The Self-Learning Agent with a Progressive Neural Network Integrated Transformer

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sivakumar, Ajay, Shalini, Raj, Vasantha, Sylvester, Sebastian
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916672852459520
author Sivakumar, Ajay
Shalini
Raj, Vasantha
Sylvester, Sebastian
author_facet Sivakumar, Ajay
Shalini
Raj, Vasantha
Sylvester, Sebastian
contents This paper introduces a self-learning agent that integrates LLaMA 3.2 with a Progressive Neural Network (PNN) for continual learning in conversational AI and code generation. The framework dynamically collects data, fine-tunes tasks with minimal samples, and leverages Meta-Learning for rapid adaptation. LoRA optimizes fine-tuning, while Elastic Weight Consolidation (EWC) enhances knowledge retention. Experimental results demonstrate improved adaptability and memory stability, positioning this approach as a scalable step toward Artificial General Intelligence (AGI).
format Preprint
id arxiv_https___arxiv_org_abs_2504_02489
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle The Self-Learning Agent with a Progressive Neural Network Integrated Transformer
Sivakumar, Ajay
Shalini
Raj, Vasantha
Sylvester, Sebastian
Artificial Intelligence
This paper introduces a self-learning agent that integrates LLaMA 3.2 with a Progressive Neural Network (PNN) for continual learning in conversational AI and code generation. The framework dynamically collects data, fine-tunes tasks with minimal samples, and leverages Meta-Learning for rapid adaptation. LoRA optimizes fine-tuning, while Elastic Weight Consolidation (EWC) enhances knowledge retention. Experimental results demonstrate improved adaptability and memory stability, positioning this approach as a scalable step toward Artificial General Intelligence (AGI).
title The Self-Learning Agent with a Progressive Neural Network Integrated Transformer
topic Artificial Intelligence
url https://arxiv.org/abs/2504.02489