Abstraction for Deep Reinforcement Learning

Abstraction for Deep Reinforcement Learning. Shanahan, M. & Mitchell, M. In Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence, pages 5588–5596, Vienna, Austria, July, 2022. International Joint Conferences on Artificial Intelligence Organization.

Paper doi abstract bibtex

We characterise the problem of abstraction in the context of deep reinforcement learning. Various well established approaches to analogical reasoning and associative memory might be brought to bear on this issue, but they present difficulties because of the need for end-to-end differentiability. We review developments in AI and machine learning that could facilitate their adoption.

@inproceedings{shanahan_abstraction_2022,
	address = {Vienna, Austria},
	title = {Abstraction for {Deep} {Reinforcement} {Learning}},
	isbn = {9781956792003},
	url = {https://www.ijcai.org/proceedings/2022/780},
	doi = {10.24963/ijcai.2022/780},
	abstract = {We characterise the problem of abstraction in the context of deep reinforcement learning. Various well established approaches to analogical reasoning and associative memory might be brought to bear on this issue, but they present difficulties because of the need for end-to-end differentiability. We review developments in AI and machine learning that could facilitate their adoption.},
	language = {en},
	urldate = {2022-07-30},
	booktitle = {Proceedings of the {Thirty}-{First} {International} {Joint} {Conference} on {Artificial} {Intelligence}},
	publisher = {International Joint Conferences on Artificial Intelligence Organization},
	author = {Shanahan, Murray and Mitchell, Melanie},
	month = jul,
	year = {2022},
	pages = {5588--5596},
}

Downloads: 0

{"_id":"8YTKJjJ9fTtHSaGZe","bibbaseid":"shanahan-mitchell-abstractionfordeepreinforcementlearning-2022","author_short":["Shanahan, M.","Mitchell, M."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","address":"Vienna, Austria","title":"Abstraction for Deep Reinforcement Learning","isbn":"9781956792003","url":"https://www.ijcai.org/proceedings/2022/780","doi":"10.24963/ijcai.2022/780","abstract":"We characterise the problem of abstraction in the context of deep reinforcement learning. Various well established approaches to analogical reasoning and associative memory might be brought to bear on this issue, but they present difficulties because of the need for end-to-end differentiability. We review developments in AI and machine learning that could facilitate their adoption.","language":"en","urldate":"2022-07-30","booktitle":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","publisher":"International Joint Conferences on Artificial Intelligence Organization","author":[{"propositions":[],"lastnames":["Shanahan"],"firstnames":["Murray"],"suffixes":[]},{"propositions":[],"lastnames":["Mitchell"],"firstnames":["Melanie"],"suffixes":[]}],"month":"July","year":"2022","pages":"5588–5596","bibtex":"@inproceedings{shanahan_abstraction_2022,\n\taddress = {Vienna, Austria},\n\ttitle = {Abstraction for {Deep} {Reinforcement} {Learning}},\n\tisbn = {9781956792003},\n\turl = {https://www.ijcai.org/proceedings/2022/780},\n\tdoi = {10.24963/ijcai.2022/780},\n\tabstract = {We characterise the problem of abstraction in the context of deep reinforcement learning. Various well established approaches to analogical reasoning and associative memory might be brought to bear on this issue, but they present difficulties because of the need for end-to-end differentiability. We review developments in AI and machine learning that could facilitate their adoption.},\n\tlanguage = {en},\n\turldate = {2022-07-30},\n\tbooktitle = {Proceedings of the {Thirty}-{First} {International} {Joint} {Conference} on {Artificial} {Intelligence}},\n\tpublisher = {International Joint Conferences on Artificial Intelligence Organization},\n\tauthor = {Shanahan, Murray and Mitchell, Melanie},\n\tmonth = jul,\n\tyear = {2022},\n\tpages = {5588--5596},\n}\n\n\n\n","author_short":["Shanahan, M.","Mitchell, M."],"key":"shanahan_abstraction_2022","id":"shanahan_abstraction_2022","bibbaseid":"shanahan-mitchell-abstractionfordeepreinforcementlearning-2022","role":"author","urls":{"Paper":"https://www.ijcai.org/proceedings/2022/780"},"metadata":{"authorlinks":{}},"downloads":0,"html":""},"bibtype":"inproceedings","biburl":"https://bibbase.org/zotero/alukina","dataSources":["Cfgnp5s4HQSBd8tAf"],"keywords":[],"search_terms":["abstraction","deep","reinforcement","learning","shanahan","mitchell"],"title":"Abstraction for Deep Reinforcement Learning","year":2022}