Computer Science And Its Impact On Our Future
September 16-18, 2019
Jerusalem, Israel
The Forum covered a broad spectrum of issues on the intersection of computer science and society, such as the impact of quantum science on technology, computational technology in life sciences, and the development and application of AI and autonomous systems.
Click here for the printed program »
Click here for the agenda »
Greetings
Nili Cohen • President, Israel Academy of Sciences and Humanities (IASH)
Ken Fulton • Executive Director, National Academy of Sciences (NAS)
The Quantum Impact on Technology and Fundamental Questions
Charles H. Bennett • IBM Thomas J. Watson Research Center
Beginning in the late 1960’s quantum effects in information processing progressed from being viewed as a nuisance to be overcome, to an enabler of new kinds of. communication and computation with potentially revolutionary consequences for information technology, to a source of insights into fundamental questions like the origin of spacetime.
Harnessing Topological Physics for Quantum Computation
Ady Stern • Weizmann Institute of Science
I will give an overview talk on topological states of matter and the way they may be useful for battling de-coherence of quantum computers. To that end, I will start by explaining what topological states of matter are and why they are a perfect example of scientific beauty. I will then explain what de-coherence is, and why it is a major obstacle on the way to realizing quantum computers. Next, I will describe how topological states may help in overcoming this obstacle, and how they may pave the way to realizing a centuries-old human desire to eat the cake and have it too. Finally I will conclude by emphasizing the practical difficulties on the way.
Complexity and Simplicity in Economic Design
Noam Nisan • The Hebrew University of Jerusalem
As more and more economic activity moves to the Internet, familiar economic mechanisms are being deployed at unprecedented scales of size, speed, and complexity. In many cases this new complexity becomes the defining feature of the deployed economic mechanism and the quantitative difference becomes a key qualitative one. The talk will discuss some of the challenges and opportunities that this complexity-in several different senses of “complexity”– entails.
Complexity-Theoretic Perspectives on Algorithmic Fairness
Omer Reingold • Stanford University
A prominent concern, in the age of machine learning and data analysis, is that left to their own devices, algorithms will propagate–even amplify–existing biases. Common definitions of fairness are group-based, typically requiring that a given statistic be equal across a few demographic groups, socially identified as deserving protection. Such definitions tend to be easy to study and satisfy but are liable to provide exceedingly weak protection from unfair discrimination. This motivated the introduction of definitions that aim at individual-level protections. Such protection is much stronger and more nuanced but harder to satisfy. We will discuss a recent sequence of results, where protection is provided to a large collection of populations, defined complexity-theoretically. This gives a surprising approach to a question that has been debated by a variety of scientific communities; which population should be protected? Our approach suggests protecting every group that can be identified given the data and our computational limitations, which in a sense is the best we can hope to do. We will discuss this approach in different contexts as well as its relation to pseudo-randomness.
Based on joint works with Cynthia Dwork, Úrsula Hébert-Johnson, Michael P. Kim, Guy Rothblum and Gal Yona.
Integrated Computational Analysis in Cancer and Precision Medicine
Ron Shamir • Tel Aviv University
Today’s large biological datasets open novel opportunities in basic science and medicine. While inquiry of each dataset separately often provides insights, integrative analysis may reveal more holistic, systems-level findings. We demonstrate the power of integrated analysis in cancer on two levels: (1) in analysis of one omic in many cancer types together, and (2) in analysis of multiple omics for the same cancer. In both cases, we develop novel methods and observe a clear advantage of the integration. We also describe a new method for identifying and ranking driver genes in an individual’s tumor, based on expression and mutation profiles.
Robot-Human Collaboration and Learning for Improving Human Well-Being
Sarit Kraus • Bar-Ilan University
Autonomous mobile robots can greatly improve human well-being. For example, robot assistance to older adults and people with physical impairments could be extremely beneficial in our aging society. Challenges in building functional robots that act in unstructured environments such as homes and hospitals are: current technological limitations in perception and action of robots in unstructured and highly variable environments; the variety of tasks at hand are complex involving man, machine and a dynamic interaction between them; the nature of functional robots that interact with people requires the development of general purpose policies that can be computed efficiently. In order to increase robot autonomy, there is a need to develop innovative algorithms and methodologies enabling a small number of human operators to work together, possibly by tele-operating the robots, with a large number of mobile robots located in different unstructured locations, such as homes. The robots will learn to improve their capabilities over time, enabling them to become more autonomous in facing new tasks and adapting to users changing needs. Operators can also assist the robots in the learning process. In this talk, I will discuss some preliminary results toward this vision.
Learning in Multi-Agent Environments
Eva Tardos • Cornell University
In many online systems, participants use data and algorithms to experiment and learn how to best use the system. Examples include traffic routing as well as online auctions. Game theory classically studies Nash equilibrium as the outcome of selfish interaction, and has many examples illustrating that selfish behavior can lead to suboptimal outcome for all participants. Over the last decade, we developed good understanding of how to quantify the impact of strategic user behavior on overall performance in Nash equilibria of games. In this talk we will focus on games where players use a form of learning that helps them adapt to the environment. We ask if the quantitative guarantees obtained for Nash equilibria extend to such out of equilibrium game play, possibly even in dynamically changing environments? Or possible even better, does learning lead the agents to outcomes that are better than the (worst) Nash equilibrium?
The Promise of Machine Learning and AI in Transforming Industries
Amnon Shashua • The Hebrew University of Jerusalem; Mobileye
The fast-moving field of artificial intelligence is transforming industries. I will focus on two examples in some details. First, the field of transportation is undergoing a seismic change with the coming introduction of autonomous driving. The technologies required to enable computer driven cars involves the latest cutting edge artificial intelligence and machine learning algorithms. I will review the various innovations and challenges associated with bringing autonomous driving to reality and the changes it will likely bring to society at large. Second, “AI as a Companion” is likely to is gaining momentum and, at some point, computers will transform from a tool to a companion. I will focus on two segments of society, the visual impaired and the hearing disabled, for which the current state of the art has a game-changing effect.
Quantum Information Science: A Computational Lens on Quantum Physics
Dorit Aharonov • The Hebrew University of Jerusalem
While the jury is still out as to when and where the impressive experimental progress on quantum gates and qubits will indeed lead one day to a full scale quantum computing machine, a new and not-less exciting development has been taking place over the past decade. Computational notions such as reductions, hardness, and completeness are quickly starting to be integrated into the very heart of the research of many body quantum systems. The computational perspective brings deep new insights into physical questions that seem completely unrelated to computers, including precision measurement and sensing, testing quantum mechanics, condensed matter physics and even black holes and quantum gravity. I will try to explain some of these intriguing connections and implications, and time permitting, will ponder about what next.
Secure Computation with Quantum Devices: From Device-Independent Cryptography to Verification of Quantum Computers
Thomas Vidick • California Institute of Technology
Quantum cryptography is at present one of the most technologically mature applications of quantum information. The discovery of the unprecedented possibilities of quantum states for secure communication predates the algorithmic discoveries of quantum computing. Yet the recourse to quantum states of matter presents unique difficulties: due to the superposition principle, quantum states are exponentially complex; due to the uncertainty principle, the information they encode cannot be read off without disturbance. Thus arises one of the most pressing challenges of the field of quantum computing: how can “classical” human beings and machines test, control and verify the behavior of quantum computing devices? In the talk I will trace the history of this question, from its origins in the study of device-independent quantum cryptography to recent discoveries on the verification of quantum devices.
Safe Machine Learning
Shafi Goldwasser • Massachusetts Institute of Technology
Cryptography and Computational Learning have shared a curious history: a scientific success for one often provided an example of an impossible task for the other. Today, the goals of the two fields are aligned. Cryptographic models and tools can and should play a role in ensuring the the safe use of machine learning. We will discuss this development with its challenges and opportunities.
Humans, Machines, and Work: The Future is Now
Moshe Vardi • Rice University
Automation, driven by technological progress, has been increasing inexorably for the past several decades. Two schools of economic thought have for many years been engaged in a debate about the potential effects of automation on jobs: will new technology spawn mass unemployment, as the robots take jobs away from humans? Or will the jobs robots take over create demand for new human jobs? I will present data that demonstrate that the concerns about automation are valid. In fact, technology has been hurting working-class people for the past 40 years. The discussion about humans, machines and work tends to be a discussion about some undetermined point in the far future. But it is time to face reality. The future is now.
Similarities and Differences between Deep Learning and our Biological Brains
Naftali Tishby • The Hebrew University of Jerusalem
The new phase of artificial intelligence, in particular the one based on artificial neural networks, is truly revolutionizing the world. Problems considered extremely difficult for machines, like human face recognition, speech recognition and natural language understanding, are now done routinely on small devices like our smartphones. The striking fact is that this amazing progress was achieved not by ingenious mathematics and engineering but by a very naive attempt to mimic the human brain. How similar is the current AI to biological intelligence? Do they both suffer from similar lack of robustness and interpretability? Are some of the deficiencies of AI unavoidable? Are we on the right track to use AI for better understanding of our own brains? I will touch on some of these issues in view of the progress made in both AI and neuroscience.
Smart Swarms in the Wild and in the Small
Gal Kaminka • Bar Ilan University
Swarms are ubiquitous in natural creatures, but rare in synthetic agents. Human pedestrians, crowds and audiences, animals moving collectively, insect colonies and hives, all act as swarms: coordinating based on limited local information, yet generating significant group-wide effects. Their synthetic counterparts are often academically fascinating, but just so. This talk will highlight two high-risk, high-payoff research efforts, where artificially-created swarms open doors to high-impact science: the use of robots to study whether insect swarms can be rational (in the decision-theoretic sense), and the automated generation of swarm-drugs, made from molecular robots (affectionately called nanobots). I will present preliminary promising results, and discuss current and future steps to fulfill the promise of synthetic swarms.
Design Principles of Physiological Circuits
Uri Alon • • Weizmann Institute of Science
To understand human physiology, it is important to find mathematical principles that can unite different systems. I will describe advances in understanding of principles that allow hormone systems to work despite huge variations in physiological parameters. These principles offer new ways to understand the origin of several diseases.
Big Data, Wellness and Disease
Leroy (Lee) Hood • Institute for Systems Biology
A systems approach to human health and the big data derived from genome and longitudinal phenome of individuals leads to analyses that are transforming our understanding of wellness and disease. We have employed genome and “deep phenotyping” to make billions of measurements on individual humans–assaying hundreds of biological networks--to generate thousands of longitudinal data clouds that have given remarkable new insights into many aspects of human wellness and disease. Further, they lead us to a 21st century medicine that is predictive, preventive, personalized and participatory (P4) and that in turn leads one to the conclusion that healthcare has two major domains–wellness and disease. The objective of 21st century medicine will be to understand deeply wellness and disease and to identify and reverse the early transitions between the two–the preventive medicine of the 21st century. These insights have far-reaching implications for the practice of 21st century medicine, some of which will be discussed.
