Sackler Forum banner

About the Sackler Forum

The Raymond and Beverly Sackler U.S.-U.K. Scientific Forum was established to help the scientific leadership of the United Kingdom and the United States forge an enduring and productive partnership on pressing topics of worldwide scientific concern with benefit to all people.  These meetings are organized jointly by the National Academy of Sciences and the Royal Society, and alternate between locations in the U.S. and the U.K.

This program is made possible by a generous gift from Raymond and Beverly Sackler.


Upcoming Forum:

The Future of Machine Learning

January 31 - February 1, 2017

National Academy of Sciences Building
Washington, DC

Meeting Program


Register to watch the Forum live online!

Webcast participants can submit questions for the speakers by e-mailing Michaelle Schwalbe at

Session outlines below:

Session 1: The Frontiers of Machine Learning

The ubiquity of data, accessibility of computing power, and algorithmic advances have enabled revolutionary progress in machine learning over the past five years. As a result, technologies such as voice recognition or image perception, which a few years ago were performing at noticeably below-human levels, can now outperform people at some tasks. Not only does machine learning now underpin many applications that have become part of daily life, the field continues to evolve quickly and has the potential to play a transformative role across a diverse range of sectors.

This session will explore the frontiers of machine learning, considering both cutting-edge technology and near-term applications. The session will include a range of case studies, possibly including examples from transportation, medicine, public services, and finance. It will also consider what high profile advances in machine learning — such as the success of AlphaGo — indicate about the state of the science.

A key aim of this session will be to establish a common understanding amongst participants about applications of machine learning, both now and in the near-term, in order to frame subsequent discussions.

Session 2: Machine learning and society

As machine learning is used in an increasing range of applications, it raises legal and ethical questions, re-frames discussions about uses of data, and poses new challenges for the governance of this technology.

  • a. Decision-making: Automated decision-making systems are already in use, but machine learning offers the possibility of extending these processes, allowing a greater range and depth of decision-making without human input. Yet many machine learning systems are ‘black boxes’ whose methods can be difficult to interpret. Although such systems can produce statistically reliable results, they will not necessarily be able to create a narrative for those results. Machine learning algorithms can make mistakes without causing serious ethical concerns, for example in recommender systems for online retailers, but what happens when individuals are denied access to a public service or targeted for specific actions as a result of predictions and decisions made by machine learning algorithms? If such causal narratives become impossible in some sectors, new approaches to accountability and transparency, or new relationships between humans and machines, may be necessary.

  • b. Statistical stereotyping: Machine learning could compound the influence of social biases, both by perpetuating bias in training data or in the algorithms. However, machine learning also offers the promise of new ways to address biases in data—through making its effects more visible, or through algorithmic approaches aimed at ensuring fairness—and could therefore help make decision-making processes more robust.

  • c. The social acceptability of machine learning applications: There may be cases where machine learning will enable accurate predictions about future events or behaviours, in cases where such predictions are considered undesirable, unreasonable, or otherwise unacceptable, regardless of whether they are based on data which has been handled with due care, and in accordance with best practice. For example, society may effectively reject policy approaches that use machine learning to predict patterns of recidivism and monitor individuals accordingly, even when these approaches may reduce human bias and provide better outcomes.

  • d. Privacy: Machine learning reframes existing questions about privacy, the use of data, and the applicability today of governance systems that were designed in an earlier environment of information scarcity. Where data are used and datasets are combined in innovative ways to generate new insights, traditional approaches to managing privacy concerns may lose their effectiveness. If these traditional approaches are no longer sufficient, what new methods are needed? The balance of risks and benefits to individuals arising from the use of their data may play out differently in different contexts, such as healthcare or retail.

This session will consider a range of issues relating to machine learning and society. It will be grounded in current approaches or best practice in addressing key governance issues, through talks by relevant experts and key players in the commercial world, which will demonstrate how these issues are handled in a practical way. It may also include issues such as liability, accountability, and public engagement.

Session 3: Machine learning in the research and commercial communities

There are enormous opportunities in machine learning in academia, research labs, and industry. While much of the research and development of machine learning to date has been done in the commercial world, each of these communities will continue to advancing this field. Establishing key research challenges and areas of commercial opportunity will therefore both be important in moving the frontiers of machine learning forward.

In light of discussions in sessions one and two, this session will explore areas of significant opportunities in both industry and research, highlight where machine learning could make a significant impact, and describe some of the skills required. It will consider what type of advances might be possible, and the roles of large corporations, research labs and start-ups, including university spin-outs, in driving these advances. The session may also enable discussions about the entrepreneurial environment, and incentives and funding for start-ups.

Powered by Blackbaud
nonprofit software