This is but a single thread in the boundless tapestry woven by mortal hands
14 Posts
2 Categories
3 Tags
Mingjing Xu
14 Posts
2 Categories
3 Tags
Mingjing Xu
CS PhD student @ Swansea University | Efficient and Trustworthy AI
Mingjing Xu is a Computer Science PhD student at Swansea University working on efficient optimization, trustworthy AI, multimodal representation learning, and edge intelligence. He is first author of PSMGD, an AAAI 2025 paper on fast stochastic multi-objective optimization, and co-first author of recent ICLR 2026 and AAAI 2026 work on neural network inheritance and tactile-language-vision alignment. His earlier work spans practical adversarial attacks on WiFi sensing at ACM MobiCom 2024, magnetic-guided medical robotics at IEEE ICRA 2024, embedded firmware for IoT radio systems at Jabil, and Android/Kotlin teaching at Temple University.
  My Projects
PSMGD
First-author AAAI 2025 paper on Periodic Stochastic Multi-Gradient Descent for fast stochastic multi-objective optimization.
TLV-CoRe
Co-first-author AAAI 2026 work on tactile-language-vision collaborative representation learning for robust multimodal robotic perception.
InherNet
Co-first-author ICLR 2026 work on neural network inheritance through asymmetric low-rank reconstruction beyond standard knowledge distillation.
A3B2
IJCAI-ECAI 2026 collaboration on adaptive asymmetric adapters for few-shot vision-language image classification under branch bias.
Personal Website
Published build of this Hexo + Matery personal website, including research notes, study logs, and personal writing.
  My Skills
AI Optimization
Trustworthy AI
Multimodal Learning
Python / PyTorch
Embedded Systems
Academic Reviewing