Artificial Intelligence
February 1, 2025

Papers accepted by leading conferences


2 Papers Accepted at AAAI 2025 – One of the world’s top AI conferences (A*-ranked, alongside NeurIPS, ICML, IJCAI, and CVPR).

Paper 1: Fragility, Robustness, and Antifragility in AI
This paper enhances deep learning security by defending against adversarial attacks, improving model robustness through parameter analysis, and optimizing deep learning models for Edge AI via compression and pruning.

Paper 2: LAGD – Local Topological Alignment & Global Semantic Deconstruction for 3D Segmentation
LAGD introduces a new framework to prevent catastrophic forgetting in incremental 3D segmentation. It also develops Interaction Topological-aware Alignment (ITA) for efficient representation transfer, making it highly relevant for medical imaging, autonomous vehicles, and Edge AI applications.