PDF Reversible Computation 10th International Conference RC 2018 Leicester UK September 12 14 2018 Proceedings Jarkko Kari Download
PDF Reversible Computation 10th International Conference RC 2018 Leicester UK September 12 14 2018 Proceedings Jarkko Kari Download
PDF Reversible Computation 10th International Conference RC 2018 Leicester UK September 12 14 2018 Proceedings Jarkko Kari Download
com
https://textbookfull.com/product/reversible-
computation-10th-international-conference-
rc-2018-leicester-uk-
september-12-14-2018-proceedings-jarkko-kari/
https://textbookfull.com/product/reversible-computation-12th-
international-conference-rc-2020-oslo-norway-
july-9-10-2020-proceedings-ivan-lanese/
textbookfull.com
https://textbookfull.com/product/learning-based-vanet-communication-
and-security-techniques-liang-xiao/
textbookfull.com
Fundamentals of Midwifery A Textbook for Students 1st
Edition Louise Lewis
https://textbookfull.com/product/fundamentals-of-midwifery-a-textbook-
for-students-1st-edition-louise-lewis/
textbookfull.com
https://textbookfull.com/product/the-mathematics-and-mechanics-of-
biological-growth-1st-edition-alain-goriely-2/
textbookfull.com
https://textbookfull.com/product/exercise-cognition-interaction-
neuroscience-perspectives-1st-edition-mcmorris/
textbookfull.com
Labour Questions In The Global South Praveen Jha
https://textbookfull.com/product/labour-questions-in-the-global-south-
praveen-jha/
textbookfull.com
Jarkko Kari
Irek Ulidowski (Eds.)
LNCS 11106
Reversible Computation
10th International Conference, RC 2018
Leicester, UK, September 12–14, 2018
Proceedings
123
Lecture Notes in Computer Science 11106
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen
Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zurich, Zurich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology Madras, Chennai, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7408
Jarkko Kari Irek Ulidowski (Eds.)
•
Reversible Computation
10th International Conference, RC 2018
Leicester, UK, September 12–14, 2018
Proceedings
123
Editors
Jarkko Kari Irek Ulidowski
University of Turku University of Leicester
Turku Leicester
Finland UK
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
This volume contains the proceedings of RC 2018, the 10th International Conference
on Reversible Computation, held in Leicester, UK, during September 12–14, 2018. RC
2018 was the tenth event in a series of annual meetings designed to gather researchers
from different scientific disciplines for the discussion and dissemination of recent
developments in all aspects of reversible computation. Previous RC events took place
in York, UK (2009), Bremen, Germany (2010), Ghent, Belgium (2011), Copenhagen,
Denmark (2012), Victoria, Canada (2013), Kyoto, Japan (2014), Grenoble, France
(2015), Bologna, Italy (2016), and Kolkata, India (2017).
Reversible computation concerns models of computation where programs or pro-
cesses are logically reversible (as, for example, in undoing of program execution for
reversible debugging), or physically reversible (as, for example, in quantum circuits
and robotics). The main areas of research presented at the conference were reversible
formal models for computation and physical systems, reversible programming lan-
guages, and reversible circuits.
The conference received 28 submissions, and we would like to thank everyone who
submitted. Each submission was reviewed by at least three reviewers, who provided
detailed evaluations as well as constructive comments and recommendations. After
careful reviewing and extensive discussions, the Program Committee (PC) accepted 13
full papers, one tutorial paper, and seven short papers for presentation at the confer-
ence. We would like to thank the PC members and all the additional reviewers for their
truly professional work and strong commitment to the success of RC 2018. We are also
grateful to the authors for taking into account the comments and suggestions provided
by the referees during the preparation of the final versions of their papers.
To mark the tenth edition of the conference, the conference program included four
invited talks. Michael P. Frank discussed “Physical Foundations of Landauer’s Prin-
ciple,” Ivan Lanese took the audience “From Reversible Semantics to Reversible
Debugging,” Norman Margolus discussed “Finite-State Classical Mechanics,” and
Nicolas Ollinger presented work “On Aperiodic Reversible Turing Machines.” The
papers that accompany the invited talks are included in these proceedings. Addition-
ally, we were honored to welcome Edward F. Fredkin at the conference who gave a talk
titled “Discrete Space-Time-State Physics”; a short abstract for the talk is included in
these proceedings.
We would like to thank everyone who contributed to the organization of RC 2018,
especially Ushma Chauhan, James Hoey, Claudio Antares Mezzina, and Emilio
Tuosto. We also thank Lisa Jungmann and Robert Wille for helping with the confer-
ence website. We thank the Department of Informatics of the University of Leicester
and the COST Action IC1405 for their financial support. Finally, we acknowledge
EasyChair for facilitating PC discussions and the production of the proceedings.
Program Committee
Gerhard Dueck University of New Brunswick, Canada
Carla Ferreira Universidade NOVA de Lisboa, Portugal
Michael P. Frank Sandia National Laboratories, USA
Anahi Gajardo Universidad de Concepción, Chile
Robert Glück University of Copenhagen, Denmark
Jarkko Kari University of Turku, Finland
Mozammel Huq Azad Khan East West University, Bangladesh
Ivan Lanese University of Bologna, Italy
Claudio Antares Mezzina IMT School for Advanced Studies Lucca, Italy
Claudio Moraga TU Dortmund University, Germany
Ulrik Pagh Schultz University of Southern, Denmark
Iain Phillips Imperial College London, UK
Krzysztof Podlaski University of Łódź, Poland
Mariusz Rawski Warsaw University of Technology, Poland
Ville Salo University of Turku, Finland
Peter Selinger Dalhousie University, Canada
Mathias Soeken Ecole Polytechnique Fédérale de Lausanne,
Switzerland
Michael Kirkedal Thomsen DIKU, University of Copenhagen, Denmark
Irek Ulidowski University of Leicester, UK
German Vidal MiST, DSIC, Universitat Politecnica de Valencia,
Spain
Robert Wille Johannes Kepler University Linz, Austria
Shigeru Yamashita Ritsumeikan University, Japan
Tetsuo Yokoyama Nanzan University, Japan
Additional Reviewers
Kaarsgaard, Robin
Klimov, Andrei
Midtgaard, Jan
Mogensen, Torben Ægidius
Palacios, Adrian
Philippou, Anna
Discrete Space-Time-State Physics
(Social Event Talk)
Edward F. Fredkin
Our hypothesis is that the process that is the most microscopic discrete
physics (perhaps underlying QM) could correspond exactly to the temporal
evolution of state of some such discrete, deterministic system. Instead of ran-
domness, at the bottom we might have Unknowable Determinism. Every correct
picture of the actual microscopic state cannot be calculated by us until after it
has arrived, naturally.
An advantage of such reversible systems to their creators is that after the
detection of an extraordinary event, the process can be reversed to enable effi-
cient study of the exact cause.
Contents
Invited Papers
Quantum Circuits
Applications
Michael P. Frank(B)
This work was supported by the Laboratory Directed Research and Development
program at Sandia National Laboratories and by the Advanced Simulation and Com-
puting program under the U.S. Department of Energy’s National Nuclear Security
Administration (NNSA). Sandia National Laboratories is a multimission labora-
tory managed and operated by National Technology and Engineering Solutions of
Sandia, LLC., a wholly owned subsidiary of Honeywell International, Inc., for NNSA
under contract DE-NA0003525. Approved for public release, SAND2018-7205 C.
This paper describes objective technical results and analysis. Any subjective views
or opinions that might be expressed in the paper do not necessarily represent the
views of the U.S. Department of Energy or the United States Government.
c Springer Nature Switzerland AG 2018
J. Kari and I. Ulidowski (Eds.): RC 2018, LNCS 11106, pp. 3–33, 2018.
https://doi.org/10.1007/978-3-319-99498-7_1
4 M. P. Frank
1 Introduction
A core motivation for the field of reversible computation is Landauer’s Principle
[1], which tells us that each bit’s worth of information that is lost from a com-
putational process results in a (permanent) increase in thermodynamic entropy
of ΔS ≥ k ln 2, where k = kB is Boltzmann’s constant,1 with the correspond-
ing dissipation of ΔE ≥ kT ln 2 energy to heat, where T is the temperature
of the heat sink. By avoiding information loss, reversible computation bypasses
this limit on the energy efficiency of computing, opening the door to a future of
potentially unlimited long-term improvements in computational efficiency.2
The correctness of Landauer’s Principle has recently been empirically val-
idated [5–8], but the results of these experiments are unsurprising, given that
the validity of Landauer’s Principle can be shown to follow as a rigorous con-
sequence of basic facts of fundamental physics that have been known for over
a century, ever since pioneering work by such luminaries as Boltzmann and
Planck revealed the fundamentally statistical nature of entropy, summarized in
the equation S = k log W that is emblazoned on Boltzmann’s tombstone.3 As
we will show in some detail, Landauer’s Principle follows directly and rigor-
ously from the modern statistical-mechanical understanding of thermodynamics
(which elaborates upon Boltzmann’s view), augmented only by a few mathemat-
ical concepts from information theory, along with the most basic understanding
of what is meant by a computational process.
However, despite this underlying simplicity, certain subtleties regarding the
proper interpretation of the Principle remain frequently misunderstood; I have
discussed some of these in earlier papers [9–13], and will elaborate upon another
one in the present paper. Issues mentioned in these works include:
is the ejection of correlated bits from the computational state, since a ther-
mal environment cannot be expected to preserve those correlations in any
way that is accessible to human modeling. So really, it is the loss of prior cor-
relations that is the ultimate basis for the consideration of information loss
and entropy increase in Landauer’s Principle. I addressed this issue briefly in
previous presentations [19,20]; in this paper, I elaborate on it in more detail.
The rest of this paper is organized as follows. Section 2 reviews some basic
mathematical concepts of entropy, information, and computation. Section 3 dis-
cusses the connection of these concepts with physics in detail, and gives exam-
ples of physical systems that illustrate the fundamental appropriateness of these
abstract concepts for modeling the practical physical circumstances that we use
them to describe. We use this understanding of basic physics to prove Landauer’s
Principle, and discuss its implications for the energy efficiency of future reversible
and irreversible computing technologies. Section 4 briefly reviews some of the
existing laboratory studies that have validated Landauer’s Principle empirically.
Section 5 concludes with some suggestions for future work.
4
Intuitively, the more different values vi there are, the more unlikely or improbable
each individual value would seem to be, proportionally—not knowing anything else
about the situation.
Physical Foundations of Landauer’s Principle 7
Fig. 1. Surprise and heaviness functions. (a) Plot of surprise s (in units of k =
log e) as a function of probability p. Note that heaviness h = ps is given by the area of
a rectangle drawn between the origin and a point on this curve—if we imagine that the
rectangle were a flat sheet of physical material of uniform density and thickness, then
its physical heaviness would indeed be proportional to its area. (b) Plot of heaviness (in
k) as a function of probability. Note that the maximum heaviness of k/e is associated
with events of improbability e.
This makes sense intuitively, since it is the (weighted) average value of the func-
tion f that we would expect to obtain if values of V were chosen at random in
proportion to their probabilities.
heavily our uncertainty concerning the actual value might weigh on our minds,
if we dearly desired to know the value, but did not yet. In simpler terms, we
might say it corresponds to a lack of knowledge or amount of uncertainty or
amount of unknown information. It is the extent to which our knowledge of the
variable’s value falls short of perfection. We’ll explain later why physical entropy
is, in fact, the very same concept.
It is easy to show that the entropy S(P ) of a probability distribution P over
any given value set V has a maximum value of S(P ) = Ŝ(V ) = Ŝ(V) = log n
(where recall n = |V|) when all of the probabilities pi are equal, corresponding
to our original scenario, where only the number n of alternative values is known.
In contrast, whenever the probability p(vi ) of a single value vi approaches 1, the
entropy of the whole probability distribution approaches its minimum of 0 (no
lack of knowledge, i.e. full knowledge of the variable’s value).
We can also write S(V ) to denote the entropy S(P ) of a discrete variable V
under a probability distribution P over the values of the variable that is implicit.
The conditional entropy of X given Y tells you the expected value of what
the entropy of your resulting probability distribution over X would become if
you learned the value of Y . That this is true is a rigorous theorem (which we’ll
call the conditional entropy theorem) that is provable from the definitions above.
if the exact value of that variable were to be learned. We just saw that the max-
imum possible entropy, in relation to a given discrete variable V with a finite
value set V, is Ŝ(V ) = log |V|, that is, the logarithm of the number of possible
values of the variable, which is the same as the surprise that would result from
learning the value, starting from no knowledge about the value. Thus, in any
given epistemological situation (characterized by a probability distribution P ) in
which the entropy may be less than that maximum, the natural definition of the
amount of knowledge that we have, or in other words the (amount of) (known)
information K(P ) = K(V ) (also called negentropy or extropy) that we have
about the value of the variable V , is simply given by the difference between the
maximum entropy, and the actual entropy, given our probability distribution P :
in other words, it measures that part of our total knowledge K(X, Y ) about
the joint distribution P (X, Y ) that is not reflected in the separate distributions
P (X) and P (Y ). It is also the difference between the total entropies of (the
probability distributions over) X and Y considered separately, and the entropy of
the two variables considered jointly. It is also a theorem that I(X; Y ) = H(X) −
H(X|Y ), the amount by which the entropy of X would be reduced by learning Y
(and vice-versa). Mutual information is always positive, and always less than or
equal to the total known information K(X, Y ) in the joint distribution P (X, Y )
over the two variables X, Y taken together. It can be considered the amount of
information that is shared or redundant between variables X and Y , in terms of
our knowledge about them. It can be considered to be a way of quantifying the
6
I gave a detailed example of this information capacity relation (Eq. 8) in [22].
Physical Foundations of Landauer’s Principle 11
7
Note that this information-theoretic concept of correlation differs from, and is more
generally applicable than, a statistical correlation coefficient between scalar numeric
variables. General discrete variables do not require any numerical interpretation.
12 M. P. Frank
constant”), to also quantify, for the first time, the number of distinguishable
microstates for a given physical system, and thereby determine the exact rela-
tionship between Boltzmann’s statistical concept of entropy, and Clausius’ pre-
existing thermodynamic concept of entropy.
Planck saw that, in the discrete case, a maximum statistical entropy S = Ŝ
can be derived and expressed as
S = k log W, (12)
Further, all of the vast amount of 20th-century experimental physics that uti-
lizes Boltzmann’s constant also fundamentally rests (directly or indirectly) on the
statistical-mechanical understanding of entropy. Moreover, the entire structure
of quantum theory rests, at its core, on the discreteness of states discovered by
Planck, which itself was derived from statistical-mechanical assumptions. Infor-
mation theory is, fundamentally, the basic language for quantifying knowledge
and uncertainty in any statistically-described system, including physical systems.
And today’s quantum physics is, at root, just the intellectual heir of Boltzmann’s
statistical physics, in its most highly-developed, modern form. That’s how deep
the connection between information theory and physics goes.
The point of reviewing this history is simply to underscore this paper’s main
message, which is that to deny the validity of Landauer’s Principle would be to
repudiate much of the progress in theoretical and applied physics that has been
made in the more than 150 years that have elapsed since Boltzmann’s earliest
papers.
ARGUMENTO
Accediendo a los ruegos de Adán, cuéntale Rafael cómo y por qué fue creado este mundo
que habiendo Dios expulsado del cielo a Satán y a sus ángeles, declaró que le placía crear otro
mundo y otras criaturas que habitasen en él; y así envía a su Hijo circundado de gloria y
acompañado de angélicos coros, para que en el espacio de seis días realice la obra de la
creación. Al compás de sus himnos celebran los ángeles esta nueva maravilla, y la reascensión
del Hijo a los cielos.
ARGUMENTO
Adán hace algunas preguntas sobre los movimientos celestes, a las que contesta el Ánge
con palabras dudosas, aconsejándole que procure informarse de cosas más dignas de saberse
Persuádese de ello Adán; pero deseoso de tener a Rafael más tiempo consigo, le refiere todo lo
que recuerda su memoria desde que fue creado, y cómo entró en el Paraíso; su conferencia
con Dios respecto a la soledad y a la compañía que pudiera convenirle; su primer encuentro y
su desposorio con Eva; y prosigue discurriendo sobre este punto con el Ángel que después de
hacerle algunas amonestaciones, regresa al cielo.