Morteza Mahdiani

I am a Master's student in Computer Science at the University of Montreal and a researcher at Mila - Quebec AI Institute, currently working in the Charest Lab. My research focuses on multimodal learning, vision-language models, and brain encoding and decoding, with a specific interest in comparing AI models with brain data. Previously, I served as a research assistant at Mila and the University of Montreal, where one of my projects contributed to the development of the platform cogcompneuro.com.

Contact:
Email: morteza.mahdiani@umontreal.ca
Homepage: morteza-mahdiani.github.io

LinkedIn  /  Scholar  /  GitHub  /  X

profile photo

Selected Research Contributions

paper1
Few-Shot Learning Fails: Limits of In-Context Reasoning in VLMs
TBD
VLMs reasoning with minimal visual changes.
paper2
VISGate: Dual-Head Encoders for Brain Alignment
Neurips 2025 Workshop
Image-to-brain encoding with LLM embeddings.
paper3
Graph-Guided Prompting for Multi-Hop QG
ICML 2025 Workshop
Zero-shot multi-hop QG with LLMs.
paper4
CLIP Predicts Object Knowledge
iScience 2024
CLIP-ViT predicts object dimensions.
paper5
Multimodal Neuroimaging Dataset
bioRxiv 2022
fMRI & EEG for object recognition.