Published:
A gradient-regularized deep learning framework is proposed for inverse airfoil design, enforcing geometric smoothness through spatial derivatives while preserving aerodynamic accuracy.
Published:
This paper introduces IndiCASA, a comprehensive dataset and evaluation framework for measuring social biases in Large Language Models specifically tailored for the Indian context using contrastive embedding similarity.
Recommended citation: Santhosh G S, Akshay Govind Srinivasan, et al. (2025). "IndiCASA: A Dataset and Bias Evaluation Framework for LLMs Using Contrastive Embedding Similarity in the Indian Context." Eighth AAAI/ACM Conference on AI, Ethics, and Society. (Accepted).
Published:
This patent covers innovative data generation techniques specifically designed for tool-based Large Language Model applications, enabling scalable and efficient training data creation.
Recommended citation: Akshay Govind Srinivasan, Sahil Girhepuje, Siva Sankar Sajeev et al. (2024). "Data Generation Techniques for Tool-Based LLM Applications." Patent Application Number: 202441079904.
Published:
This patent describes systems and methods for implementing agentic approaches in retrieval-augmented generation, enhancing AI system capabilities through intelligent agent-based architectures.
Recommended citation: Akshay Govind Srinivasan, Ryan Jacob George et al. (2025). "Systems and Methods for Agentic Based Approach in Retrieval Augmented Generation." Patent Application Number: 202541024879.
Published:
This paper presents a novel approach to enhance financial retrieval-augmented generation using agentic AI systems and Multi-HyDE methodology to reduce hallucinations and improve information retrieval accuracy.
Recommended citation: Akshay Govind Srinivasan, Ryan George et al. (2025). "Enhancing Financial RAG using Agentic AI and Multi-HyDE: A Novel Approach to Information Retrieval and Hallucination Reduction." Conference on Language Modelling. (Under Review).
Published:
This patent describes systems and methods for implementing agentic approaches in retrieval-augmented generation, enhancing AI system capabilities through intelligent agent-based architectures.
Recommended citation: Akshay Govind Srinivasan, Ryan Jacob George et al. (2025). "Systems and Methods for Agentic Based Approach in Retrieval Augmented Generation." Patent Application Number: 202541024879.
Published:
This paper proposes RE-GAINS and EnChAnT, two novel frameworks that empower Large Language Models to tackle complex user queries through intelligent tool invocation and chaining using external APIs.
Recommended citation: Sahil Girhepuje, Siva Sankar Sajeev, Akshay Govind Srinivasan, et al. (2024). "RE-GAINS & EnChAnT: Intelligent Tool Manipulation Systems For Enhanced Query Responses." arXiv preprint. arXiv:2401.15724. https://arxiv.org/pdf/2401.15724.pdf