Early goal-detection for black-box environment poisoning attacks
This project aimed to develop a Goal Recognition (GR) system to enhance the applicability of existing Environment Poisoning Attacks (EPAs). Specifically, it focused on increasing the early accuracy and robustness against environmental changes for an existing GR solution. An architecture based on the...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2023
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/165988 |
| _version_ | 1826123220850311168 |
|---|---|
| author | Iyengar, Varun Srikant |
| author2 | Zinovi Rabinovich |
| author_facet | Zinovi Rabinovich Iyengar, Varun Srikant |
| author_sort | Iyengar, Varun Srikant |
| collection | NTU |
| description | This project aimed to develop a Goal Recognition (GR) system to enhance the applicability of existing Environment Poisoning Attacks (EPAs). Specifically, it focused on increasing the early accuracy and robustness against environmental changes for an existing GR solution. An architecture based on the Machine Theory of Mind (ToMnet) was implemented and trained on a large dataset of an EPA-specific environment. Two evaluation techniques were proposed to test the quality of character embeddings and predictions. The model was evaluated on 10,000 trajectories of Q-learning agents of varying optimality on randomly generated environments. The model provided weak evaluation results when trained on a basic algorithm. Hence, two improved training algorithms were proposed to combat local overfitting and improve the model's ability to differentiate between goals. These training algorithms increased the model's evaluation accuracy by over 10\% and successfully provided its top predictions as early as 200 victim epochs. Evaluation performances were analyzed to investigate ties between model performance and train data, trajectory quality, and environment patterns that may be points of failure for the current model. Overall, the project successfully sets a baseline score for early goal detection without any knowledge of environment dynamics and agent policy provided to the model. |
| first_indexed | 2024-10-01T06:01:06Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/165988 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T06:01:06Z |
| publishDate | 2023 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1659882023-04-21T15:38:04Z Early goal-detection for black-box environment poisoning attacks Iyengar, Varun Srikant Zinovi Rabinovich School of Computer Science and Engineering Computational Intelligence Lab zinovi@ntu.edu.sg Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence This project aimed to develop a Goal Recognition (GR) system to enhance the applicability of existing Environment Poisoning Attacks (EPAs). Specifically, it focused on increasing the early accuracy and robustness against environmental changes for an existing GR solution. An architecture based on the Machine Theory of Mind (ToMnet) was implemented and trained on a large dataset of an EPA-specific environment. Two evaluation techniques were proposed to test the quality of character embeddings and predictions. The model was evaluated on 10,000 trajectories of Q-learning agents of varying optimality on randomly generated environments. The model provided weak evaluation results when trained on a basic algorithm. Hence, two improved training algorithms were proposed to combat local overfitting and improve the model's ability to differentiate between goals. These training algorithms increased the model's evaluation accuracy by over 10\% and successfully provided its top predictions as early as 200 victim epochs. Evaluation performances were analyzed to investigate ties between model performance and train data, trajectory quality, and environment patterns that may be points of failure for the current model. Overall, the project successfully sets a baseline score for early goal detection without any knowledge of environment dynamics and agent policy provided to the model. Bachelor of Science in Data Science and Artificial Intelligence 2023-04-18T05:26:28Z 2023-04-18T05:26:28Z 2023 Final Year Project (FYP) Iyengar, V. S. (2023). Early goal-detection for black-box environment poisoning attacks. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/165988 https://hdl.handle.net/10356/165988 en SCSE22-0146 application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence Iyengar, Varun Srikant Early goal-detection for black-box environment poisoning attacks |
| title | Early goal-detection for black-box environment poisoning attacks |
| title_full | Early goal-detection for black-box environment poisoning attacks |
| title_fullStr | Early goal-detection for black-box environment poisoning attacks |
| title_full_unstemmed | Early goal-detection for black-box environment poisoning attacks |
| title_short | Early goal-detection for black-box environment poisoning attacks |
| title_sort | early goal detection for black box environment poisoning attacks |
| topic | Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence |
| url | https://hdl.handle.net/10356/165988 |
| work_keys_str_mv | AT iyengarvarunsrikant earlygoaldetectionforblackboxenvironmentpoisoningattacks |