Online Training of Large Language Models: Learn while chatting

TL;DR

This paper proposes a new online training paradigm for Large Language Models that enables real-time, persistent updates and customization through external interactions, addressing current limitations in flexibility and user accessibility.

Contribution

It introduces a novel interaction paradigm that combines persistent online training with external knowledge sources for improved LLM customization.

Findings

Enables real-time model updates during user interactions

Allows personalized model customization via external knowledge bases

Improves flexibility and user accessibility in LLM training

Abstract

Large Language Models(LLMs) have dramatically revolutionized the field of Natural Language Processing(NLP), offering remarkable capabilities that have garnered widespread usage. However, existing interaction paradigms between LLMs and users are constrained by either inflexibility, limitations in customization, or a lack of persistent learning. This inflexibility is particularly evident as users, especially those without programming skills, have restricted avenues to enhance or personalize the model. Existing frameworks further complicate the model training and deployment process due to their computational inefficiencies and lack of user-friendly interfaces. To overcome these challenges, this paper introduces a novel interaction paradigm-'Online Training using External Interactions'-that merges the benefits of persistent, real-time model updates with the flexibility for individual customization through external interactions such as AI agents or online/offline knowledge bases.