Publications & Talks
Research papers, articles, and conference presentations
Publications
Deep Learning Approaches for Time Series Forecasting in Financial Markets
Mehran P., et al.
International Conference on Machine Learning (ICML)
This paper presents a novel approach to time series forecasting using attention-based LSTM networks for financial market prediction.
Efficient Neural Architecture Search for Computer Vision Tasks
Mehran P., Smith J., Johnson A.
Journal of Artificial Intelligence Research
We propose an efficient NAS method that reduces search time by 70% while maintaining state-of-the-art performance on ImageNet.
Transfer Learning for Low-Resource Natural Language Processing
Mehran P., Brown K.
Conference on Empirical Methods in NLP (EMNLP)
A comprehensive study on transfer learning techniques for NLP tasks with limited training data.
Explainable AI: Interpretability in Deep Learning Models
Mehran P.
AI Magazine
An overview of current techniques for interpreting and explaining deep learning model predictions.
Talks & Presentations
Building Production-Ready ML Systems
2023PyData Conference
Virtual
Discussed best practices for deploying machine learning models in production environments.
The Future of NLP: Transformers and Beyond
2023AI Summit
San Francisco, CA
Explored recent advances in natural language processing and future research directions.
Data Science for Social Good
2022DataCon
New York, NY
Presented case studies on using data science to solve social and environmental challenges.