Title of the paper
Two-line summary of the contribution and scope.
Publications
Search by concept, filter by research direction, then expand any paper for a concise abstract and links. Full list on Google Scholar.
For hiring and collaboration discussions, start with Universal Weight Subspace and Name That Part.
Cognitive & Structured Perception
Perceptual Taxonomy: Evaluating and Guiding Hierarchical Scene Reasoning in Vision-Language Models
A taxonomy for evaluating hierarchical scene reasoning in vision-language models.
Name That Part: 3D Part Segmentation and Naming
Joint 3D part segmentation and semantic naming.
Learning Part Segmentation from Synthetic Animals
Part segmentation learned from synthetic animal data.
Abstract, links & BibTeX
Preprint link coming soon.
Animal3D: A Comprehensive Dataset of 3D Animal Pose and Shape
A comprehensive dataset for 3D animal pose and shape.
Abstract, links & BibTeX
Preprint link coming soon.
Efficient & Continual Learning
The Universal Weight Subspace Hypothesis
Shared weight-subspace structure for efficient adaptation and continual learning.
EigenLoRAx: Recycling Adapters to Find Principal Subspaces for Resource-Efficient Adaptation and Inference
Resource-efficient adaptation and inference via adapter recycling.
Abstract, links & BibTeX
Shared LoRA Subspaces for Almost Strict Continual Learning
A shared-subspace approach toward almost strict continual learning for large models.
Understanding Catastrophic Forgetting and Remembering in Continual Learning with Optimal Relevance Mapping
Continual learning with optimal relevance mapping.
iNeMo: Incremental Neural Mesh Models for Robust Class-Incremental Learning
Incremental neural mesh models for class-incremental learning.
Adaptive Neural Connections for Sparsity Learning
Adaptive neural connectivity for sparsity-aware learning.
Robust & Compositional Vision / 3D
Can These Views Be One Scene? Evaluating Multiview 3D Consistency when 3D Foundation Models Hallucinate
Failure-aware multiview 3D consistency metrics for hallucination-prone foundation reconstruction models.
Scaling 3D Compositional Models for Robust Classification and Pose Estimation
Scaling compositional models for robust 3D classification and pose.
Source-Free and Image-Only Unsupervised Domain Adaptation for Category-Level Object Pose Estimation
Source-free domain adaptation for category-level pose estimation.
A Bayesian Approach to OOD Robustness in Image Classification
Bayesian OOD robustness in image classification.
Generative & Alignment
TriDiff-4D: Fast 4D Generation through Diffusion-based Triplane Re-posing
Fast 4D generation through diffusion-based triplane re-posing.
Gaussian Scenes: Pose-Free Sparse-View Scene Reconstruction using Depth-Enhanced Diffusion Priors
Pose-free sparse-view scene reconstruction using diffusion priors.
Progressive Prompt Detailing for Improved Alignment in Text-to-Image Generative Models
Improved alignment in text-to-image generative models.
Security & Systems
An Offline Outdoor Navigation System with Full Privacy
Offline outdoor navigation with full privacy.
Abstract, links & BibTeX
Preprint link coming soon.
Timing Attack Analysis on AES on Modern Processors
Timing attack analysis on AES on modern processors.
Abstract, links & BibTeX
Preprint link coming soon.
Radar as a Security Measure: Real-Time Neural Model-Based Human Detection and Behavior Classification
Real-time neural model-based human detection and behavior classification.
Abstract, links & BibTeX
Preprint link coming soon.
2026
Can These Views Be One Scene? Evaluating Multiview 3D Consistency when 3D Foundation Models Hallucinate
Failure-aware multiview 3D consistency metrics for hallucination-prone foundation reconstruction models.
Name That Part: 3D Part Segmentation and Naming
Joint 3D part segmentation and semantic naming.
Shared LoRA Subspaces for Almost Strict Continual Learning
A shared-subspace approach toward almost strict continual learning for large models.
2025
Perceptual Taxonomy: Evaluating and Guiding Hierarchical Scene Reasoning in Vision-Language Models
A taxonomy for evaluating hierarchical scene reasoning in vision-language models.
The Universal Weight Subspace Hypothesis
Shared weight-subspace structure for efficient adaptation and continual learning.
Scaling 3D Compositional Models for Robust Classification and Pose Estimation
Scaling compositional models for robust 3D classification and pose.
TriDiff-4D: Fast 4D Generation through Diffusion-based Triplane Re-posing
Fast 4D generation through diffusion-based triplane re-posing.
Gaussian Scenes: Pose-Free Sparse-View Scene Reconstruction using Depth-Enhanced Diffusion Priors
Pose-free sparse-view scene reconstruction using diffusion priors.
Progressive Prompt Detailing for Improved Alignment in Text-to-Image Generative Models
Improved alignment in text-to-image generative models.
EigenLoRAx: Recycling Adapters to Find Principal Subspaces for Resource-Efficient Adaptation and Inference
Resource-efficient adaptation and inference via adapter recycling.
Abstract, links & BibTeX
2024
iNeMo: Incremental Neural Mesh Models for Robust Class-Incremental Learning
Incremental neural mesh models for class-incremental learning.
Learning Part Segmentation from Synthetic Animals
Part segmentation learned from synthetic animal data.
Abstract, links & BibTeX
Preprint link coming soon.
Source-Free and Image-Only Unsupervised Domain Adaptation for Category-Level Object Pose Estimation
Source-free domain adaptation for category-level pose estimation.
A Bayesian Approach to OOD Robustness in Image Classification
Bayesian OOD robustness in image classification.
2023
Animal3D: A Comprehensive Dataset of 3D Animal Pose and Shape
A comprehensive dataset for 3D animal pose and shape.
Abstract, links & BibTeX
Preprint link coming soon.
2021
Understanding Catastrophic Forgetting and Remembering in Continual Learning with Optimal Relevance Mapping
Continual learning with optimal relevance mapping.
2020
Adaptive Neural Connections for Sparsity Learning
Adaptive neural connectivity for sparsity-aware learning.
2019
Radar as a Security Measure: Real-Time Neural Model-Based Human Detection and Behavior Classification
Real-time neural model-based human detection and behavior classification.
Abstract, links & BibTeX
Preprint link coming soon.
2017
An Offline Outdoor Navigation System with Full Privacy
Offline outdoor navigation with full privacy.
Abstract, links & BibTeX
Preprint link coming soon.
Timing Attack Analysis on AES on Modern Processors
Timing attack analysis on AES on modern processors.
Abstract, links & BibTeX
Preprint link coming soon.
No matching publications. Try a broader term or switch filters.