Active Inference With Empathy Mechanism for Socially Behaved Artificial Agents in Diverse Situations

hello and welcome it's March 13th 2024 and we're here in active inference guest stream number 75.1 active inference with empathy mechanism for socially behaved artificial agent in diverse situations with t Yuki matsumura kanako isaki and huki Mizuno so thank you all for joining and looking forward to hearing your talk and discussion okay thank you for inviting us to such a good opportunity and I'm T and my co researcher kanako and huki also join this discussion we are developing AI especially AI which is inspired by humanik autonomous agent to make social artificial agent recently we published a paper related to the idea in Journal of a life the title is active imp with empathy mechanism for socially behaved artificial agent in diverse situations first I'd like to explain the paper this is the motivation of This research at the title of the paper shows the motivation of This research is to develop socially behaved artificial agent we believed that sociality will become an important point to apply AI in our daily life the difficulty to realize social agent is that an appropriate social behavior depend on a given situation or culture as described in the bottom figure we have sociality even in just working case for example how fast should we walk or how much distance from others orah blah blah so if we Implement sociality with rule based program we have to design a lot of magic number for each situations this is impossible because there are a white variety of situations due to this unified mechanism or principle for social behavior is required this is the challenge we assume that we humans has such unified behavior model because we humans can behave socially in diverse situations and we develop social Eng inspired by human behavior model namely three principle three energy principle and active inflence in this research we use the active inflence for the basis of the behavior model of human life agent as you know active inference is a concept proposed in the free energy principle and free energy principle is a unified hypo for our variety of cognitive activities in the FP our cognitive activities are explained by such that first we have internal models for predicting future of surrounding environment and second we always try to minimize the free energy of the model free energy is defined by this equation and it intuitively represent uncertainty of prediction the interesting point of the FP is that the internal model if not only the hidden state but also actions for the agent that is the action to be taken is the action which minimize the free energy of the agent this is the basic idea of active inference as I understand it intentional actions also can be explained in the active inference in this case we consider expected free energy which is the free energy for the future State when considering much time Step Ahead expected energy at much time Step Ahead is considered as described in the right figure based on the expected free energy for each action the distribution the action is determined such that the smaller the expected free energy the higher the probability of section when we consider intentional Behavior intentions are considered as PRS of future observations and reward according to the Future observation are encoded into the first time into the uh in the last equation these are my understanding of the standard active inference and the behavior model of the humanik agent based on it we extend this active inference to generate the social behavior the idea of the extension is explained using a situation in which there are two agents called I and other the agent called I is the subject of the action in this explanation the biological agent under the AC inference makes predictions about the environment and take some action to minimize the uncertainty of the prediction in this figure I predict future beh I predict future behavior of the other because the a is an environment for I namely I act to reduce uncertainty for the for…