Research Articles

My published research in machine learning, artificial intelligence, and related fields, contributing to the scientific community and advancing technological innovation.

Nature

Adaptive Neural Architectures for Resource-Constrained Environments

Published June 2023

Deep Learning Model Optimization Edge Computing 42 citations

This paper presents a novel approach to neural network architecture adaptation for deployment in resource-constrained environments. We introduce a dynamic pruning methodology that automatically adjusts model complexity based on available computational resources while maintaining high accuracy.

ACL

Cross-Lingual Transfer Learning for Low-Resource Languages Using Transformer Architectures

Published May 2023

NLP Transformers Transfer Learning 28 citations

We propose a novel approach to cross-lingual transfer learning that significantly improves natural language understanding for low-resource languages. By leveraging shared linguistic features across related languages, our method achieves state-of-the-art performance on several benchmark tasks.

CVPR

Self-Supervised Visual Representation Learning Through Geometric Transformations

Published December 2022

Computer Vision Self-Supervised Learning Representation Learning 35 citations

This paper introduces a novel approach to self-supervised visual representation learning by exploiting geometric transformations. Our method enables models to learn robust visual features without labeled data, achieving comparable performance to supervised approaches on various downstream tasks.

Elsevier

Fairness and Accountability Frameworks for Deployed AI Systems: A Practical Approach

Published October 2022

AI Ethics Algorithmic Fairness Policy 19 citations

This research proposes practical frameworks for ensuring fairness, accountability, and transparency in deployed AI systems. We present a comprehensive methodology for auditing algorithms and mitigating bias, along with case studies demonstrating successful implementation in high-stakes domains.

NeurIPS

Probabilistic Graph Neural Networks for Uncertainty Quantification in Molecular Property Prediction

Published December 2022

Graph Neural Networks Bayesian Methods Drug Discovery 23 citations

We introduce a probabilistic extension to graph neural networks that enables accurate uncertainty quantification in molecular property prediction. Our approach provides reliable confidence estimates alongside predictions, crucial for high-stakes applications like drug discovery and materials science.

EMNLP

Detecting and Mitigating Language Model Biases: A Comparative Analysis of Debiasing Techniques

Published November 2021

NLP AI Ethics Bias Mitigation 31 citations

This study presents a comprehensive analysis of bias detection and mitigation techniques for large language models. We evaluate various debiasing approaches across different bias types and propose an ensemble method that outperforms existing techniques while maintaining model utility.

Publication Impact

15+

Research Papers

250+

Citations

8

Top Conferences

Research Collaborations

Stanford University
MIT
University of Toronto
Google Research
DeepMind