Data Efficient Paradigms for Personalized Assessment of Taskable AI Systems. Verma, P. In Proceedings of the 20th ICAPS Doctoral Consortium (ICAPS DC 2022), pages 18–22, 2022. ![Data Efficient Paradigms for Personalized Assessment of Taskable AI Systems [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper ![Data Efficient Paradigms for Personalized Assessment of Taskable AI Systems [link]](https://bibbase.org/img/filetypes/link.svg "link")
Presentation abstract bibtex 15 downloads The vast diversity of internal designs of taskable black-box AI systems and their nuanced zones of safe functionality make it difficult for a layperson to use them without unintended side effects. The focus of my dissertation is to develop algorithms and requirements of interpretability that would enable a user to assess and understand the limits of an AI system’s safe operability. We develop a personalized AI assessment module that lets an AI system execute instruction sequences in simulators and answer the queries about its execution of sequences of actions. Our results show that such a primitive query-response capability is sufficient to efficiently derive a user-interpretable model of the system’s capabilities in fully observable, and deterministic settings.
@InProceedings{ICAPSDC2022paper-05,
author = {Pulkit Verma},
title = {Data Efficient Paradigms for Personalized Assessment of Taskable AI Systems},
booktitle = {Proceedings of the 20th ICAPS Doctoral Consortium (ICAPS DC 2022)},
year = {2022},
pages = {18--22},
abstract = {The vast diversity of internal designs of taskable black-box AI systems and their nuanced zones of safe functionality make it difficult for a layperson to use them without unintended side effects. The focus of my dissertation is to develop algorithms and requirements of interpretability that would enable a user to assess and understand the limits of an AI system’s safe operability. We develop a personalized AI assessment module that lets an AI system execute instruction sequences in simulators and answer the queries about its execution of sequences of actions. Our results show that such a primitive query-response capability is sufficient to efficiently derive a user-interpretable model of the system’s capabilities in fully observable, and deterministic settings.},
url_paper = {https://icaps22.icaps-conference.org/dc/ICAPS_2022_paper_363.pdf},
url_presentation = {https://www.youtube.com/watch?v=7jwlOX-0qrE&list=PLj-ZdQ5rfSEqD1SztBXJppdjIE9CQQfzV}
}
Downloads: 15
{"_id":"e6HT6XpyewQxTHSEx","bibbaseid":"verma-dataefficientparadigmsforpersonalizedassessmentoftaskableaisystems-2022","author_short":["Verma, P."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["Pulkit"],"propositions":[],"lastnames":["Verma"],"suffixes":[]}],"title":"Data Efficient Paradigms for Personalized Assessment of Taskable AI Systems","booktitle":"Proceedings of the 20th ICAPS Doctoral Consortium (ICAPS DC 2022)","year":"2022","pages":"18–22","abstract":"The vast diversity of internal designs of taskable black-box AI systems and their nuanced zones of safe functionality make it difficult for a layperson to use them without unintended side effects. The focus of my dissertation is to develop algorithms and requirements of interpretability that would enable a user to assess and understand the limits of an AI system’s safe operability. We develop a personalized AI assessment module that lets an AI system execute instruction sequences in simulators and answer the queries about its execution of sequences of actions. Our results show that such a primitive query-response capability is sufficient to efficiently derive a user-interpretable model of the system’s capabilities in fully observable, and deterministic settings.","url_paper":"https://icaps22.icaps-conference.org/dc/ICAPS_2022_paper_363.pdf","url_presentation":"https://www.youtube.com/watch?v=7jwlOX-0qrE&list=PLj-ZdQ5rfSEqD1SztBXJppdjIE9CQQfzV","bibtex":"@InProceedings{ICAPSDC2022paper-05,\n author = {Pulkit Verma},\n title = {Data Efficient Paradigms for Personalized Assessment of Taskable AI Systems},\n booktitle = {Proceedings of the 20th ICAPS Doctoral Consortium (ICAPS DC 2022)},\n year = {2022},\n pages = {18--22},\n abstract = {The vast diversity of internal designs of taskable black-box AI systems and their nuanced zones of safe functionality make it difficult for a layperson to use them without unintended side effects. The focus of my dissertation is to develop algorithms and requirements of interpretability that would enable a user to assess and understand the limits of an AI system’s safe operability. We develop a personalized AI assessment module that lets an AI system execute instruction sequences in simulators and answer the queries about its execution of sequences of actions. Our results show that such a primitive query-response capability is sufficient to efficiently derive a user-interpretable model of the system’s capabilities in fully observable, and deterministic settings.},\n url_paper = {https://icaps22.icaps-conference.org/dc/ICAPS_2022_paper_363.pdf},\n url_presentation = {https://www.youtube.com/watch?v=7jwlOX-0qrE&list=PLj-ZdQ5rfSEqD1SztBXJppdjIE9CQQfzV}\n}\n\n","author_short":["Verma, P."],"key":"ICAPSDC2022paper-05","id":"ICAPSDC2022paper-05","bibbaseid":"verma-dataefficientparadigmsforpersonalizedassessmentoftaskableaisystems-2022","role":"author","urls":{" paper":"https://icaps22.icaps-conference.org/dc/ICAPS_2022_paper_363.pdf"," presentation":"https://www.youtube.com/watch?v=7jwlOX-0qrE&list=PLj-ZdQ5rfSEqD1SztBXJppdjIE9CQQfzV"},"metadata":{"authorlinks":{}},"downloads":15,"html":""},"bibtype":"inproceedings","biburl":"https://icaps22.icaps-conference.org/dc/dc-papers.bib","dataSources":["g2C7Q7wimFzyqrFAt"],"keywords":[],"search_terms":["data","efficient","paradigms","personalized","assessment","taskable","systems","verma"],"title":"Data Efficient Paradigms for Personalized Assessment of Taskable AI Systems","year":2022,"downloads":15}