deletions | additions
diff --git a/12discussion.tex b/12discussion.tex
index 34815cb..ccba9d6 100644
--- a/12discussion.tex
+++ b/12discussion.tex
...
by the particular context.''
\end{quote}
%
%% Their article describes a number of criteria relevant to writing
%% good design patterns, e.g. \emph{Clear target audience},
%% \emph{Visible forces}, and \emph{Relationship to other patterns}.
%
Applying the solution achieves this resolution of forces, and a design
pattern shows how this works. The design pattern itself achieves
something further: it encapsulates knowledge in a brief, shareable
form, often with meaningful relationships to other patterns. Tracing
the steps involved, we see that the creation of a new design pattern
is always somewhat serendipitous (Figure \ref{fig:pattern-schematic};
compare Figure \ref{fig:1b}).
\vspace{-.3cm}
\begin{figure}
\input{pattern-schematic-tikz.tex}
\caption{The components of design patterns mapped to our process schematic\label{fig:pattern-schematic}}
\end{figure}
\vspace{-.3cm} To van Andel's assertion that ``The very moment I can
plan or programme `serendipity' it cannot be called serendipity
anymore,'' we reply that we can certainly describe patterns -- and
programs -- with built-in indeterminacy. We can foster circumstances
that may make an unexpected happy outcome more likely. Figure
\ref{fig:va-pattern-figure} illustrates this with one van Andel's
patterns of serendipity, rewritten using the standard design pattern
template. In future work, we intend to build a more complete
serendipity pattern language -- and to use this within autonomous
programming systems that transform raw data into ``strategic data.''
\setlist[description]{font=\normalfont\itshape}
\begin{figure}[!h]
\setlist[description]{font=\normalfont\itshape}
\begin{mdframed}
\vspace{2mm}
\textbf{\emph{Successful error}}~
...
\caption{Standard design pattern template applied to van Andel's \em{Successful error}\label{fig:va-pattern-figure}}
\end{figure}
%
%% Their article describes a number of criteria relevant to writing
%% good design patterns, e.g. \emph{Clear target audience},
%% \emph{Visible forces}, and \emph{Relationship to other patterns}.
%
Applying the solution achieves this resolution of forces, and a design
pattern shows how this works. The design pattern itself achieves
something further: it encapsulates knowledge in a brief, shareable
form, often with meaningful relationships to other patterns. Tracing
the steps involved, we see that the creation of a new design pattern
is always somewhat serendipitous (Figure \ref{fig:pattern-schematic};
compare Figure \ref{fig:1b}).
\vspace{-.3cm}
\begin{figure}
\input{pattern-schematic-tikz.tex}
\caption{The components of design patterns mapped to our process schematic\label{fig:pattern-schematic}}
\end{figure}
\vspace{-.3cm} To van Andel's assertion that ``The very moment I can
plan or programme `serendipity' it cannot be called serendipity
anymore,'' we reply that we can certainly describe patterns -- and
programs -- with built-in indeterminacy. We can foster circumstances
that may make an unexpected happy outcome more likely. Figure
\ref{fig:va-pattern-figure} illustrates this with one van Andel's
patterns of serendipity, rewritten using the standard design pattern
template. In future work, we intend to build a more complete
serendipity pattern language -- and to use this within autonomous
programming systems that transform raw data into ``strategic data.''
% Is ``having a stretch goal'' an example of a serendipity pattern? I think so!
diff --git a/1introduction.tex b/1introduction.tex
index 84f8e27..6e3799f 100644
--- a/1introduction.tex
+++ b/1introduction.tex
...
in this field -- even though serendipity has played a well-documented
role in historical instances of scientific and technical creativity.
One reason for this omission may be that the field of computational
creativity has tended to focus on artistic creativity, conceptualised in such a way that creative outputs are largely under the direct control of the creative agent. However, serendipity is increasingly seen as relevant
within in the arts
\cite{mckay-serendipity} \cite{mckay-serendipity}, in the tech industry \cite{rao2015breaking}, and
other enterprises, where it is elsewhere. In the workplace,
serendipity may be encouraged with methods drawn from
fields ranging from architecture
to and data science \cite{kakko2009homo,engineering-serendipity}.
An interdisciplinary perspective on the phenomenon of serendipity
promises further illumination.
This paper follows and expands
\citeA{pease2013discussion}, where many of the ideas that are developed here were first presented. \citeA{pease2013discussion}. The current paper reassesses and updates this earlier work, developing a robust computational characterisation of serendipity for computer modelling and system evaluation. New claims are advanced, positioning serendipity as a fundamental concept in computational creativity, with
exciting the potential to play a key role in computational intelligence more broadly. There is particularly interesting potential for serendipity within computational systems whose processes involve interaction with users, and autonomous systems that make use of a multi-agent framework.\footnote{It should not be assumed that a system that can accommodate user interaction would directly lead to serendipity; take for example the use of a calculator, where potential for serendipity through user interaction is minimal at best.}
Serendipity is centred on reevaluation. For example, a
non-sticky ``superglue'' that no one was quite sure how to use turned
diff --git a/2a-etymology.tex b/2a-etymology.tex
index 85762a7..bf67312 100644
--- a/2a-etymology.tex
+++ b/2a-etymology.tex
...
% \section{Literature review}
\subsection{Etymology and selected definitions} \label{sec:overview-serendipity}
The English term ``serendipity'' derives from the 1302 long poem \emph{Eight Paradises}, written in Persian by the Sufi poet Am\={\i}r Khusrow in Uttar Pradesh, India.\footnote{\url{http://en.wikipedia.org/wiki/Hasht-Bihisht}} In the English-speaking world, its first chapter became known as ``The Three Princes of Serendip'', where ``Serendip'' represents the Old Tamil-Malayalam word for Sri Lanka (%{\tam சேரன்தீவு},
\emph{Cerantivu}, island of the Ceran kings).
%
The term ``serendipity'' is first found in a 1757 letter by Horace Walpole to Horace Mann:
\begin{quote}
\emph{``This discovery is almost of that kind which I call serendipity, a very expressive
word} \ldots \emph{You will understand it better by the derivation than by the
definition. I once read a silly fairy tale, called The Three Princes of Serendip:
as their Highness travelled, they were always making discoveries, by accidents
\& sagacity, of things which they were not in quest of}[.]''~\cite[p. 633]{van1994anatomy}
\end{quote}
The term became more widely known in the 1940s through studies of serendipity as a factor in scientific discovery, surveyed by Robert Merton and Elinor Barber \citeyear{merton} in ``The Travels and Adventures of Serendipity, A Study in Historical Semantics and the Sociology of Sciences''. Merton \citeyear{merton1948bearing} \cite[pp. 195--196]{merton} describes a generalised ``serendipity pattern'' and its constituent parts:
\begin{quote}
``\emph{The serendipity pattern refers to the fairly common experience of observing an \emph{unanticipated}, \emph{anomalous} \emph{and strategic} datum which becomes the occasion for developing a new theory or for extending an existing theory.}''~\cite[p. 506]{merton1948bearing}~{[}emphasis in original{]}
%% The datum [that exerts a pressure for initiating theory] is, first of all, unanticipated. A research directed toward the test of one hypothesis yields a fortuitous by-product, an unexpected observation which bears upon theories not in question when the research was begun.
%% Secondly, the observation is anomalous, surprising, either because it seems inconsistent with prevailing theory or with other established facts. In either case, the seeming inconsistency provokes curiosity; it stimulates the investigator to "make sense of the datum," to fit it into a broader frame of knowledge....
%% And thirdly, in noting that the unexpected fact must be "strategic," i. e., that it must permit of implications which bear upon generalized theory, we are, of course, referring rather to what the observer brings to the datum than to the datum itself. For it obviously requires a theoretically sensitized observer to detect the universal in the particular.
\end{quote}
In 1986, Philippe Qu\'eau described serendipity as ``the art of
finding what we are not looking for by looking for what we are not
finding'' (\citeNP{eloge-de-la-simulation}, as quoted in
\citeNP[p. 121]{Campos2002}). Campbell
\citeyear{campbell2005serendipity} defines it as ``the rational
exploitation of chance observation, especially in the discovery of
something useful or beneficial.'' Pek van Andel
\citeyear[p. 631]{van1994anatomy} describes it simply as ``the art of
making an unsought finding.''
Roberts \citeyear[pp. 246--249]{roberts} records 30 entries for the term ``serendipity'' from English language dictionaries dating from 1909 to 1989.
%
Classic definitions require the investigator not to be aware of the problem they serendipitously solve, but this criterion has largely dropped from dictionary definitions. Only 5 of Roberts' collected definitions explicitly say ``not sought for.'' Roberts characterises ``sought findings'' in which an accident leads to a discovery with the term \emph{pseudoserendipity} \cite{chumaceiro1995serendipity}.
%
While Walpole initially described serendipity as an event
(i.e., a kind of discovery), it has
since been reconceptualised as a psychological attribute, a matter of
sagacity on the part of the discoverer: a ``gift'' or ``faculty'' more
than a ``state of mind.'' Only one of the collected definitions, from
1952, defined it solely as an event, while five define it as both
event and attribute.
However, numerous historical examples exhibit features of
serendipity and develop on a social scale rather than an individual
scale. For instance, between Spencer Silver's creation of high-tack,
low-adhesion glue in 1968, the invention of a sticky bookmark in 1973,
and the eventual launch of the distinctive canary yellow re-stickable
notes in 1980, there were many opportunities for
Post-it\texttrademark\ Notes \emph{not} to have come to be
\cite{tce-postits}. Merton and Barber argue that the
psychological perspective needs to be integrated with a
\emph{sociological} one.\footnote{ ``For if chance favours prepared
minds, it particularly favours those at work in microenvironments
that make for unanticipated sociocognitive interactions between
those prepared minds. These may be described as serendipitous
sociocognitive microenvironments'' \cite[p. 259--260]{merton}.}
Large-scale scientific and technical projects generally rely on the
convergence of interests of key actors and on other cultural factors.
For example, Umberto Eco \citeyear{eco2013serendipities} describes the
historical role of serendipitous mistakes and falsehoods in the
production of knowledge.
It is important to note that serendipity is usually discussed within
the context of \emph{discovery}, rather than \emph{creativity},
although in typical parlance these terms are closely related
\cite{jordanous12jims}. In the definition of serendipity that we present in Section \ref{sec:our-model}, we make use
of Henri Bergson's distinction:
\begin{quote}
%% \emph{``La d\'ecouverte porte sur ce qui existe d\'ej\`a, actuellement
%% ou virtuellement ; elle \'etait donc s\^ure de venir t\^ot ou
%% tard. L'invention donne l'\^etre \`a ce qui n'\'etait pas, elle
%% aurait pu ne venir jamais.''}
``\emph{Discovery, or uncovering, has to do with what already exists,
actually or virtually; it was therefore certain to happen sooner
or later. Invention gives being to what did not exist; it might
never have happened.}''~\cite[p. 58]{bergson2010creative}
\end{quote}
As we have indicated, serendipity would seem to require features of
both discovery and invention: that is, the discovery of something
unexpected and the invention of an application for the same. Both
processes can be seen as ongoing and diverse, which underscores the
relationship between serendipity and creativity. According to Arthur
Cropley \citeyear{cropley2006praise}, creative thinking involves
``novelty generation followed by (or accompanied by) exploration of
the novelty from the point of view of workability, acceptability, or
similar criteria, in order to determine if it is effective.''
Following \citeA{austin2003chase}, Cropley understands serendipity to
describe the case of a person who ``stumbles upon something novel and
effective when not looking for it.'' Nearby categories are
\emph{blind luck}, the \emph{luck of the diligent} (or
pseudoserendipity) and \emph{self-induced luck}; however, Cropley
questions ``whether it is a matter of luck at all'' because of the
work and knowledge involved in the process of assessment.
%
The perspective developed here would sharpen these understandings in two ways:
firstly, we point out that work is involved in both discovery and
invention even when chance plays a role, and secondly, we defer true ``novelty'' to the invention phase.
%% In other words, serendipity involves creative making. Furthermore, we
%% emphasise the importance of active, agential discernment over more
%% passive stumbling.
diff --git a/2literature.tex b/2literature.tex
index 4353fe0..0c4ed61 100644
--- a/2literature.tex
+++ b/2literature.tex
...
as their Highness travelled, they were always making discoveries, by accidents
\& sagacity, of things which they were not in quest of}[.]''~\cite[p. 633]{van1994anatomy}
\end{quote}
The
same story formed part of the inspiration for Voltaire's \emph{Zadig}, and ``the method of Zadig'' was referenced in late 19th Century philosophy of science \cite{huxley1894science}.
Walpole's term
``serendipity'' became more widely known in the 1940s through studies
on the sociology of
serendipity as a factor in scientific discovery, surveyed science by Robert Merton and Elinor
Barber \citeyear{merton} Barber, collected in
``The \emph{The Travels and Adventures of Serendipity, A Study in Historical Semantics and the Sociology of
Sciences''. Sciences}. Merton \citeyear{merton1948bearing} \cite[pp. 195--196]{merton} describes a generalised ``serendipity pattern'' and its constituent parts:
\begin{quote}
``\emph{The serendipity pattern refers to the fairly common experience of observing an
\emph{unanticipated}, \emph{anomalous} \emph{and strategic} \emph{\textbf{unanticipated}}, \emph{\textbf{anomalous}} \emph{\textbf{and strategic}} datum which becomes the occasion for developing a new theory or for extending an existing theory.}''~\cite[p. 506]{merton1948bearing}~{[}emphasis in original{]}
%% The datum [that exerts a pressure for initiating theory] is, first of all, unanticipated. A research directed toward the test of one hypothesis yields a fortuitous by-product, an unexpected observation which bears upon theories not in question when the research was begun.
%% Secondly, the observation is anomalous, surprising, either because it seems inconsistent with prevailing theory or with other established facts. In either case, the seeming inconsistency provokes curiosity; it stimulates the investigator to "make sense of the datum," to fit it into a broader frame of knowledge....
%% And thirdly, in noting that the unexpected fact must be "strategic," i. e., that it must permit of implications which bear upon generalized theory, we are, of course, referring rather to what the observer brings to the datum than to the datum itself. For it obviously requires a theoretically sensitized observer to detect the universal in the particular.
...
It is important to note that serendipity is usually discussed within
the context of \emph{discovery}, rather than \emph{creativity},
although in
typical everyday parlance these terms are closely related
\cite{jordanous12jims}. In the definition of serendipity that we present in Section \ref{sec:our-model}, we make use
of Henri Bergson's distinction:
\begin{quote}
...
unexpected and the invention of an application for the same. Both
processes can be seen as ongoing and diverse, which underscores the
relationship between serendipity and creativity. According to Arthur
Cropley \citeyear{cropley2006praise}, Cropley, creative thinking involves
``novelty \begin{quote}
``{[}N{]}\emph{ovelty generation followed by (or accompanied by) exploration of the novelty from the point of view of workability, acceptability, or similar criteria, in order to determine if it is
effective.'' effective.}'' \citeyear{cropley2006praise}
\end{quote}
Following \citeA{austin2003chase}, Cropley understands serendipity to
describe the case of a person who ``stumbles upon something novel and
effective when not looking for it.'' Nearby categories are
...
questions ``whether it is a matter of luck at all'' because of the
work and knowledge involved in the process of assessment.
%
The perspective developed here
would sharpen sharpens these understandings in two ways:
firstly, we point out that work is involved in both
discovery and
invention even phases of the process (even when chance plays a
role, role), and secondly, we defer true ``novelty'' to the invention phase.
%% In other words, serendipity involves creative making. Furthermore, we
%% emphasise the importance of active, agential discernment over more
%% passive stumbling.
...
investigator also needs to have a suitable frame of mind, one that
is ready to make a jump into the unknown as the world changes. In a
certain sense it is necessary to be able to ``overcome'' situated
cognition, or at least be ready to revise the
plan approach as the
situation changes \cite{bereiter1997situated}.
\end{itemize}
\begin{itemize}
...
\subsection{Related work} \label{sec:related}
An active research community investigating computational models of serendipity exists in the field of information
retrieval, retrieval \cite{foster2003serendipity}, and specifically, in recommender systems \cite{Toms2000}. In
this the recommender system domain, \citeA{Herlocker2004} and \citeA{McNee2006} view serendipity as an important factor for user satisfaction, alongside accuracy and diversity. Serendipity in recommendations is
understood to imply that the system suggests \emph{unexpected} items, which the user considers to be \emph{useful}, \emph{interesting}, \emph{attractive} or \emph{relevant}.
% \cite{Herlocker2004} \cite{Lu2012},\cite{Ge2010}.
Definitions differ as to the requirement of \emph{novelty}; \citeA{Adamopoulos2011}, for example, describe systems that suggest items that may already be known, but are still unexpected in the current context. While standardised measures such as the $F_1$-score or the (R)MSE are used to determine the \emph{accuracy} of a recommendation (i.e.~whether the recommended item is very close to what the user is already known to prefer), there is no common agreement on a measure for serendipity yet, although there are several proposals \cite{Murakami2008, Adamopoulos2011, McCay-Peet2011,iaquinta2010can}.
...
Recent work has examined related topics of \emph{curiosity}
\cite{wu2013curiosity} and \emph{surprise} \cite{grace2014using} in
computing. This latter work seeks to ``adopt methods from the field
of computational creativity [$\ldots$] to the generation of scientific hypotheses.'' This is
an a useful example of an effort focused on computational \emph{invention}.
Paul Andr{\'e} et al.~\citeyear{andre2009discovery} have examined
serendipity from a design perspective. Like us, these authors
proposed a two-part model encompassing ``the chance encountering of
information, and the sagacity to derive insight from the encounter.''
According to Andr\'e et al., the first phase is the one that has most
frequently been automated -- but they suggest that computational
systems should be developed that support both aspects. They
specifically suggest to focus on representational features:
\emph{domain expertise} and a \emph{common language model}.
These features seem to exemplify
aspects of the \emph{prepared mind}. However, as we mentioned above,
the \emph{bridge} is a distinct process that mental preparation can
support, but that it does not necessarily fully determine. For example, participants in
a poetry workshop may possess a very limited understanding of each
other's aims or of the work they are critiquing, and may as a
consequence talk past one another to a greater or lesser degree --
while nevertheless finding the overall process of participating in the
workshop illuminating and rewarding (often precisely because
such misunderstandings elucidate poor communication choices!).
Various social strategies, ranging from Writers Workshops to open
source software, pair programming, and design charettes
\cite[p. 11]{gabriel2002writer} have been developed to exploit similar
emergent effects and to develop \emph{new} shared language. We have
recently investigated the feasibility of using
designs of this sort in multi-agent systems that learn by sharing and
discussing partial understandings \cite{corneli2015computational,corneli2015feedback}.
\citeA{robot-rendezvous} develop a discussion of serendipitous
rendezvous in a multi-agent system for a graph exploration problem, in
which ``[h]aving more data about their colleagues, better decisions
are made about the potential serendipity path.'' This has some
similarity to the discursive scenario described above, and shows that
\emph{asymmetric partial knowledge} can support serendipitious
findings. These examples suggest that a distinction between emergent
knowledge of other actors and knowledge about an underlying domain may
be useful -- although the distinction may be less relevant if
the underlying domain itself has dynamic and emergent features.
\emph{Social coordination} among human users of information systems is
a current research topic. \citeA{rubin2010everyday} point out that
naive end users often \emph{talk about} serendiptious occurrences,
which presents another route for research and evaluation.
The {\sf SerenA} system developed by Deborah Maxwell et al.~\citeyear{maxwell2012designing} offers a case study of several of the points discussed above.
This system is designed to support serendipitous discovery for its (human) users
\cite{forth2013serena}. The authors rely on a process-based model of
serendipity \cite{Makri2012,Makri2012a} that is derived from user
studies, including interviews with 28 researchers, looking for
instances of serendipity from both their personal and professional
lives. This material was coded along three dimensions:
\emph{unexpectedness}, \emph{insightfulness}, and \emph{value}. This
research aims to support the process of forming bridging connections
from an unexpected encounter to a previously unanticipated but valuable
outcome. The theory focuses on the acts of \emph{reflection}
that foment both the creation of a bridge and estimates of the
potential value of the result.
%
While this description touches on all Creativity and serendipity are closely linked. Indeed, Cropley's
definition of
the features creative thinking quoted in Section
\ref{sec:by-example} has a lot in common with understanding of
serendipity that is developed in this paper. However, in
standard usage the terms are hardly synonymous.
%
On our
model, {\sf
SerenA} largely matches understanding, serendipity is neither a system trait (like
skill or autonomy), nor an attribute of an artefact (like novelty
or usefulness). Rather, it the
term is a description
offered by of a
certain kind of process.
Like us, Paul Andr{\'e} et al.~\citeyear{andre2009discovery}
propose a two-part model of
discovery-focused systems, in which
the user experiences an ``aha'' moment serendipity encompassing ``the chance
encountering of information, and
takes the
creative steps sagacity to
realise derive insight
from the
result. {\sf SerenA}'s primary computational method is encounter.'' According to
search outside of Andr\'e et al., the
normal search parameters in order first
phase is the one that has most frequently been automated -- but
these authors suggest that computational systems should be
developed that support both aspects. They specifically suggest
to
engineer
potentially serendipitous (or at least pseudo-serendipitous)
encounters.
%% Another
%% earlier related example of pursue this
sort of system is {\sf Max}, created
%% work by
Figueiredo and Campos \citeyear{Campos2002}. The user emailed {\sf
%% Max} developing systems with
an existing list of interests better
representational features: \emph{domain expertise} and
{\sf Max} would return a
%% web page \emph{common language model}.
These features seem to exemplify aspects of the \emph{prepared
mind}. However, as we mentioned above, the \emph{bridge} is a
distinct step in the process that
might also be mental preparation can support,
but that it does not necessarily fully determine. For instance,
persons involved in a dialogue may understand each other quite
poorly, while nevertheless finding the conversation interesting
and rewarding experience. For one thing, misunderstandings can
present learning opportunities! Various social strategies,
ranging from Writers Workshops to open source software and
from design charettes to group therapy have been
developed to exploit similar emergent effects and to develop
\emph{new} shared language
\cite{gabriel2002writer,seikkula2014open}. We have recently
investigated the feasibility of using designs of
interest. Other this sort in
multi-agent systems
with similar
%% support for serendipitous discovery involve searching for analogies
%% \cite{Donoghue2002,Donoghue2012}) as well as content \cite{Iaquinta2008}. that learn by sharing and discussing partial
understandings
\cite{corneli2015computational,corneli2015feedback}.
Figueiredo and Campos \citeyear{Figueiredo2001} describe serendipitous ``moves'' from one
problem to another, which transform a problem that cannot be solved
...
(or get their systems to record) what problem a given system solves,
and the degree to which the computer was responsible for coming up
with this problem.
%
As \citeA[p. 69]{pease2013discussion} remark, anomaly detection and
outlier analysis are part of the standard machine learning toolkit --
but recognising \emph{new} patterns and defining \emph{new} problems
is more ambitious. Establishing complex analogies between evolving
problems and solutions is one of the key strategies used by teams of
human designers \cite{Analogical-problem-evolution-DCC}.
Kazjon Grace
\citeyear{kaz-thesis} presents a computational model of the The creation
of new
concepts and interpretations, but this work did include the
ability to create new higher order relationships necessary for complex
analogies. New patterns and higher-order analogies were considered in
Hofstadter and Mitchell's
work on {\sf Copycat} and the subsequent
{\sf Metacat}, but these systems
operated operate in a simple and fairly
abstract ``microdomain''
\cite{hofstadter1994copycat,DBLP:journals/jetai/Marshall06}.
%% More
%% recent work %
Serendipity and problem-creation are related, and turning over
increasing responsibility to the machine will be important if we
want to foster the possibility of genuine surprises.
The {\sf SerenA} system developed by Deborah Maxwell et
al.~\citeyear{maxwell2012designing} offers a case study in
this tradition some
of these concepts. This system is
surveyed in
%% \cite{eric-nichols-thesis}. designed to support
serendipitous discovery for its (human) users
\cite{forth2013serena}. The
relationship between authors rely on a process-based
model of serendipity \cite{Makri2012,Makri2012a} that is derived
from user studies that draws on interviews with 28 researchers,
who were asked to look for instances of serendipity
from both
their personal and
novel problems receives
considerable attention professional lives. The research aims to
support the formation of bridging connections from an unexpected
encounter to a previously unanticipated but valuable outcome.
The theory focuses on the acts of reflection that support both
the creation of a bridge, and the estimation of the potential
value of the result.
%
While this description touches on all of the features of our model, {\sf
SerenA} largely matches the description offered by Andr{\'e} et
al.~\citeyear{andre2009discovery} of discovery-focused systems, in
which
the
current work, since user experiences an ``aha'' moment and takes the
creative steps to realise the result. {\sf SerenA}'s primary computational method is to
search outside of the normal search parameters in order to engineer
potentially serendipitous (or at least pseudo-serendipitous)
encounters.
Computer-supported serendipity has been well-studied, but purely
computational serendipity, much less so. This may partly be due
to the absence of clear criteria (which something we
want address in
the current paper) but also due to
sociological factors in
computing research. Existing standards for assessing
computational creativity have historically focused on product
evaluations \cite{ritchie07}, and are increasingly
turn over responsibility complemented
by methods that consider process
\cite{colton2008creativity,colton-assessingprogress} or a
combination of product and process \cite{jordanous:12}. However,
processes that arise outside of the control of the system (and
ultimately, the researcher) are not often considered. They may
be seen to ``invalidate'' research into creativity, rendering it
unscientific. But we would argue we need more research into
autonomous creative systems. This requires a different mindset,
and a different approach to system building and evaluation.
\begin{quote}
``\emph{Tinkering is a process of serendipity-seeking that does not just tolerate uncertainty and ambiguity, it requires it. When conditions for
creating it are right, the result is a snowballing effect where pleasant surprises lead to more pleasant surprises.}'' \cite{rao2015breaking}
%% What makes this a problem-solving mechanism is diversity of individual perspectives coupled with the law of large numbers (the statistical idea that rare events can become highly probable if there are enough trials going on). If an increasing number of highly diverse individuals operate this way, the chances of any given problem getting solved via a serendipitous new idea slowly rises. This is the luck of networks.
%% Serendipitous solutions are not just cheaper than goal-directed ones. They are typically more creative and elegant, and
maintaining require much less conflict. Sometimes they are so creative, the fact that they even solve a
prepared mind particular problem becomes hard to
recognize. For example, telecommuting and video-conferencing do more to “solve” the problem of fossil-fuel dependence than many alternative energy technologies, but are usually understood as technologies for flex-work rather than energy savings.
%% Ideas born of tinkering are not targeted solutions aimed at specific problems, such as “climate change” or “save the middle class,” so they can be applied more broadly. As a result, not only do current problems get solved in unexpected ways, but new value is created through surplus and spillover. The clearest early sign of such serendipity at work is unexpectedly rapid growth in the adoption of a new capability. This indicates that it is being used in many unanticipated ways, solving both seen and unseen problems, by both design and “luck”.
\end{quote}
If we control the system, at bottom the
machine. best we can hope for is
``pleasant unsurprises.'' Understanding serendipity may help
build autonomous systems without ``unpleasant surprises,'' a
topic of considerable current concern
\cite{philosophy-machine-morality,machine-ethics-status}.
diff --git a/3model.tex b/3model.tex
index d9122de..bc6f418 100644
--- a/3model.tex
+++ b/3model.tex
...
``serendipitous.'' Summarising, we propose the following:
\begin{ndef}
\emph{(1) Within a system with a prepared mind, a previously uninteresting serendipity trigger arises due to circumstances that the system does not control, and is classified as interesting by the system; and,}
\emph{(2) The
system, by subsequently processing this system uses the trigger and
background information prior preparation, together with relevant
reasoning, computational processing, networking,
or and experimental techniques,
obtains to obtain a novel result that is evaluated favourably by the system or by external sources.}
\end{ndef}
diff --git a/8cc.tex b/8cc.tex
index bce2ba6..4a329cc 100644
--- a/8cc.tex
+++ b/8cc.tex
...
discovery. Section \ref{sec:our-model} distilled these elements into a computational model,
culminating in a method for evaluating computational serendipity in Section \ref{specs-overview}.
In order to apply these criteria, it is important to clearly delineate
the scope of the system being evaluated. For example, a standard
spell-checking system might suggest a substitution that the user deems
especially fortuitous; we may agree that serendipity has occurred, but
we would not attribute serendipity to the spell-checker.
\citeA{pease2013discussion} used an earlier variant
these the SPECS criteria
to analyse three examples of potentially serendipitous behaviour:
dynamic investigation problems, model generation, and poetry
flowcharts.
Three additional Using our updated criteria, we discuss two new examples
are discussed below using the revised
criteria. below, and revisit poetry flowcharts as a third example. As Campbell
\citeyear{campbell2005serendipity} writes, ``serendipity presupposes a
smart mind,'' and these examples suggest potential directions for
further work in computational intelligence.
Before describing these examples, as a baseline, we introduce the
notion of \emph{minimally serendipitous systems}. According to our
standards, there are various ways to achieve a result with little or
no serendipity: if the observation was likely, if further developments
happened with little skill, and if the the value of the result was
low, then we would not say the outcome was serendipitous. We would be
prepared to attribute ``minimal serendipity'' to cases where the
observation was \emph{moderately} likely, \emph{some} skill or effort
was involved, and the result was only \emph{fairly good}. However,
for computational systems, if most of the skill involved lies with the
user, then there is little reason to call the system's operation
serendipitous -- even if it consistently does its job very well. For
example, machines can learn to recognise or approximate certain types
of patterns, but it is surprising when a computational system
independently finds an entirely new kind of pattern. Furthermore, the
position of the evaluator is important: a spell-checking system might
suggest a particularly fortuitous substitution, but we would not
expect the spell-checker to know when it was being clever. In such a
case, we may say serendipity has occurred, but not that we have a
serendipitous system.
%% If the system learns an $N$th fact or
%% If applied to a system which could be described as minimally
...
\end{quote}
In future versions of the system, autonomous evaluation could take over for the human evaluator. Once the interesting samples have been found, a \textbf{bridge} is then built to new results through the creation of new Improvisors. The \textbf{results} are the various musical improvisations produced by the fittest Improvisors (as well as, perhaps, the parameters that have been considered fittest).
%% The likelihood of serendipitous evolution is greatly enhanced by the
%% use of random mutation and crossover operations within the genetic
%% algorithm, which increase the diversity of the search space covered by
%% the system during evolution.
The probability of encountering any
particular pair of Improvisor and user evaluation is vanishingly low,
given the massive dimensions of this search space. However,
due to
the way the system's evolution works,
there will always be a
highest-scoring Improviser, whose parameters will be used to seed the
next round. So, in fact, the \textbf{chance} of the system
encountering a
serendipity trigger is high. The evolution of
Improvisors captures a sense of \textbf{curiosity} about how to
satisfy the musical tastes of a particular human user who identifies
certain Improvisors as interesting. The \textbf{sagacity} of the
system's system corresponds to its methods for enhancing the likelihood that
the user will appreciate a given Improvisor's music (or similar music)
over time. However, with little basis for comparison, we can only say
that these dimensions present to ``typical'' degree. The aim of the
...
measure of
$\mathit{high}\times\mathit{moderate}\times\mathit{moderate}$, with
outcomes of moderate value, so that the system as a whole is ``not
very serendipitous.''
If However, if individual threads in the search
process were given more independence, they could be evaluated
separately, and some might prove to be more serendipitous than others.
In According to the
current case, SPECS criteria, we are not required to progress to
Part
C, C here, but for completeness, we note the following. The {\sf
GAmprovising} system does operate in \textbf{dynamic world},
assuming that the user's tastes may change. A more elaborate version
of the system that could cater to multiple users is not yet
implemented, but would be occupied with a considerably more complex
problem, spanning and integrating \textbf{multiple contexts}. The
system clearly engages with \textbf{multiple tasks}, but these are
largely separate, for instance, one global fitness function is used,
rather than evolving a local fitness function for each user along with
their ratings. \textbf{Multiple influences} are present but currently
only at compile time, in the design of the fitness function, and the
selection of musical parameters that can later be set. Greater
dynamism in future versions of the system would be likely to increase
its potential for serendipity.
\subsection{Case Study: Next-generation recommender systems} \label{sec:nextgenrec}
% Stress distinction between serendipity on the system- vs. serendipity on the user's side.
As discussed in Section \ref{sec:related}, recommender systems are one
of the primary contexts in computing where serendipity is currently discussed. Serendipity, for current recommender systems, means suggesting items to a user that will be likely to introduce new ideas that are unexpected, but close to what the user is already interested in. As we noted, these systems mostly focus on supporting \emph{discovery} for the user -- but some architectures also seem to take account of \emph{invention} of new methods for making recommendations, e.g.~using Bayesian methods, as surveyed in \citeNP{shengbo-guo-thesis}. In light of our working definition of serendipity, we need distinguish serendipity on the user side from serendipity in the system itself.
Current recommendation techniques associate less popular items with high unexpectedness \cite{Herlocker2004,Lu2012}, and use clustering to discover latent structures in the search space, e.g., partitioning users into clusters of common interests, or clustering users and domain objects \cite{Kamahara2005,Onuma2009,Zhang2011}. But even in the Bayesian case, the system has limited autonomy.
Nevertheless, a A case for
developing giving more
autonomous autonomy to recommender systems can be made, especially in
complex complex, rapidly evolving, domains where hand-tuning is cost-intensive or infeasible.
With this challenge in mind, we investigate how serendipity could be achieved on the system side. In terms of our model, current systems have at least the makings of a \textbf{prepared mind}, comprising both a user- and a domain model, both of which can be updated dynamically. User behaviour (e.g.~following certain recommendations) or changes to the domain (e.g.~adding a new product) may serve as a potential \textbf{trigger} that could ultimately cause the system to discover a new way to make recommendations in the future. Note, however, that it is unexpected behaviour in aggregate, rather than a one-off event, that is
most likely to provide grounds for a \textbf{focus shift}. A \textbf{bridge} to a new kind of recommendation could be created by looking at exceptional patterns as they appear over time. For instance, new elements may have been introduced into the domain that do not cluster well, and clusters may appear in the user model that do not have obvious connections between them. A new approach that helps to address the organisational mission would constitute a serendipitous \textbf{result} on the system side.
The system has only imperfect knowledge of user preferences and
interests. At least relative to current recommender systems, the
...
tuned to search for flowcharts that generate poetry, as we discuss in
\cite{corneli2015computational}.
There is a strong role for The \textbf{chance}
even in the early
prototype, regarding of finding a novel successful combination of nodes
is fairly low, as this depends on both the output from certain nodes,
and in terms the combinatorial search strategy itself.
The chances Compared to
humans users of
{\sf FloWr}, the seems exceptionally \textbf{curious}
about finding
a novel
successful combination of nodes is fairly low. The system's
ongoing effort to discover new combinations
that work could be
interpreted as \textbf{curiosity}. The fact that it remembers of nodes. Remembering viable
combinations and
doesn't try avoiding combinations that are known not to work
could be thought presents only a modest degree of
as moderate \textbf{sagacity}. At the moment,
the
threshold system's criterion for
attributing \textbf{value} is
low, simply
a new that
the combination of nodes
that produces generates non-empty
output. If this can be accompanied by output; however an
explanation, external evaluator is not likely to judge these combinations as
useful. The associated ``likelihood score'' is
$\mathit{low}\times\mathit{low}\times\mathit{high}$, but the
value system
should be seen only as ``potentially serendipitous'' until there is
higher. a
more discriminating way to judge value.
The \textbf{dynamic world} the system operates in is dynamic in two
ways: first, in the straightforward sense that some of the input
...
\item The next-generation recommender systems we've envisioned need to
be able to make inferences from aggregate user behaviour. This
points to long-term considerations that go beyond the unique
serendipitous event. How
much leeway ``curious'' should these
system have systems be? One
obvious criterion is that short-term value should be allowed to
experiment? suffer as long as expected value is still higher.
\item The clearest way to enhance the serendipity of results from the
flowchart assembly process would be to impose more stringent, and
more meaningful, criteria for value. In addition to raising
challenges for autonomous evaluation (as in the evolutionary music
system case), this requirement would impose more sophisticated
constaints on processing in earlier
steps. steps, which would require the
system to become more sagacious.
\end{enumerate}
diff --git a/bibliography/biblio.bib b/bibliography/biblio.bib
index 15122dd..7b6b43a 100644
--- a/bibliography/biblio.bib
+++ b/bibliography/biblio.bib
...
Year = {2012}}
@book{zadig,
Author
= {Voltaire}, ={Voltaire},
Place = {London},
Publisher = {John Brindley},
Title = {Zadig, or the Book of Fate},
...
booktitle={Proceedings of the {F}irst {I}nternational {W}orkshop on {AI} and {F}eedback},
editor={Nardine Osman and Matthew Yee-King},
year={2015},
url={http://ceur-ws.org/Vol-1407/AInF2015paper2.pdf}}
@incollection{huxley1894science,
title={On the method of {Z}adig},
booktitle={Science and {H}ebrew {T}radition: {E}ssays},
author={Huxley, Thomas Henry},
volume={4},
year={1894},
publisher={Appleton}
}
@inproceedings{colton2008creativity,
title={Creativity {V}ersus the {P}erception of {C}reativity in {C}omputational Systems},
author={Colton, Simon},
booktitle={{AAAI} Spring Symposium: Creative Intelligent Systems},
pages={14--20},
year={2008}
}
@book{seikkula2014open,
title={{O}pen {D}ialogues and {A}nticipations -- {R}especting {O}therness in the {P}resent {M}oment},
author={Seikkula, Jaakko and Arnkil, Tom Erik},
year={2014},
location={Helsinki},
publisher={National Institute for Health and Welfare}
}
@book{rao2015breaking,
title={Breaking {S}mart: {S}eeking serendipity through technology},
author={Rao, Vekatesh},
year={2015},
publisher={Ribbonfarm, Inc.},
url={http://breakingsmart.com}
}
@article{bryson2015artificial,
title={Artificial Intelligence and Pro-Social Behaviour},
author={Bryson, Joanna J},
year={2015},
publisher={Springer}
}
@article{philosophy-machine-morality,
year={2014},
issn={2210-5433},
journal={Philosophy \& Technology},
volume={27},
number={1},
title={{M}achine {M}orality: {T}he {M}achine as {M}oral {A}gent and {P}atient {[}{S}pecial issue{]}},
url={http://dx.doi.org/10.1007/s13347-014-0151-1},
publisher={Springer Netherlands},
keywords={Artificial intelligence; Ethics; Moral agency; Moral patiency; Roboethics},
editor={Gunkel, David J. and Bryson, Joanna},
language={English}
}
@article{machine-ethics-status,
year={2007},
issn={0924-6495},
journal={Minds and Machines},
volume={17},
number={1},
doi={10.1007/s11023-007-9053-7},
title={The status of machine ethics: a report from the {AAAI} {S}ymposium},
url={http://dx.doi.org/10.1007/s11023-007-9053-7},
publisher={Kluwer Academic Publishers},
keywords={Artificial intelligence; Machine ethics},
author={Anderson, Michael and Anderson, Susan Leigh},
pages={1-10},
language={English}
}
