Loading…
Research at the intersection of AI, software engineering and distributed systems.
FaRAG-Guard explores lightweight, explainable hallucination detection in Persian RAG. The manuscript is under review.
This research direction focuses on trustworthy generated answers in Persian. FaRAG-Guard uses XLM-RoBERTa and LoRA for three-way answer verification with extractive evidence.
The manuscript is under review. The work explores hallucination detection in Persian RAG and evidence that helps assess generated answers.
Combine reinforcement learning and graph neural networks to study compute-resource decisions.
This area examines adaptive resource allocation in distributed and cloud environments. RL-GNN studies graph representations alongside reinforcement learning and compares MLP, GCN, GAT and GraphSAGE architectures.
The related DCHPC 2026 paper and the project GitHub repository provide the route into the technical details.
Study LLM training and deployment costs through multi-objective optimization and adaptive methods.
This direction studies the relationship between model quality and runtime, memory and compute cost. Related papers examine multi-objective training optimization and reinforcement learning for language-model deployment.
The AI deployment survey also summarizes techniques, challenges and practices in resource-constrained environments.
Defect prediction, intelligent testing and code evolution connect AI with software quality and performance.
This research collection examines machine learning and deep learning across software development: early fault prediction, code-dependency graph analysis, automated testing and code evolution with language models.
Runtime and memory optimization is another topic in this area. Supplied abstracts and paper files are available through the related links.
Explore device trust and privacy alongside energy optimization in distributed edge environments.
Papers in this area address deep learning for IoT security and trust, blockchain for privacy and machine learning for energy optimization in edge computing.
These are presented as research topics; the description and abstract of each work are available in its own article record.
