Deep Learning Speech-Text Chatbot for High School Advising
High school is a crucial stage for students as they begin to shape their future. During this time, students need to identify their strengths and interests, choose the right curriculum, and prepare for university applications, including admission tests and selecting majors. However, not all high scho...
Saved in:
| Main Author: | |
|---|---|
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://bspace.buid.ac.ae/handle/1234/2765 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1862980612383571968 |
|---|---|
| author | Assayed, Suha Khalil |
| author_facet | Assayed, Suha Khalil |
| author_role | author |
| dc.contributor.none.fl_str_mv | Dr Manar AlKhatib |
| dc.creator.none.fl_str_mv | Assayed, Suha Khalil |
| dc.date.none.fl_str_mv | 2024-11 2025-01-28T08:50:58Z 2025-01-28T08:50:58Z |
| dc.format.none.fl_str_mv | application/pdf |
| dc.identifier.none.fl_str_mv | 21002774 https://bspace.buid.ac.ae/handle/1234/2765 |
| dc.language.none.fl_str_mv | en |
| dc.publisher.none.fl_str_mv | The British University in Dubai (BUiD) |
| dc.subject.none.fl_str_mv | ASR, chatbot, high school, LSTM, college-career guidance, speech, Seq2Seq, ChatGPT |
| dc.title.none.fl_str_mv | Deep Learning Speech-Text Chatbot for High School Advising |
| dc.type.none.fl_str_mv | Thesis |
| description | High school is a crucial stage for students as they begin to shape their future. During this time, students need to identify their strengths and interests, choose the right curriculum, and prepare for university applications, including admission tests and selecting majors. However, not all high schools can afford college-career advisers to assist students with these important decisions, leaving some students less prepared for their future compared to those who receive guidance. This thesis addresses the challenges faced by both students and advisers in schools, proposing a novel, affordable bilingual speech-text chatbot designed to provide equal support to all high school students, including those in underprivileged schools. We explored various deep neural network models to determine the most effective model for this task. The proposed architecture integrates an encoder-encoder framework with different layers of deep recurrent neural networks (RNN), such as LSTM, BiLSTM, and stacked LSTM layers. Additionally, automatic speech recognition (ASR) is incorporated to convert spoken inquiries into text, allowing the chatbot to generate effective responses. Evaluation using the ROUGE metric showed that the BiLSTM layer achieved the highest performance, particularly in precision. A qualitative study comparing our chatbot (HSGAdviser) with ChatGPT revealed that students preferred our chatbot, especially for Yes/No questions, demonstrating its potential to provide equal, accessible advising for all students. |
| id | budr_fd4f73aa58b6a4800b3e7741051076fa |
| identifier_str_mv | 21002774 |
| language_invalid_str_mv | en |
| network_acronym_str | budr |
| network_name_str | The British University in Dubai repository |
| oai_identifier_str | oai:bspace.buid.ac.ae:1234/2765 |
| publishDate | 2024 |
| publisher.none.fl_str_mv | The British University in Dubai (BUiD) |
| repository.mail.fl_str_mv | |
| repository.name.fl_str_mv | |
| repository_id_str | |
| spelling | Deep Learning Speech-Text Chatbot for High School AdvisingAssayed, Suha KhalilASR, chatbot, high school, LSTM, college-career guidance, speech, Seq2Seq, ChatGPTHigh school is a crucial stage for students as they begin to shape their future. During this time, students need to identify their strengths and interests, choose the right curriculum, and prepare for university applications, including admission tests and selecting majors. However, not all high schools can afford college-career advisers to assist students with these important decisions, leaving some students less prepared for their future compared to those who receive guidance. This thesis addresses the challenges faced by both students and advisers in schools, proposing a novel, affordable bilingual speech-text chatbot designed to provide equal support to all high school students, including those in underprivileged schools. We explored various deep neural network models to determine the most effective model for this task. The proposed architecture integrates an encoder-encoder framework with different layers of deep recurrent neural networks (RNN), such as LSTM, BiLSTM, and stacked LSTM layers. Additionally, automatic speech recognition (ASR) is incorporated to convert spoken inquiries into text, allowing the chatbot to generate effective responses. Evaluation using the ROUGE metric showed that the BiLSTM layer achieved the highest performance, particularly in precision. A qualitative study comparing our chatbot (HSGAdviser) with ChatGPT revealed that students preferred our chatbot, especially for Yes/No questions, demonstrating its potential to provide equal, accessible advising for all students.The British University in Dubai (BUiD)Dr Manar AlKhatib2025-01-28T08:50:58Z2025-01-28T08:50:58Z2024-11Thesisapplication/pdf21002774https://bspace.buid.ac.ae/handle/1234/2765enoai:bspace.buid.ac.ae:1234/27652025-01-28T23:00:20Z |
| spellingShingle | Deep Learning Speech-Text Chatbot for High School Advising Assayed, Suha Khalil ASR, chatbot, high school, LSTM, college-career guidance, speech, Seq2Seq, ChatGPT |
| title | Deep Learning Speech-Text Chatbot for High School Advising |
| title_full | Deep Learning Speech-Text Chatbot for High School Advising |
| title_fullStr | Deep Learning Speech-Text Chatbot for High School Advising |
| title_full_unstemmed | Deep Learning Speech-Text Chatbot for High School Advising |
| title_short | Deep Learning Speech-Text Chatbot for High School Advising |
| title_sort | Deep Learning Speech-Text Chatbot for High School Advising |
| topic | ASR, chatbot, high school, LSTM, college-career guidance, speech, Seq2Seq, ChatGPT |
| url | https://bspace.buid.ac.ae/handle/1234/2765 |