Scribd
Upload
2 views20 pages

Mini Project Final PPT

This document outlines the development of a Conversational AI system designed to assist students with college-related inquiries in real time using Natural Language Processing (NLP). The system aims to enhance accessibility, reduce administrative workload, and provide instant, accurate responses through an intuitive interface. Key features include integration with a local large language model, context-aware conversations, and a user-friendly web interface built with Streamlit.

Uploaded by

sasank.kodamarthy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
2 views20 pages

Mini Project Final PPT

This document outlines the development of a Conversational AI system designed to assist students with college-related inquiries in real time using Natural Language Processing (NLP). The system aims to enhance accessibility, reduce administrative workload, and provide instant, accurate responses through an intuitive interface. Key features include integration with a local large language model, context-aware conversations, and a user-friendly web interface built with Streamlit.

Uploaded by

sasank.kodamarthy
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 20

Abstract

This project presents a


Conversational AI system designed
to handle college-related enquiries
in real time. Using Natural
Language Processing (NLP), the
chatbot understands and responds to
student queries about courses,
admissions, facilities, and more. The
system aims to improve
accessibility, reduce manual
workload, and provide instant,
accurate information to users
through an intuitive interface.
Introduction
In recent years, Conversational
AI has emerged as a powerful
tool for automating information
services across various sectors,
including education. Colleges
often receive a high volume of
repetitive queries from students
regarding admissions, course
offerings, eligibility, and
facilities. Addressing these
manually can be time-
consuming and inefficient.
LITERATURE REVIEW
Paper Author Title Method Used Demerit

Rule-based Limited to
approach using predefined
A. Hussain, N. Chatbot for University Enquiry AIML (Artificial responses; lacks
1 Ameen System Intelligence contextual
Markup understanding
Language)

Intent
recognition via Depends heavily on
2 S. Patel, D. Mehta College Enquiry Bot using NLP and internet and
Dialogflow Dialogflow predefined intents
integration

ML-based intent Poor handling of


R. Sharma, K. AI-based Virtual Assistant for classification ambiguous queries
3 Gupta Student Helpdesk with voice and lacks
support personalization
Existing System

• College enquiry systems often rely on static FAQs or manual help desks, which are not
scalable.
• Response delays and limited availability of staff reduce the efficiency of student support.
• Existing chatbots in some institutions are rule-based and fail to handle natural, varied
queries.
• Many current systems lack contextual understanding and cannot maintain a meaningful
conversation.
• Language support is often limited to English, excluding students more comfortable with
regional languages.
• Most tools do not provide personalized answers based on specific user data (e.g., course
preferences or eligibility).
• There is little to no integration with college databases, making real-time and accurate
responses difficult.
• No support for voice-based interaction, which could enhance accessibility for visually
impaired users.
System Architecture
UML Class Diagram
Use Case Diagram
Sequence Diagram
Collaboration Diagram
Activity
Diagram
Data Flow Diagram
CODE IMPLEMENTATION
To deploy the Conversational AI for College Enquiries chatbot, we utilized Streamlit
for creating a simple, interactive web interface, and LangChain for managing
memory and chaining prompts with a large language model (LLM).

Frontend:
• Streamlit is used to develop a web-based chat interface.
• It allows real-time interaction, session management, and dynamic rendering of
messages.
Backend & LLM:

• The backend is powered by LangChain, integrated with LlamaCpp,


a lightweight LLM interface for local model inference.
• The model used is Mistral-7B-Instruct, downloaded from Hugging
Face and loaded using the hf_hub_download() function.
Custom Chain Construction:
• Inputs and conversation history are passed through a series of LangChain components:
• ChatPromptTemplate to structure prompts
• RunnableLambda and RunnablePassThrough for chaining and post-processing.
SYSTEM TESTING
1. Functional Testing
Verified that the chatbot could handle various types of student
queries (e.g., courses offered, admission process, eligibility
criteria).
2. Interface Testing
• Ensured seamless interaction between the Streamlit
interface and the LangChain-powered backend.
• Verified dynamic message rendering and user input
handling through st.chat_input() and st.chat_message().
3. Model Testing
• Evaluated the quality and relevance of responses from the
Mistral-7B-Instruct model
• Checked for accurate prompt formatting and proper use of
the system prompt to guide the chatbot's tone and domain
specificity
4. Performance Testing
• Measured response generation times under different loads.
• Confirmed that the chatbot performs efficiently even with
memory accumulation across extended conversations.
OUTPUT SCREENSHOTS
CONCLUSION
The Conversational AI for College Enquiries project successfully streamlines student
interaction with academic institutions by providing an intelligent, real-time chatbot that
responds to college-related queries using Natural Language Processing (NLP).

By integrating a local large language model (LLM) with a user-friendly Streamlit interface, the
system offers:
• Instant responses to admission, course, and facility-related questions,
• Context-aware conversation with memory retention, and
• A customizable and easily deployable chatbot framework.

This solution reduces the workload on administrative staff, improves accessibility for
students, and ensures consistent, accurate information delivery.

You might also like