Nuo Wi Noel Tay, János Botzheim, C. Loo, N. Kubota
{"title":"Robust face recognition via transfer learning for robot partner","authors":"Nuo Wi Noel Tay, János Botzheim, C. Loo, N. Kubota","doi":"10.1109/RIISS.2014.7009163","DOIUrl":"https://doi.org/10.1109/RIISS.2014.7009163","url":null,"abstract":"Face recognition is crucial for human-robot interaction. Robot partners are required to work in real-time under unconstrained condition, yet, do not restrict the personal freedom of human occupants. On the other hand, due to its limited computational capability, a tradeoff between accuracy and computational load needs to be made. This tradeoff can be alleviated via the introduction of informationally structured space. For this paper, transfer learning is employed to perform unconstrained face recognition, where templates are constructed from domains acquired from various image-capturing devices, which is a subset of sensors from the informationally structured space. Given the environmental conditions, appropriate templates are used for recognition. Currently, different database images are used to simulate different environmental conditions. The templates can be easily learned and merged via a reformulated joint probabilistic face verification method, which reduces significantly the processing load. Tested on standard databases, experimental studies show that specific and small target domain samples can boost the recognition performance without imposing strain on computation.","PeriodicalId":270157,"journal":{"name":"2014 IEEE Symposium on Robotic Intelligence in Informationally Structured Space (RiiSS)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122268178","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}
{"title":"Autonomous motion primitive segmentation of actions for incremental imitative learning of humanoid","authors":"Farhan Dawood, C. Loo","doi":"10.1109/RIISS.2014.7009169","DOIUrl":"https://doi.org/10.1109/RIISS.2014.7009169","url":null,"abstract":"During imitation learning or learning by demon-stration/observation, a crucial element of conception involves segmenting the continuous flow of motion into simpler units ÂĂŗ- motion primitives -ÂĂŗ by identifying the boundaries of an action. Secondly, in realistic environment the robot must be able to learn the observed motion patterns incrementally in a stable adaptive manner. In this paper, we propose an on-line and unsupervised motion segmentation method rendering the robot to learn actions by observing the patterns performed by other partner through Incremental Slow Feature Analysis. The segmentation model directly operates on the images acquired from the robot's vision sensor (camera) without requiring any kinematic model of the demonstrator. After segmentation, the spatio-temporal motion sequences are learned incrementally through Topological Gaussian Adaptive Resonance Hidden Markov Model. The learning model dynamically generates the topological structure in a self-organizing and self-stabilizing manner.","PeriodicalId":270157,"journal":{"name":"2014 IEEE Symposium on Robotic Intelligence in Informationally Structured Space (RiiSS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129921773","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}
{"title":"A computational approach to parameter identification of spatially distributed nonlinear systems with unknown initial conditions","authors":"J. Kasać, V. Milić, J. Stepanic, G. Mester","doi":"10.1109/RIISS.2014.7009170","DOIUrl":"https://doi.org/10.1109/RIISS.2014.7009170","url":null,"abstract":"In this paper, a high-precision algorithm for parameter identification of nonlinear multivariable dynamic systems is proposed. The proposed computational approach is based on the following assumptions: a) system is nonlinearly parameterized by a vector of unknown system parameters; b) only partial measurement of system state is available; c) there are no state observers; d) initial conditions are unknown except for measurable system states. The identification problem is formulated as a continuous dynamic optimization problem which is discretized by higher-order Adams method and numerically solved by a backward-in-time recurrent algorithm which is similar to the backpropagation-through-time (BPTT) algorithm. The proposed algorithm is especially effective for identification of homogenous spatially distributed nonlinear systems what is demonstrated on the parameter identification of a multi-degree-of-freedom torsional system with nonlinearly parameterized elastic forces, unknown initial velocities and positions measurement only.","PeriodicalId":270157,"journal":{"name":"2014 IEEE Symposium on Robotic Intelligence in Informationally Structured Space (RiiSS)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133894047","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}
{"title":"Robot team learning enhancement using Human Advice","authors":"Justin Girard, M. Emami","doi":"10.1109/RIISS.2014.7009184","DOIUrl":"https://doi.org/10.1109/RIISS.2014.7009184","url":null,"abstract":"The paper discusses the augmentation of the Concurrent Individual and Social Learning (CISL) mechanism with a new Human Advice Layer (HAL). The new layer is characterized by a Gaussian Mixture Model (GMM), which is trained on human experience data. The CISL mechanism consists of the Individual Performance and Task Allocation Markov Decision Processes (MDP), and the HAL can provide preferred action selection policies to the individual agents. The data utilized for training the GMM is collected using a heterogeneous team foraging simulation. When leveraging human experience in the multi-agent learning process, the team performance is enhanced significantly.","PeriodicalId":270157,"journal":{"name":"2014 IEEE Symposium on Robotic Intelligence in Informationally Structured Space (RiiSS)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2014-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121473915","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}