Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network
It is crucial to ask how agents can achieve goals by generating action plans using only partial models of the world acquired through habituated sensory-motor experiences. Although many existing robotics studies use a forward model framework, there are generalization issues with high degrees of freed...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2020-05-01
|
| Series: | Entropy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1099-4300/22/5/564 |
| _version_ | 1797567635174457344 |
|---|---|
| author | Takazumi Matsumoto Jun Tani |
| author_facet | Takazumi Matsumoto Jun Tani |
| author_sort | Takazumi Matsumoto |
| collection | DOAJ |
| description | It is crucial to ask how agents can achieve goals by generating action plans using only partial models of the world acquired through habituated sensory-motor experiences. Although many existing robotics studies use a forward model framework, there are generalization issues with high degrees of freedom. The current study shows that the predictive coding (PC) and active inference (AIF) frameworks, which employ a generative model, can develop better generalization by learning a prior distribution in a low dimensional latent state space representing probabilistic structures extracted from well habituated sensory-motor trajectories. In our proposed model, learning is carried out by inferring optimal latent variables as well as synaptic weights for maximizing the evidence lower bound, while goal-directed planning is accomplished by inferring latent variables for maximizing the estimated lower bound. Our proposed model was evaluated with both simple and complex robotic tasks in simulation, which demonstrated sufficient generalization in learning with limited training data by setting an intermediate value for a regularization coefficient. Furthermore, comparative simulation results show that the proposed model outperforms a conventional forward model in goal-directed planning, due to the learned prior confining the search of motor plans within the range of habituated trajectories. |
| first_indexed | 2024-03-10T19:44:53Z |
| format | Article |
| id | doaj.art-804ab67f10bc425b945a03a2b480b0b5 |
| institution | Directory Open Access Journal |
| issn | 1099-4300 |
| language | English |
| last_indexed | 2024-03-10T19:44:53Z |
| publishDate | 2020-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Entropy |
| spelling | doaj.art-804ab67f10bc425b945a03a2b480b0b52023-11-20T00:52:29ZengMDPI AGEntropy1099-43002020-05-0122556410.3390/e22050564Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural NetworkTakazumi Matsumoto0Jun Tani1Okinawa Institute of Science and Technology, Okinawa 904-0495, JapanOkinawa Institute of Science and Technology, Okinawa 904-0495, JapanIt is crucial to ask how agents can achieve goals by generating action plans using only partial models of the world acquired through habituated sensory-motor experiences. Although many existing robotics studies use a forward model framework, there are generalization issues with high degrees of freedom. The current study shows that the predictive coding (PC) and active inference (AIF) frameworks, which employ a generative model, can develop better generalization by learning a prior distribution in a low dimensional latent state space representing probabilistic structures extracted from well habituated sensory-motor trajectories. In our proposed model, learning is carried out by inferring optimal latent variables as well as synaptic weights for maximizing the evidence lower bound, while goal-directed planning is accomplished by inferring latent variables for maximizing the estimated lower bound. Our proposed model was evaluated with both simple and complex robotic tasks in simulation, which demonstrated sufficient generalization in learning with limited training data by setting an intermediate value for a regularization coefficient. Furthermore, comparative simulation results show that the proposed model outperforms a conventional forward model in goal-directed planning, due to the learned prior confining the search of motor plans within the range of habituated trajectories.https://www.mdpi.com/1099-4300/22/5/564goal directed planningactive inferencepredictive codingvariational bayesrecurrent neural network |
| spellingShingle | Takazumi Matsumoto Jun Tani Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network Entropy goal directed planning active inference predictive coding variational bayes recurrent neural network |
| title | Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network |
| title_full | Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network |
| title_fullStr | Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network |
| title_full_unstemmed | Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network |
| title_short | Goal-Directed Planning for Habituated Agents by Active Inference Using a Variational Recurrent Neural Network |
| title_sort | goal directed planning for habituated agents by active inference using a variational recurrent neural network |
| topic | goal directed planning active inference predictive coding variational bayes recurrent neural network |
| url | https://www.mdpi.com/1099-4300/22/5/564 |
| work_keys_str_mv | AT takazumimatsumoto goaldirectedplanningforhabituatedagentsbyactiveinferenceusingavariationalrecurrentneuralnetwork AT juntani goaldirectedplanningforhabituatedagentsbyactiveinferenceusingavariationalrecurrentneuralnetwork |