Under Review 2024
Strada-LLM: Graph LLM for Traffic Prediction
17% RMSE improvement for long-term forecasting
Under Review, 2024
01. About Me
I am a 3rd year PhD student (2023–2027) at the Department of Electronics and Informatics (ETRO), Vrije Universiteit Brussel (VUB), under the supervision of Prof. Adrian Munteanu.
My research focuses on advancing spatio-temporal forecasting through deep generative models. I develop methods that leverage diffusion models, flow matching, and large language models to improve probabilistic predictions on graph-structured data—with applications in traffic forecasting and beyond.
Prior to my PhD, I earned my Master’s in Computer Science from Amirkabir University of Technology (Tehran Polytechnic) in 2022.
My work has been published at top venues including NeurIPS Workshops and IEEE MDM, and I actively review for conferences such as KDD, NeurIPS, and ICASSP.
4
Publications
NeurIPS
Top Venue
4
Venues Reviewed
02. Research Interests
Spatio-Temporal Modeling
Modeling complex dependencies across space and time in structured data for accurate forecasting.
Time Series Forecasting
Long-term and short-term prediction methods with state-of-the-art accuracy and efficiency.
Graph Neural Networks
Learning on graph-structured data to capture relational and topological dependencies.
Deep Learning
Neural network architectures for representation learning and generalization in sequential data.
Probabilistic Forecasting
Uncertainty-aware predictions via generative models for robust decision-making.
Diffusion Models
Score-based generative models for high-quality probabilistic forecasting on structured data.
Flow Matching
Efficient continuous normalizing flows for lightweight, fast generative forecasting.
LLMs for Time Series
Leveraging large language models for reasoning and prediction in temporal data domains.
03. Publications
Under Review 2024
GINTRIP: Interpretable Temporal Graph Regression using Information Bottleneck and Prototype-based Method
Under Review (submitted to IEEE), 2024
NeurIPS 2025 Workshop
FreqFlow: Long-term Forecasting using Lightweight Flow Matching
~89K parameters · 7% RMSE improvement
NeurIPS 2025 Workshop (PriGM / SPIGM)
IEEE MDM 2024
STRADA: Spatial-Temporal Dashboard for Traffic Forecasting
IEEE International Conference on Mobile Data Management (MDM), 2024, pp. 251–254
04. Reviewing Experience
Serving as a reviewer for the following venues:
KDD 2025
ACM SIGKDD Conference on Knowledge Discovery and Data Mining
NeurIPS
Conference on Neural Information Processing Systems
ICASSP
IEEE International Conference on Acoustics, Speech and Signal Processing
Transportation Research Part C
Emerging Technologies — Elsevier Journal
05. Education
2023 – 2027 (expected)
PhD in Engineering Sciences
Vrije Universiteit Brussel (VUB)
Department of Electronics and Informatics (ETRO)
Supervisor: Prof. Adrian Munteanu
Research: Spatio-Temporal Modeling, Deep Generative Models, Graph Neural Networks
Brussels, Belgium
2020 – 2022
MSc in Computer Science
Amirkabir University of Technology
Tehran Polytechnic
Tehran, Iran
06. Contact & Links
Feel free to reach out for collaborations, research discussions, or opportunities.