DROPS

Document

Rethinking the Role of Bayesianism in the Age of Modern AI (Dagstuhl Seminar 24461)

Authors: Vincent Fortuin, Mohammad Emtiyaz Khan, Mark van der Wilk, Zoubin Ghahramani, and Katharine Fisher

Published in: Dagstuhl Reports, Volume 14, Issue 11 (2025)

Abstract

Despite the recent success of large-scale deep learning, these systems still fall short in terms of their reliability and trustworthiness. They often lack the ability to estimate their own uncertainty in a calibrated way, encode meaningful prior knowledge, avoid catastrophic failures, and also reason about their environments to avoid such failures. Since its inception, Bayesian deep learning (BDL) has harbored the promise of achieving these desiderata by combining the solid statistical foundations of Bayesian inference with the practically successful engineering solutions of deep learning methods. This was intended to provide a principled mechanism to add the benefits of Bayesian learning to the framework of deep neural networks. However, compared to its promise, BDL methods often do not live up to the expectation and underdeliver in terms of real-world impact. This is due to many fundamental challenges related to, for instance, computation of approximate posteriors, unavailability of flexible priors, but also lack of appropriate testbeds and benchmarks. To make things worse, there are also numerous misconceptions about the scope of Bayesian methods, and researchers often end up expecting more than what they can get out of Bayes. By bringing together researchers from diverse communities, such as machine learning, statistics, and deep learning practice, in a personal and interactive seminar environment featuring debates, round tables, and brainstorming sessions, our Dagstuhl Seminar "Rethinking the Role of Bayesianism in the Age of Modern AI" (24461) has discussed these questions from a variety of angles and charted a path for future research to innovate, enhance, and strengthen meaningful real-world impact of Bayesian deep learning.

Cite as

Vincent Fortuin, Mohammad Emtiyaz Khan, Mark van der Wilk, Zoubin Ghahramani, and Katharine Fisher. Rethinking the Role of Bayesianism in the Age of Modern AI (Dagstuhl Seminar 24461). In Dagstuhl Reports, Volume 14, Issue 11, pp. 40-59, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2025)

Copy BibTex To Clipboard

@Article{fortuin_et_al:DagRep.14.11.40,
  author =	{Fortuin, Vincent and Khan, Mohammad Emtiyaz and van der Wilk, Mark and Ghahramani, Zoubin and Fisher, Katharine},
  title =	{{Rethinking the Role of Bayesianism in the Age of Modern AI (Dagstuhl Seminar 24461)}},
  pages =	{40--59},
  journal =	{Dagstuhl Reports},
  ISSN =	{2192-5283},
  year =	{2025},
  volume =	{14},
  number =	{11},
  editor =	{Fortuin, Vincent and Khan, Mohammad Emtiyaz and van der Wilk, Mark and Ghahramani, Zoubin and Fisher, Katharine},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/DagRep.14.11.40},
  URN =		{urn:nbn:de:0030-drops-228200},
  doi =		{10.4230/DagRep.14.11.40},
  annote =	{Keywords: Bayesian machine learning, deep learning, foundation models, model selection, uncertainty estimation}
}

Document

DOI: 10.4230/DagRep.14.2.182

AI for Social Good (Dagstuhl Seminar 24082)

Authors: Claudia Clopath, Ruben De Winne, Mohammad Emtiyaz Khan, and Jacopo Margutti

Published in: Dagstuhl Reports, Volume 14, Issue 2 (2024)

Abstract

Progress in the field of Artificial intelligence (AI) and machine learning (ML) has not slowed down in recent years. Long-standing challenges like Go have fallen and the technology has entered daily use via the vision, speech or translation capabilities in billions of smartphones. The pace of research progress shows no signs of slowing down, and demand for talent is unprecedented. AI for Social Good in general is trying to ensure that the social good does not become an afterthought, but that society benefits as a whole. In this Dagstuhl Seminar, which can be considered a follow-up edition of Dagstuhl Seminars 19082 and 22091 with the same title, we brought together AI and machine learning researchers with non-governmental organisations (NGOs), as they already pursue a social good goal, have rich domain knowledge, and vast networks with (non-)governmental actors in developing countries. Such collaborations benefit both sides: on the one hand, the new techniques can help with prediction, data analysis, modelling, or decision making. On the other hand, the NGOs’ domains contain many non-standard conditions, like missing data, side-effects, or multiple competing objectives, all of which are fascinating research challenges in themselves. And of course, publication impact is substantially enhanced when a method has real-world impact. In this seminar, researchers and practitioners from diverse areas of machine learning joined stakeholders from a range of NGOs to spend a week together. We first pursued an improved understanding of each side’s challenges and established a common language, via presentations and discussion groups. Building on this foundation, we organised a hackathon around some existing technical questions within the NGOs to scope the applicability of AI methods and seed collaborations. Finally, we discussed topics that cut across the AI for social good field, such as how to properly evaluate AI models that are used for good.

Cite as

Claudia Clopath, Ruben De Winne, Mohammad Emtiyaz Khan, and Jacopo Margutti. AI for Social Good (Dagstuhl Seminar 24082). In Dagstuhl Reports, Volume 14, Issue 2, pp. 182-190, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2024)

Copy BibTex To Clipboard

@Article{clopath_et_al:DagRep.14.2.182,
  author =	{Clopath, Claudia and De Winne, Ruben and Khan, Mohammad Emtiyaz and Margutti, Jacopo},
  title =	{{AI for Social Good (Dagstuhl Seminar 24082)}},
  pages =	{182--190},
  journal =	{Dagstuhl Reports},
  ISSN =	{2192-5283},
  year =	{2024},
  volume =	{14},
  number =	{2},
  editor =	{Clopath, Claudia and De Winne, Ruben and Khan, Mohammad Emtiyaz and Margutti, Jacopo},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/DagRep.14.2.182},
  URN =		{urn:nbn:de:0030-drops-205036},
  doi =		{10.4230/DagRep.14.2.182},
  annote =	{Keywords: artificial intelligence, interdisciplinary, machine learning, non-governmental organizations, social good}
}

Document

DOI: 10.4230/DagRep.9.2.111

AI for the Social Good (Dagstuhl Seminar 19082)

Authors: Claudia Clopath, Ruben De Winne, Mohammad Emtiyaz Khan, and Tom Schaul

Published in: Dagstuhl Reports, Volume 9, Issue 2 (2019)

Abstract

Artificial intelligence (AI) and machine learning (ML) have made impressive progress in the last few years. Long-standing challenges like Go have fallen and the technology has entered daily use via the vision, speech or translation capabilities in billions of smartphones. The pace of research progress shows no signs of slowing down, and demand for talent is unprecedented. AI for Social Good in general is trying to ensure that the social good does not become an afterthought, but that society benefits as a whole. In this Dagstuhl seminar, we brought together AI and machine learning researchers with non-governmental organisations (NGOs), as they already pursue a social good goal, have rich domain knowledge, and vast networks with (non)-governmental actors in developing countries. Such collaborations benefit both sides: on the one hand, the new techniques can help with prediction, data analysis, modelling, or decision making. On the other hand, the NGOs' domains contain many non-standard conditions, like missing data, side-effects, or multiple competing objectives, all of which are fascinating research challenges in themselves. And of course, publication impact is substantially enhanced when a method has real-world impact. In this workshop, researchers and practitioners from diverse areas of machine learning joined stakeholders from a range of NGOs to spend a week together. We first pursued an improved understanding of each side's challenges and established a common language, via presentations and discussion groups. We identified ten key challenges for AI for Social Good initiatives. To make matters concrete, we organised a hackathon around some existing technical questions within the NGOs to scope the applicability of AI methods and seed collaborations. Finally, we defined guidelines and next steps for future AI for Social Good initiatives.

Cite as

Claudia Clopath, Ruben De Winne, Mohammad Emtiyaz Khan, and Tom Schaul. AI for the Social Good (Dagstuhl Seminar 19082). In Dagstuhl Reports, Volume 9, Issue 2, pp. 111-122, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2019)

Copy BibTex To Clipboard

@Article{clopath_et_al:DagRep.9.2.111,
  author =	{Clopath, Claudia and De Winne, Ruben and Khan, Mohammad Emtiyaz and Schaul, Tom},
  title =	{{AI for the Social Good (Dagstuhl Seminar 19082)}},
  pages =	{111--122},
  journal =	{Dagstuhl Reports},
  ISSN =	{2192-5283},
  year =	{2019},
  volume =	{9},
  number =	{2},
  editor =	{Clopath, Claudia and De Winne, Ruben and Khan, Mohammad Emtiyaz and Schaul, Tom},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/DagRep.9.2.111},
  URN =		{urn:nbn:de:0030-drops-108620},
  doi =		{10.4230/DagRep.9.2.111},
  annote =	{Keywords: Machine Learning, Artificial Intelligence, Social Good, NGO, sustainable development goals, Non-governmental organisation}
}

Search Results

Documents authored by Khan, Mohammad Emtiyaz

Rethinking the Role of Bayesianism in the Age of Modern AI (Dagstuhl Seminar 24461)

Abstract

Cite as

AI for Social Good (Dagstuhl Seminar 24082)

Abstract

Cite as

AI for the Social Good (Dagstuhl Seminar 19082)

Abstract

Cite as

Thanks for your feedback!

Could not send message