Towards a Framework for Openness in Foundation Models. Basdevant, A., François, C., Storchan, V., Bankston, K., Bdeir, A., Debbah, M., Kapoor, S., LeCun, Y., Surman, M., King-Turvey, H., Lambert, N., Maffulli, S., Marda, N., Shivkumar, G., & Tunney, J. abstract bibtex Over the past year, there has been a robust debate about the benefits and risks of “open sourcing” foundation models. However, this discussion has often taken place at a high level of generality or with a narrow focus on specific technical attributes. In part, this is because defining "open source" for foundation models has proven tricky, given its significant differences from traditional software development. In order to inform more practical and nuanced decisions about opening AI systems, including foundation models, this paper presents a framework for grappling with openness across the AI stack. It summarizes previous work on this topic, analyzes the various potential reasons to pursue openness, and outlines how openness varies in different parts of the AI stack, both at the model and at the system level. In doing so, we hope to provide a common descriptive framework to deepen a nuanced and rigorous understanding of openness in AI and enable further work around definitions of openness and safety in AI.
@article{basdevant_towards_nodate,
title = {Towards a {Framework} for {Openness} in {Foundation} {Models}},
abstract = {Over the past year, there has been a robust debate about the benefits and risks of “open sourcing” foundation models. However, this discussion has often taken place at a high level of generality or with a narrow focus on specific technical attributes. In part, this is because defining "open source" for foundation models has proven tricky, given its significant differences from traditional software development. In order to inform more practical and nuanced decisions about opening AI systems, including foundation models, this paper presents a framework for grappling with openness across the AI stack. It summarizes previous work on this topic, analyzes the various potential reasons to pursue openness, and outlines how openness varies in different parts of the AI stack, both at the model and at the system level. In doing so, we hope to provide a common descriptive framework to deepen a nuanced and rigorous understanding of openness in AI and enable further work around definitions of openness and safety in AI.},
language = {en},
author = {Basdevant, Adrien and François, Camille and Storchan, Victor and Bankston, Kevin and Bdeir, Ayah and Debbah, Merouane and Kapoor, Sayash and LeCun, Yann and Surman, Mark and King-Turvey, Helen and Lambert, Nathan and Maffulli, Stefano and Marda, Nik and Shivkumar, Govind and Tunney, Justine},
}
Downloads: 0
{"_id":"BkanjtTAznDRAuWxE","bibbaseid":"basdevant-franois-storchan-bankston-bdeir-debbah-kapoor-lecun-etal-towardsaframeworkforopennessinfoundationmodels","author_short":["Basdevant, A.","François, C.","Storchan, V.","Bankston, K.","Bdeir, A.","Debbah, M.","Kapoor, S.","LeCun, Y.","Surman, M.","King-Turvey, H.","Lambert, N.","Maffulli, S.","Marda, N.","Shivkumar, G.","Tunney, J."],"bibdata":{"bibtype":"article","type":"article","title":"Towards a Framework for Openness in Foundation Models","abstract":"Over the past year, there has been a robust debate about the benefits and risks of “open sourcing” foundation models. However, this discussion has often taken place at a high level of generality or with a narrow focus on specific technical attributes. In part, this is because defining \"open source\" for foundation models has proven tricky, given its significant differences from traditional software development. In order to inform more practical and nuanced decisions about opening AI systems, including foundation models, this paper presents a framework for grappling with openness across the AI stack. It summarizes previous work on this topic, analyzes the various potential reasons to pursue openness, and outlines how openness varies in different parts of the AI stack, both at the model and at the system level. In doing so, we hope to provide a common descriptive framework to deepen a nuanced and rigorous understanding of openness in AI and enable further work around definitions of openness and safety in AI.","language":"en","author":[{"propositions":[],"lastnames":["Basdevant"],"firstnames":["Adrien"],"suffixes":[]},{"propositions":[],"lastnames":["François"],"firstnames":["Camille"],"suffixes":[]},{"propositions":[],"lastnames":["Storchan"],"firstnames":["Victor"],"suffixes":[]},{"propositions":[],"lastnames":["Bankston"],"firstnames":["Kevin"],"suffixes":[]},{"propositions":[],"lastnames":["Bdeir"],"firstnames":["Ayah"],"suffixes":[]},{"propositions":[],"lastnames":["Debbah"],"firstnames":["Merouane"],"suffixes":[]},{"propositions":[],"lastnames":["Kapoor"],"firstnames":["Sayash"],"suffixes":[]},{"propositions":[],"lastnames":["LeCun"],"firstnames":["Yann"],"suffixes":[]},{"propositions":[],"lastnames":["Surman"],"firstnames":["Mark"],"suffixes":[]},{"propositions":[],"lastnames":["King-Turvey"],"firstnames":["Helen"],"suffixes":[]},{"propositions":[],"lastnames":["Lambert"],"firstnames":["Nathan"],"suffixes":[]},{"propositions":[],"lastnames":["Maffulli"],"firstnames":["Stefano"],"suffixes":[]},{"propositions":[],"lastnames":["Marda"],"firstnames":["Nik"],"suffixes":[]},{"propositions":[],"lastnames":["Shivkumar"],"firstnames":["Govind"],"suffixes":[]},{"propositions":[],"lastnames":["Tunney"],"firstnames":["Justine"],"suffixes":[]}],"bibtex":"@article{basdevant_towards_nodate,\n\ttitle = {Towards a {Framework} for {Openness} in {Foundation} {Models}},\n\tabstract = {Over the past year, there has been a robust debate about the benefits and risks of “open sourcing” foundation models. However, this discussion has often taken place at a high level of generality or with a narrow focus on specific technical attributes. In part, this is because defining \"open source\" for foundation models has proven tricky, given its significant differences from traditional software development. In order to inform more practical and nuanced decisions about opening AI systems, including foundation models, this paper presents a framework for grappling with openness across the AI stack. It summarizes previous work on this topic, analyzes the various potential reasons to pursue openness, and outlines how openness varies in different parts of the AI stack, both at the model and at the system level. In doing so, we hope to provide a common descriptive framework to deepen a nuanced and rigorous understanding of openness in AI and enable further work around definitions of openness and safety in AI.},\n\tlanguage = {en},\n\tauthor = {Basdevant, Adrien and François, Camille and Storchan, Victor and Bankston, Kevin and Bdeir, Ayah and Debbah, Merouane and Kapoor, Sayash and LeCun, Yann and Surman, Mark and King-Turvey, Helen and Lambert, Nathan and Maffulli, Stefano and Marda, Nik and Shivkumar, Govind and Tunney, Justine},\n}\n\n\n\n\n\n\n\n\n\n\n\n","author_short":["Basdevant, A.","François, C.","Storchan, V.","Bankston, K.","Bdeir, A.","Debbah, M.","Kapoor, S.","LeCun, Y.","Surman, M.","King-Turvey, H.","Lambert, N.","Maffulli, S.","Marda, N.","Shivkumar, G.","Tunney, J."],"key":"basdevant_towards_nodate","id":"basdevant_towards_nodate","bibbaseid":"basdevant-franois-storchan-bankston-bdeir-debbah-kapoor-lecun-etal-towardsaframeworkforopennessinfoundationmodels","role":"author","urls":{},"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/zotero/abhishek-p","dataSources":["h7kKWXpJh2iaX92T5"],"keywords":[],"search_terms":["towards","framework","openness","foundation","models","basdevant","françois","storchan","bankston","bdeir","debbah","kapoor","lecun","surman","king-turvey","lambert","maffulli","marda","shivkumar","tunney"],"title":"Towards a Framework for Openness in Foundation Models","year":null}