Applying interpolation-constrained autoencoders to world models approach reinforcement learning
World Models Approach Reinforcement Learning helps to tackle complex problems by breaking down the learning task to Vision Model, Memory Model, and Controller Model. Variational Autoencoder (VAE) is commonly used for Vision Model. However, there has been development of other variants of Autoencoders...
| Autor principal: | |
|---|---|
| Outros Autores: | |
| Formato: | Final Year Project (FYP) |
| Idioma: | English |
| Publicado em: |
Nanyang Technological University
2021
|
| Assuntos: | |
| Acesso em linha: | https://hdl.handle.net/10356/148004 |
| _version_ | 1826117071361015808 |
|---|---|
| author | Kevin Winata |
| author2 | Zinovi Rabinovich |
| author_facet | Zinovi Rabinovich Kevin Winata |
| author_sort | Kevin Winata |
| collection | NTU |
| description | World Models Approach Reinforcement Learning helps to tackle complex problems by breaking down the learning task to Vision Model, Memory Model, and Controller Model. Variational Autoencoder (VAE) is commonly used for Vision Model. However, there has been development of other variants of Autoencoders, with one prominent example is ACAI (Adversarially Constrained Autoencoder Interpolations). In this paper, we propose to substitute ACAI to VAE which might improve the performance on Open AI Car Racing environment. Unfortunately, the ingenuity of ACAI does not apply well to the Car Racing environment because of how ACAI is modeled. |
| first_indexed | 2024-10-01T04:21:45Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/148004 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T04:21:45Z |
| publishDate | 2021 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1480042021-04-22T03:05:16Z Applying interpolation-constrained autoencoders to world models approach reinforcement learning Kevin Winata Zinovi Rabinovich School of Computer Science and Engineering Computational Intelligence Lab zinovi@ntu.edu.sg Engineering::Computer science and engineering World Models Approach Reinforcement Learning helps to tackle complex problems by breaking down the learning task to Vision Model, Memory Model, and Controller Model. Variational Autoencoder (VAE) is commonly used for Vision Model. However, there has been development of other variants of Autoencoders, with one prominent example is ACAI (Adversarially Constrained Autoencoder Interpolations). In this paper, we propose to substitute ACAI to VAE which might improve the performance on Open AI Car Racing environment. Unfortunately, the ingenuity of ACAI does not apply well to the Car Racing environment because of how ACAI is modeled. Bachelor of Engineering (Computer Science) 2021-04-22T03:05:16Z 2021-04-22T03:05:16Z 2021 Final Year Project (FYP) Kevin Winata (2021). Applying interpolation-constrained autoencoders to world models approach reinforcement learning. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/148004 https://hdl.handle.net/10356/148004 en SCSE20-0486 application/pdf Nanyang Technological University |
| spellingShingle | Engineering::Computer science and engineering Kevin Winata Applying interpolation-constrained autoencoders to world models approach reinforcement learning |
| title | Applying interpolation-constrained autoencoders to world models approach reinforcement learning |
| title_full | Applying interpolation-constrained autoencoders to world models approach reinforcement learning |
| title_fullStr | Applying interpolation-constrained autoencoders to world models approach reinforcement learning |
| title_full_unstemmed | Applying interpolation-constrained autoencoders to world models approach reinforcement learning |
| title_short | Applying interpolation-constrained autoencoders to world models approach reinforcement learning |
| title_sort | applying interpolation constrained autoencoders to world models approach reinforcement learning |
| topic | Engineering::Computer science and engineering |
| url | https://hdl.handle.net/10356/148004 |
| work_keys_str_mv | AT kevinwinata applyinginterpolationconstrainedautoencoderstoworldmodelsapproachreinforcementlearning |