ToM for Teams最新文献

Using Features at Multiple Temporal and Spatial Resolutions to Predict Human Behavior in Real Time 利用多时间和空间分辨率的特征实时预测人类行为

ToM for Teams Pub Date : 2022-11-12 DOI: 10.48550/arXiv.2211.06721

L. Zhang, Justin Lieffers, A. Pyarelal

{"title":"Using Features at Multiple Temporal and Spatial Resolutions to Predict Human Behavior in Real Time","authors":"L. Zhang, Justin Lieffers, A. Pyarelal","doi":"10.48550/arXiv.2211.06721","DOIUrl":"https://doi.org/10.48550/arXiv.2211.06721","url":null,"abstract":". When performing complex tasks, humans naturally reason at multiple temporal and spatial resolutions simultaneously. We contend that for an artiﬁcially intelligent agent to eﬀectively model human teammates, i.e., demonstrate computational theory of mind (ToM), it should do the same. In this paper, we present an approach for integrating high and low-resolution spatial and temporal information to predict human behavior in real time and evaluate it on data collected from human subjects performing simulated urban search and rescue (USAR) missions in a Minecraft-based environment. Our model composes neural networks for high and low-resolution feature extraction with a neural network for behavior prediction, with all three networks trained simultaneously. The high-resolution extractor encodes dynamically changing goals robustly by taking as input the Manhattan distance diﬀerence between the humans’ Minecraft avatars and candidate goals in the environment for the latest few actions, computed from a high-resolution gridworld representation. In contrast, the low-resolution extractor encodes participants’ historical behavior using a historical state matrix computed from a low-resolution graph representation. Through supervised learning, our model acquires a robust prior for human behavior prediction, and can eﬀectively deal with long-term observations. Our experimental results demonstrate that our method signiﬁcantly improves prediction accuracy compared to approaches that only use high-resolution information. layer combined with a batch normalization layer as a basic building block for our three neural networks. The output FC layers in the prediction network ( g ( e lr , e hr )) are passed through softmax and sigmoid functions to obtain the probabilities of the agent’s goal ( O gp ) and the likelihood that the next victim is triaged ( O vp ), respectively.","PeriodicalId":119585,"journal":{"name":"ToM for Teams","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130502257","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}