EValueAction: a proposal for policy evaluation in simulation to support interactive imitation learning. Sibona, F., Luijkx, J., van der Heijden, B., Ferranti, L., & Indri, M. In IEEE INDIN 2023, 2023.
Paper abstract bibtex 3 downloads The up-and-coming concept of Industry 5.0 foresees human-centric flexible production lines, where collaborative robots support human workforce. In order to allow a seamless collaboration between intelligent robots and human workers, designing solutions for non-expert users is crucial. Learning from demonstration emerged as the enabling approach to address such a problem. However, more focus should be put on finding safe solutions which optimize the cost associated with the demonstrations collection process. This paper introduces a preliminary outline of a system, namely EValueAction (EVA), designed to assist the human in the process of collecting interactive demonstrations taking advantage of simulation to safely avoid failures. A policy is pre-trained with human-demonstrations and, where needed, new informative data are interactively gathered and aggregated to iteratively improve the initial policy. A trial case study further reinforces the relevance of the work by demonstrating the crucial role of informative demonstrations for generalization.
@inproceedings{sibona_evalueaction_2023,
title = {{EValueAction}: a proposal for policy evaluation in simulation to support interactive imitation learning},
url = {paper=https://r2clab.com/wp-content/uploads/2023/06/Paper_EVA_2023_acks.pdf},
abstract = {The up-and-coming concept of Industry 5.0 foresees
human-centric flexible production lines, where collaborative
robots support human workforce. In order to allow a seamless
collaboration between intelligent robots and human workers,
designing solutions for non-expert users is crucial. Learning from
demonstration emerged as the enabling approach to address such
a problem. However, more focus should be put on finding safe
solutions which optimize the cost associated with the demonstrations
collection process. This paper introduces a preliminary outline
of a system, namely EValueAction (EVA), designed to assist
the human in the process of collecting interactive demonstrations
taking advantage of simulation to safely avoid failures. A policy
is pre-trained with human-demonstrations and, where needed,
new informative data are interactively gathered and aggregated
to iteratively improve the initial policy. A trial case study further
reinforces the relevance of the work by demonstrating the crucial
role of informative demonstrations for generalization.},
booktitle = {{IEEE} {INDIN} 2023},
author = {Sibona, F. and Luijkx, J. and van der Heijden, B. and Ferranti, L. and Indri, M.},
year = {2023},
}
Downloads: 3
{"_id":"fTx3eZka9vt3PnvPS","bibbaseid":"sibona-luijkx-vanderheijden-ferranti-indri-evalueactionaproposalforpolicyevaluationinsimulationtosupportinteractiveimitationlearning-2023","author_short":["Sibona, F.","Luijkx, J.","van der Heijden, B.","Ferranti, L.","Indri, M."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"EValueAction: a proposal for policy evaluation in simulation to support interactive imitation learning","abstract":"The up-and-coming concept of Industry 5.0 foresees human-centric flexible production lines, where collaborative robots support human workforce. In order to allow a seamless collaboration between intelligent robots and human workers, designing solutions for non-expert users is crucial. Learning from demonstration emerged as the enabling approach to address such a problem. However, more focus should be put on finding safe solutions which optimize the cost associated with the demonstrations collection process. This paper introduces a preliminary outline of a system, namely EValueAction (EVA), designed to assist the human in the process of collecting interactive demonstrations taking advantage of simulation to safely avoid failures. A policy is pre-trained with human-demonstrations and, where needed, new informative data are interactively gathered and aggregated to iteratively improve the initial policy. A trial case study further reinforces the relevance of the work by demonstrating the crucial role of informative demonstrations for generalization.","booktitle":"IEEE INDIN 2023","author":[{"propositions":[],"lastnames":["Sibona"],"firstnames":["F."],"suffixes":[]},{"propositions":[],"lastnames":["Luijkx"],"firstnames":["J."],"suffixes":[]},{"propositions":["van","der"],"lastnames":["Heijden"],"firstnames":["B."],"suffixes":[]},{"propositions":[],"lastnames":["Ferranti"],"firstnames":["L."],"suffixes":[]},{"propositions":[],"lastnames":["Indri"],"firstnames":["M."],"suffixes":[]}],"year":"2023","bibtex":"@inproceedings{sibona_evalueaction_2023,\n\ttitle = {{EValueAction}: a proposal for policy evaluation in simulation to support interactive imitation learning},\n\turl = {paper=https://r2clab.com/wp-content/uploads/2023/06/Paper_EVA_2023_acks.pdf},\n\tabstract = {The up-and-coming concept of Industry 5.0 foresees\nhuman-centric flexible production lines, where collaborative\nrobots support human workforce. In order to allow a seamless\ncollaboration between intelligent robots and human workers,\ndesigning solutions for non-expert users is crucial. Learning from\ndemonstration emerged as the enabling approach to address such\na problem. However, more focus should be put on finding safe\nsolutions which optimize the cost associated with the demonstrations\ncollection process. This paper introduces a preliminary outline\nof a system, namely EValueAction (EVA), designed to assist\nthe human in the process of collecting interactive demonstrations\ntaking advantage of simulation to safely avoid failures. A policy\nis pre-trained with human-demonstrations and, where needed,\nnew informative data are interactively gathered and aggregated\nto iteratively improve the initial policy. A trial case study further\nreinforces the relevance of the work by demonstrating the crucial\nrole of informative demonstrations for generalization.},\n\tbooktitle = {{IEEE} {INDIN} 2023},\n\tauthor = {Sibona, F. and Luijkx, J. and van der Heijden, B. and Ferranti, L. and Indri, M.},\n\tyear = {2023},\n}\n\n","author_short":["Sibona, F.","Luijkx, J.","van der Heijden, B.","Ferranti, L.","Indri, M."],"urlpaper":"https://r2clab.com/wp-content/uploads/2023/06/Paper_EVA_2023_acks.pdf","key":"sibona_evalueaction_2023","id":"sibona_evalueaction_2023","bibbaseid":"sibona-luijkx-vanderheijden-ferranti-indri-evalueactionaproposalforpolicyevaluationinsimulationtosupportinteractiveimitationlearning-2023","role":"author","urls":{"Paper":"https://r2clab.com/wp-content/uploads/2023/06/Paper_EVA_2023_acks.pdf"},"metadata":{"authorlinks":{}},"downloads":3},"bibtype":"inproceedings","biburl":"https://api.zotero.org/groups/4723267/items?key=08YweqYIT8pOivc6EEeHdsB6&q=Ferranti&format=bibtex&sort=date&limit=1000","dataSources":["aZARmtpiqBtCBxMT4","5kt9xtsBSuRiKB4Ns","83WoChruhNBrARyJE","ckSAQwMWRg82iJe5D"],"keywords":[],"search_terms":["evalueaction","proposal","policy","evaluation","simulation","support","interactive","imitation","learning","sibona","luijkx","van der heijden","ferranti","indri"],"title":"EValueAction: a proposal for policy evaluation in simulation to support interactive imitation learning","year":2023,"downloads":3}