTML v1 Website
  5. eflux.jpeg
    EigenFlux: Parameter Efficient Continual Finetuning via Low-Rank Shared Subspace Adaptation
    Prakhar Kaushik, Ankit Vaidya, and Alan Yuille
    Strict Continual Learning for large Vision and Language Models. Code and Draft coming soon. , 2025

2024

  1. iNeMo: Incremental Neural Mesh Models for Robust Class-Incremental Learning
    Tom Fischer, Yaoyao Liu, Artur Jesslen, and 6 more authors
    In European Conference on Computer Vision , 2024
    Code
  2. Learning Part Segmentation from Synthetic Animals
    Jiawei Peng, Ju He, Prakhar Kaushik, and 3 more authors
    In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , 2024
  3. poseuda.jpg
    Source-Free and Image-Only Unsupervised Domain Adaptation for Category Level Object Pose Estimation
    Prakhar Kaushik, Aayush Mishra, Adam Kortylewski, and 1 more author
    In International Conference on Learning Representations (ICLR) , 2024
    HTML Website
  4. ugt.jpeg
    A Bayesian Approach to OOD Robustness in Image Classification
    Prakhar Kaushik, Adam Kortylewski, and Alan Yuille
    In The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , 2024
    Code Website

2023

  1. Animal3d: A comprehensive dataset of 3d animal pose and shape
    Jiacong Xu, Yi Zhang, Jiawei Peng, and 8 more authors
    In Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023

2021

  1. Understanding catastrophic forgetting and remembering in continual learning with optimal relevance mapping
    Prakhar Kaushik, Alex Gain, Adam Kortylewski, and 1 more author
    In 2021 NeurIPS Fifth Workshop on Meta-Learning; arXiv preprint arXiv:2102.11343 , 2021
    HTML Code

2020

  1. Adaptive neural connections for sparsity learning
    Alex Gain*, Prakhar Kaushik*, and Hava Siegelmann
    In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , 2020
    HTML

2019

  1. Radar as a Security Measure-Real Time Neural Model based Human Detection and Behaviour Classification
    Prakhar Kaushik
    In 2019 IEEE Global Conference on Signal and Information Processing (GlobalSIP) , 2019

2017

  1. An Offline Outdoor Navigation System with Full Privacy.
    Prakhar Kaushik, Vishal Saraswat, and Francesco Buccafurri
    In WINSYS , 2017
  2. Timing attack analysis on AES on modern processors
    Prakhar Kaushik, and Rana Majumdar
    In 2017 6th International Conference on Reliability, Infocom Technologies and Optimization (Trends and Future Directions)(ICRITO) , 2017