Chat Generative Pre-Trained Transformer (ChatGPT): Comprehending its Operational Structure, AI Techniques, Working, Features and Limitations

Ishita Naik; Dishita Naik; Nitin Naik

doi:10.1109/ICTBIG59752.2023.10456201

Chat Generative Pre-Trained Transformer (ChatGPT): Comprehending its Operational Structure, AI Techniques, Working, Features and Limitations

Ishita Naik, Dishita Naik, Nitin Naik

Software Engineering & Cybersecurity

Research output: Chapter in Book/Published conference output › Conference publication

Abstract

ChatGPT has transformed the way machines interact with humans, as it is a significant improvement in the field of generative AI and conversational AI. ChatGPT is a combination of generative AI and conversational AI, and it is built on an underlying large language model which can create new contents similar to human-generated contents and simulate human-like conversations. It has the ability to understand the context of a conversation and respond accordingly, and learn to adapt from user feedback. ChatGPT is a new AI tool and is still evolving very rapidly, its successive models are gradually becoming more advanced and are capable of generating highly complex, coherent and credible text, thus making it a valuable tool for a wide range of applications. Therefore, this paper will present a comprehensive study of ChatGPT covering its operational structure, key AI techniques, working, features and limitations. This comprehensive study will provide users a detailed insight of ChatGPT in order to understand its core concepts.

Original language	English
Title of host publication	2023 IEEE International Conference on ICT in Business Industry & Government (ICTBIG)
Publisher	IEEE
ISBN (Electronic)	9798350343274
ISBN (Print)	9798350343281
DOIs	https://doi.org/10.1109/ICTBIG59752.2023.10456201
Publication status	Published - 19 Mar 2024
Event	IEEE 3rd International Conference on ICT in Business Industry & Government - Indore, India Duration: 8 Dec 2023 → 9 Dec 2023

Conference

Conference	IEEE 3rd International Conference on ICT in Business Industry & Government
Abbreviated title	ICTBIG 2023
Country/Territory	India
City	Indore
Period	8/12/23 → 9/12/23

Keywords

ChatGPT
Generative Pre-training Transformer
GPT
LLM
Large Language Model
Reinforcement Learning from Human Feedback
RLHF
SFT
Supervised Fine-Tuning
NLP
Natural Language Processing
Generative AI,
Conversational AI
PPO
Proximal Policy Optimization
Unsupervised Pre-training
Reward Model

Access to Document

10.1109/ICTBIG59752.2023.10456201

Cite this

@inproceedings{c0d7d04fdbc74b7a845c78ff5ebf0288,

title = "Chat Generative Pre-Trained Transformer (ChatGPT): Comprehending its Operational Structure, AI Techniques, Working, Features and Limitations",

abstract = "ChatGPT has transformed the way machines interact with humans, as it is a significant improvement in the field of generative AI and conversational AI. ChatGPT is a combination of generative AI and conversational AI, and it is built on an underlying large language model which can create new contents similar to human-generated contents and simulate human-like conversations. It has the ability to understand the context of a conversation and respond accordingly, and learn to adapt from user feedback. ChatGPT is a new AI tool and is still evolving very rapidly, its successive models are gradually becoming more advanced and are capable of generating highly complex, coherent and credible text, thus making it a valuable tool for a wide range of applications. Therefore, this paper will present a comprehensive study of ChatGPT covering its operational structure, key AI techniques, working, features and limitations. This comprehensive study will provide users a detailed insight of ChatGPT in order to understand its core concepts.",

keywords = "ChatGPT, Generative Pre-training Transformer, GPT, LLM, Large Language Model, Reinforcement Learning from Human Feedback, RLHF, SFT, Supervised Fine-Tuning, NLP, Natural Language Processing, Generative AI,, Conversational AI, PPO, Proximal Policy Optimization, Unsupervised Pre-training, Reward Model",

author = "Ishita Naik and Dishita Naik and Nitin Naik",

year = "2024",

month = mar,

day = "19",

doi = "10.1109/ICTBIG59752.2023.10456201",

language = "English",

isbn = "9798350343281",

booktitle = "2023 IEEE International Conference on ICT in Business Industry & Government (ICTBIG)",

publisher = "IEEE",

address = "United States",

note = "IEEE 3rd International Conference on ICT in Business Industry & Government, ICTBIG 2023 ; Conference date: 08-12-2023 Through 09-12-2023",

}

Naik, I, Naik, D & Naik, N 2024, Chat Generative Pre-Trained Transformer (ChatGPT): Comprehending its Operational Structure, AI Techniques, Working, Features and Limitations. in 2023 IEEE International Conference on ICT in Business Industry & Government (ICTBIG). IEEE, IEEE 3rd International Conference on ICT in Business Industry & Government, Indore, India, 8/12/23. https://doi.org/10.1109/ICTBIG59752.2023.10456201

Chat Generative Pre-Trained Transformer (ChatGPT): Comprehending its Operational Structure, AI Techniques, Working, Features and Limitations. / Naik, Ishita; Naik, Dishita; Naik, Nitin.
2023 IEEE International Conference on ICT in Business Industry & Government (ICTBIG). IEEE, 2024.

Research output: Chapter in Book/Published conference output › Conference publication

TY - GEN

T1 - Chat Generative Pre-Trained Transformer (ChatGPT): Comprehending its Operational Structure, AI Techniques, Working, Features and Limitations

AU - Naik, Ishita

AU - Naik, Dishita

AU - Naik, Nitin

PY - 2024/3/19

Y1 - 2024/3/19

N2 - ChatGPT has transformed the way machines interact with humans, as it is a significant improvement in the field of generative AI and conversational AI. ChatGPT is a combination of generative AI and conversational AI, and it is built on an underlying large language model which can create new contents similar to human-generated contents and simulate human-like conversations. It has the ability to understand the context of a conversation and respond accordingly, and learn to adapt from user feedback. ChatGPT is a new AI tool and is still evolving very rapidly, its successive models are gradually becoming more advanced and are capable of generating highly complex, coherent and credible text, thus making it a valuable tool for a wide range of applications. Therefore, this paper will present a comprehensive study of ChatGPT covering its operational structure, key AI techniques, working, features and limitations. This comprehensive study will provide users a detailed insight of ChatGPT in order to understand its core concepts.

AB - ChatGPT has transformed the way machines interact with humans, as it is a significant improvement in the field of generative AI and conversational AI. ChatGPT is a combination of generative AI and conversational AI, and it is built on an underlying large language model which can create new contents similar to human-generated contents and simulate human-like conversations. It has the ability to understand the context of a conversation and respond accordingly, and learn to adapt from user feedback. ChatGPT is a new AI tool and is still evolving very rapidly, its successive models are gradually becoming more advanced and are capable of generating highly complex, coherent and credible text, thus making it a valuable tool for a wide range of applications. Therefore, this paper will present a comprehensive study of ChatGPT covering its operational structure, key AI techniques, working, features and limitations. This comprehensive study will provide users a detailed insight of ChatGPT in order to understand its core concepts.

KW - ChatGPT

KW - Generative Pre-training Transformer

KW - GPT

KW - LLM

KW - Large Language Model

KW - Reinforcement Learning from Human Feedback

KW - RLHF

KW - SFT

KW - Supervised Fine-Tuning

KW - NLP

KW - Natural Language Processing

KW - Generative AI,

KW - Conversational AI

KW - PPO

KW - Proximal Policy Optimization

KW - Unsupervised Pre-training

KW - Reward Model

UR - https://ieeexplore.ieee.org/document/10456201

UR - http://www.scopus.com/inward/record.url?scp=85189245262&partnerID=8YFLogxK

U2 - 10.1109/ICTBIG59752.2023.10456201

DO - 10.1109/ICTBIG59752.2023.10456201

M3 - Conference publication

SN - 9798350343281

BT - 2023 IEEE International Conference on ICT in Business Industry & Government (ICTBIG)

PB - IEEE

T2 - IEEE 3rd International Conference on ICT in Business Industry & Government

Y2 - 8 December 2023 through 9 December 2023

ER -

Chat Generative Pre-Trained Transformer (ChatGPT): Comprehending its Operational Structure, AI Techniques, Working, Features and Limitations

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this