1. Gambhir, & Gupta, V. (2016).” Recent automatic text summarization techniques: a survey”, The Artificial Intelligence Review, 47(1), 1–66., [Available online]
https://doi.org/10.1007/s10462-016-9475-9 [Accessed: September 14, 2022].
2. Pourvali, & Mohammad, P. D. (2012). “A new graph-based text segmentation using Wikipedia for automatic text summarization. International Journal of
Advanced Computer Science & Applications,
https://doi.org/10.14569/IJACSA.2012.030105 [Accessed: September 12, 2022].
3. Fakhrezi, Bijaksana, M. A., & Huda, A. F. (2021). “Implementation of Automatic Text Summarization with TextRank Method in the Development of Al-Qur’an
Vocabulary Encyclopedia”. Procedia Computer Science, 179, 391–398.
Scalability:
To assess the system's scalability, we ran tests to measure its efficiency under various loads and with varying numbers of users. We ran large-scale usage scenarios and measured reaction times and error rates.
We also looked at how the system performed when dealing with big amounts of data. We used a large dataset of news stories and academic papers to test the system's ability to handle large amounts of data and quickly and correctly produce summaries and mind
maps.
Amazon Cognito was set up to manage user authentication and authorization. This entailed establishing a user pool and assigning user permissions to ensure that users could only access the data and functions to which they were authorized. AWS Lambda was used to build a set of AWS Lambda functions to handle the project's backend processing. These functions, built in JavaScript and TypeScript, were in charge of retrieving user input, generating summaries and mind maps, and storing data in the proper data stores.
Audio and Text Summarization with Graphical
Representation of the Summarized Article
Advisor: Andrew Bond
Introduction
Information is available in abundance on the internet. More information translates to more knowledge. All there is now to consume this information. But with so much information available there isn’t enough time for the reader to go through everything and grasp everything. Time is of the essence and the speed at which readers can consume this abundance of information is a limitation. In most problems when the throughput is limited due to less bandwidth the solution is to try and increase the bandwidth, which isn’t possible in this case. The solution is to maximize the information and minimize the time taken. Summarization is one technique that can help with this.
Project Architecture and Tech Description
Analysis and Results
Summary/Conclusions
Key References
Acknowledgements
This project presents a novel method for document summarization and mind map generation this effort is built on deep learning and NLP methods. Various AWS services, including Cognito, S3, AWS Amplify, Lambda, and DynamoDB, were used to execute
our strategy, and Angular was used to create the front end. This approach produced comparable results in terms of performance and standards, showing its efficacy and
promise for real-world use in the areas of knowledge management and natural language processing.
In conclusion, the developed system is extremely flexible, effective, and simple to use. It can be used in a number of industries, for example in the education industry the system could be used to depict visual information for children to understand broad topics.
The authors are deeply indebted to Professor Andrew Bond for his invaluable comments and assistance in the preparation of this study. | |
| |
A summary needs to present the main idea of a document in fewer words. Therefore, if all the sentences of
a document are of equal importance, then it would be hard to generate a shorter summary without losing important information. Usually, information in documents
appears in bursts and hence makes it possible to identify informative segments.
Project Architecture:
Architecture Subsystems
The system is made up of several architecture subsystems that work together to accomplish the system's overall goals. The system's primary subsystems are as follows:
Frontend:
The front-end subsystem oversees offering an intuitive and user-friendly interface for users to interact with the system. Angular, a popular front-end programming framework, is used to create the front end. The front-end subsystem interacts with the backend subsystems via REST APIs to obtain system summaries and mind maps.
Backend:
For the backend, we are using Node.js and Express for API development to interact with the DynamoDB database. The backend subsystem is in charge of processing user data and producing summaries and mind maps.
The backend subsystem is made up of several components, including summarization, mind map creation, and storage components. The summarization component generates high-quality summaries from the original text using ML techniques.
Authentication and Authorization Subsystem:
The authentication and authorization subsystem handles individual authentication and access management. The system handles user authentication and permission through AWS Cognito, ensuring that only authorized users can view their summaries and mind maps.
Deployment and Scaling Subsystem:
The deployment and scaling subsystem is in charge of handling the system's deployment and scaling on AWS Lambda. AWS Lambda is serverless, event-driven compute service that allows you to run code for almost any type of application or backend service without the need for server provisioning or management.
Project Architecture & Implementation
Speech and Text conversion subsystem:
Amazon Polly is a service that converts text into natural-sounding speech, allowing you to create talking applications and create entirely new categories of speech-enabled
products. Polly's Text-to-Speech (TTS) service synthesizes natural-sounding human speech using advanced deep-learning technologies.
To assess our system's performance, we ran a number of tests and benchmarks to measure its speed, accuracy, and scalability. To guarantee that our findings were representative of the system's performance under various usage scenarios, we used a combination of
synthetic and real-world data.
We used a set of benchmark documents of various lengths and complexities to assess the system's speed and accuracy. Academic papers, news stories, and business reports were among the documents included. We timed the system's generation of a summary and a
mind map for each document and compared it to the durations of other document summarization systems.
We also contrasted the precision of our system's summaries and mind maps to that of
other cutting-edge systems. To do so, we ourselves acted as human evaluators to evaluate the quality of our system's summaries and mind maps, as well as those of other systems. The summaries and mind maps were rated by the evaluators based on variables such as coherence, relevance, and completeness.
The results of these tests revealed that our system was capable of rapidly and correctly generating summaries and mind maps.
Computer Engineering Department
Rajpal, Chirag Biharilal (MS Software Engineering) Shukla, Pratiksha (MS Computer Engineering)
Yadav, Abhishek (MS Software Engineering)
TEMPLATE DESIGN © 2008
www.PosterPresentations.com