A data auditing tool based on explainable artificial intelligence
With the rapid growth of Information Technology and the number of internet users increasing each year, the amount of data being generated each day on the internet is increasing. Artificial intelligence requires enormous amounts of data to learn and improve its decision-making process. The incr...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
Nanyang Technological University
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/156661 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | With the rapid growth of Information Technology and the number of internet users
increasing each year, the amount of data being generated each day on the internet is
increasing. Artificial intelligence requires enormous amounts of data to learn and
improve its decision-making process. The increase of data and data is much more
available, this leads to the popularity of data-hungry technologies such as artificial
intelligence. It is estimated that the market for artificial intelligence is $62 billion in 2020
and is expected to expand at a compound annual growth rate of 40.2% from 2021 to
2028 [1].
In recent years, artificial intelligence has impacted the daily lives of many. Artificial
intelligence is everywhere. For Example, using Face ID to unlock the smartphone, using
google search, Recommendation Systems such as Amazon product recommendations
and Netflix show recommendations and self-driving cars [2].
However not all data are treated evenly, some data are more useful than others which
helps to improve artificial intelligence decision-making processes while others might
worsen artificial intelligence decision-making processes. By using explainable artificial
intelligence, Hypergradient Data Relevance Analysis for Interpreting Deep Neural
Network (HYDRA), we can gauge the usefulness of the data. A higher positive
contribution result indicates that the data is useful in the training of its decision-making
processes while the lowest negative contribution indicates that the data worsens its
decision-making processes when training [3].
The objective of this project is to demonstrate the use of Hypergradient Data Relevance
Analysis for Interpreting Deep Neural networks using Web applications |
|---|