deletions | additions       diff --git a/12discussion.tex b/12discussion.tex  index 3b49347..7c42306 100644  --- a/12discussion.tex  +++ b/12discussion.tex 
...
\subsection{Challenges for future research} \label{sec:recommendations}  Viewing Reviewing  the concepts components of serendipity introduced  in Section \ref{sec:by-example} through and crystalised in the definition of serendipity  presented in Section \ref{sec:our-model} in light of  the practice scenarios we have discussed, discussed in Section \ref{sec:computational-serendipity},  we can describe the following challenges for research in computational serendipity. \begin{itemize}  \item \textbf{Autonomy}: The essential issues were drawn out in Section \ref{sec:related}, and are expanded here.  \paragraph{\textbf{Autonomy}.}  Our case studies in Section \ref{sec:computational-serendipity} highlight the potential value of  increased autonomy on the system side.  The search for connections that make raw data into ``strategic data'' 
...
shall be allowed to stipulate his own purpose.'' \emph{A  primary challenge for the serendipitous operation of computers is  developing computational agents that specify their own problems.}  \end{itemize}  \begin{itemize}  \item \textbf{Learning}: \paragraph{\textbf{Learning}.}  Each of the case studies considered in Section \ref{sec:computational-serendipity} describes a system that  is able, in one way or another, to learn from experience. As we  considered ways to enhance the measure of serendipity in these  examples, we were led to consider computational agents that  participate more meaningfully in ``our world'' rather than in a  circumscribed microdomain. Knowledge-intensive development work may  often be unavoidable, but understanding unavoidable. Understanding  how to foster serendipity is a  particularly  important step,  because it points to the potential of systems learning on their own. \emph{A second challenge is for computational agents to learn more and more about the world we live in.} \end{itemize}  \begin{itemize}  \item \textbf{Sociality}: \paragraph{\textbf{Sociality}.}  We may be aided in our pursuit of the ``smart mind'' \cite{campbell2005serendipity}  required for serendipity by recalling Turing's proposal that computers should ``be able to converse with each other to sharpen their wits'' \cite{turing-intelligent}.Turing  recognised that computers would have to be coached in the direction  of social learning, but that once they attain that standard they  will learn much more quickly.  The four supportive factors for serendipity described in this paper -- a  \emph{dynamic world}, \emph{multiple contexts},  \emph{multiple tasks}, and \emph{multiple  influences} --  resemble nothing more thanour  experience of  social reality. For example, In  our analysis of {\sf GAmprovising} we  suggested that it a future version of the system  should interact more with the listener listener,  and that individual Improvisers should allow themselves to be influenced by each other, rather than working in a digital silo. \emph{A third challenge is for computational agents to interact in a recognisably social way with us and with each other, resulting in emergent effects.} \end{itemize}  \begin{itemize}  \item \textbf{Embedded evaluation}: \paragraph{\textbf{Embedded evaluation}.}  \citeA{stakeholder-groups-bookchapter} outline a general programme for computational creativity, and examine perceptions of  computational creativity among members of the general public,  computational creativity researchers, and existing creative  communities. We should now add a fourth important ``stakeholder''  group in computational creativity research: computer systems  themselves. System designers need to teach their systems how to  make evaluations in a way evaluations. We saw  that this  is both reasonable and ethical. This  condition a crucial issue in designing  an automated programming experiment. The guiding light  isexemplified by the preference for  a ``non-zero sum'' criterion for value introduced conception of value. This can be extended to  situations  in Section \ref{sec:by-example}.  Similar judgements apply when which the  the ``product'' is a new process. process or action.  Within a Kantian framework ``an agent's moral maxims are instances  of universally-quantified propositions which could serve as moral  laws -- ones holding for any agent'' \cite{powers2005deontological}. Embedded evaluation is of immediately has  pragmatic as well asbroader  philosophical implacations; thus, for example, the latest implementation of {\sf GAmprovising} is limited because it is ``poor at using reasoned self-evaluation'' and ``does not generate novel aesthetic measures'' \cite[pp.~189, 288]{jordanous2012evaluating}. \emph{A fourth challenge is for computational agents to evaluate their own creative process and products.}\end{itemize}  \subsection{Future Work} \label{sec:futurework} \label{sec:hatching} 
...
\vspace{-1mm}  \end{mdframed}  }  \caption{Standard \caption{A standard  design pattern template applied to van Andel's serendipity pattern,  \em{Successful error}\label{fig:va-pattern-figure}} \end{figure}  \noindent There are many ways to describe a solution. Meszaros and Doble remark, 
...
To van Andel's assertion that ``The very moment I can plan or  programme `serendipity' it cannot be called serendipity anymore,'' we  reply thatwe can certainly describe  patterns -- and programs -- with can include  built-in indeterminacy. Moreover, we can foster circumstances that may make unexpected happy outcomes more likely, by designing and  developing systems that increasingly address the challenges outlined in Section \ref{sec:recommendations}. Such systems would have a chance of encountering will encounter  unexpected stimuli, becoming become  curious about them, sagaciously pursuing enquiry together with others, within a social context,  and assessing assess  the value of any outcomes. %  Figure \ref{fig:va-pattern-figure} shows one approach example of a design  pattern that can be used  to planning \emph{plan  for serendipity, serendipity},  based on rewriting one of van Andel's  serendipity patterns using the standard design a  pattern template. identified by van Andel: ``\emph{Successful Error}''.  In future work, we intend to build a more complete serendipity pattern  language -- and put it to work within autonomous programming systems.  diff --git a/13conclusion.tex b/13conclusion.tex  index c66d6f4..ada97b9 100644  --- a/13conclusion.tex  +++ b/13conclusion.tex 
...
\section{Conclusion} \label{sec:conclusion}  %  We began the paper  by surveying ``serendipity'', developing a broad the  historical view, meaning of  ``serendipity'', drawing on previous theories  and describing several collecting  historical examples from various domains. We used this survey to  develop a collection of  criteria which we propose to be computationally salient.We reviewed related work; like  \citeA{andre2009discovery}, we propose a two-part definition of  serendipity: \emph{discovery} followed by \emph{invention}.  %  Adapting the ``Standardised Procedure for Evaluating Creative  Systems'' (SPECS) model from \citeA{jordanous:12}, we developed proposed  a detailed set of evaluation standards for serendipity. %  We used this model to analyse the potential for serendipity in case  studies of evolutionary computing, recommender systems, and automated  programming. We saw that the proposed framework can be used  retrospectively, as an evaluation tool, and prospectively, as a design  tool.  %  We used this model to analyse potential for serendipity with case studies in evolutionary computing, recommender systems, and automated programming. then reviewed related work: like  \citeA{andre2009discovery}, we propose a two-part definition of  serendipity: \emph{discovery} followed by \emph{invention}.  %  We then reflected back over our definition and analyses, and outlined a programme for serendipitous computing in the pursuit of \emph{autonomy}, \emph{learning}, \emph{sociality}, and \emph{embedded evaluation} that would tackle the following challenges: %  \begin{itemize}  \item \emph{A primary challenge for the serendipitous operation of 
...
%   We indicate several possible further directions for implementation  work in each of our case studies. We have also drawn attention to  broader design considerations. considerations in system design.  Our examples show that serendipity is not foreign to computing practice. There are further gains to be had for research in computing by planning -- and programming -- for serendipity. %  diff --git a/2literature.tex b/2literature.tex  index 50e55f5..92dd42f 100644  --- a/2literature.tex  +++ b/2literature.tex 
...
latter scenario is provided by the Swiss company Freitag, which was  started by design students who built a business around ``upcycling''  used truck tarpaulins into bags and backpacks. Thanks in part to  clever marketing \cite[pp. 54--55, 68--69,]{russo2010companies}, 68--69]{russo2010companies},  their product is now a global brand. Wherever possible, we prefer  to make use of independent judgements of value, which helps to  capture a non-zero sum notion of value. 
...
\begin{itemize}  \item \textbf{Multiple tasks}: Einstein's work at the patent office  seems to have been fortuitous not because it gave him ideas, but  because it gave him time to work on his ideas, famously resulting ideas. Famously, this  resulted  in four fundamental papers in 1905. Two decades later, translating his correspondent Satyendra Nath Bose's paper from English to German, Einstein learned a calculation method that produced accurate physical results, despite results -- and that  implicitly making made  nonstandard physical assumptions \cite{delbruck1980bose}. Subsequentfurther  examination of these ideas led to fundamentally new insights into physics. the discovery of  Bose-Einstein statistics, which describes particles that do not obey  the Pauli exclusion principle.  The potential for interrelationships between the tasks -- and is especially relevant, along with  the ability to carry the tasks through-- seems  to be more important  than the raw number of tasks. a satisfactory degree.  In John Barth's \citeyearpar{barth1992last} novel \emph{The Last Voyage of Somebody  the Sailor}, the author gives himself three main tasks: writing  what reads as a straightforward semi-autobiographical fiction,  writing an historical fantasy, novel,  adapting a classic fairy tale,  and interweaving (and finally merging) these two stories. The result includes numerous examples  of what the text refers to as ``logistically assisted serendipity''  \cite[p.~311]{barth1992last}, through repeated or varied the use of  images and plot points. points that repeat-with-variation.  \end{itemize}  \begin{itemize}  \item \textbf{Multiple influences}: The bridge from trigger to result  is often found by making use of a social network. For example, Arno  Penzias and Robert Wilson, working at Bell Labs, used a large  antenna to detect radio waves that were relayed by bouncing off satellites. After they had removed interference effects due to  radar, radio, and heat, they found residual ambient noise that  couldn't be eliminated. They were mystified, and only understood  diff --git a/3model.tex b/3model.tex  index b46311c..30eaea4 100644  --- a/3model.tex  +++ b/3model.tex 
...
\section{A computational model of serendipity} \label{sec:our-model}  Figure \ref{fig:model} recapitulates the ideas from the previous  section, integrating them into a computationally meaningful model of  process. process   model.  The model is summarised in text form at the end of this section, in our working definition of serendipity.  Figure \ref{fig:1a} is a heuristic map of the features of serendipity 
...
Figure \ref{fig:1b} removes these unserendipitous paths to focus on  the key features of ``successful'' serendipity.  %  The \textbf{serendipity trigger} \textbf{trigger}  is denoted here by $T$. %  The \textbf{prepared mind} corresponds to those preparations, labeled  $p$ and $p^{\prime}$, that are relevant to the discovery and invention 
...
from there, to applications.  The ellipses at the end of the workflow in Figure \ref{fig:1c} are  intended to suggest that applications are open-ended; however, an one  important class of applications will result in changes to one or more  of the system's modules, for example either  by expanding the knowledge base that they have available. it draws on, or adjusting its processing.  Note that earlier components of the workflow cannot, in general, anticipate what the subsequent phases will produce  or achieve. If the system's next steps could be anticipated, we would  not say that the behaviour was serendipitous. In other words,  serendipity serendipitous, and this global condition pushes back on each of the component modules.   Serendipity  does not adhere to one specific part of the system, but to its operations as a whole. Although Figures \ref{fig:1b} and \ref{fig:1c}  treat the case of successful serendipity, as indicated in Figure  \ref{fig:1a}, each step is fallible, as is the system as a whole.  Thus, for example, a trigger that has been initially tagged as interesting may prove to be fruitless.  Similarly a system that implements all of the steps in Figure \ref{fig:1c}, but that for whatever reason  never achieves results of value does not cannot be said to  have potential for serendipity. However, a system only produces results of high value would also be  suspect, since it would indicate a tight coupling between trigger  and outcome. Fallibility is a ``meta-criterion'' that transcends the criteria from Section \ref{sec:by-example}. 
...
\begin{mdframed}  \begin{ndef}  \emph{(1) Within a system with a prepared mind, a previously uninteresting trigger arises due to circumstances that the system does not control and cannot predict, control,  and is classified as interesting by the system; and,} \emph{(2) The system uses the trigger and prior preparation, together with relevant computational processing, networking, and experimental techniques, to obtain a novel result that is evaluated favourably by the system or by external sources.}  \end{ndef}  \end{mdframed}  diff --git a/6SPECS.tex b/6SPECS.tex  index a51bda6..c7cfa0c 100644  --- a/6SPECS.tex  +++ b/6SPECS.tex 
...
\subsection{Heuristics}\label{specs-heuristics}  Do How can we  we estimate the chance of the trigger appearing according to the  trigger's uniqueness, or some other feature? appearing, if every  trigger is unique?  Consider de Mestral's encounter with burrs. The probability of encountering burrs while out walking is high, and high:  many people have had a similar experience before  and since. that experience.  The unique features of de Mestral's experience are that he had the curiosity to investigate the burrs under a microscope, and the sagacity (and tenacity) to turn what he discovered into a successful product. The details of the particular burrs that were encountered effectively irrelevant. In the genera case, we are not interested in the chance of encountering a particular object or set of data, which may be vanishingly small. Rather, we are interested the chance of encountering a trigger that could precipitate an interested response.  The trigger itself may be a complex object or event that takes place over a period of time; in other words, it may be a pattern, rather than a fact. Noticing that a pattern exists is a key aspect of sagacity. diff --git a/8cc.tex b/8cc.tex  index 4b7e7c0..fab18c0 100644  --- a/8cc.tex  +++ b/8cc.tex 
...
However, in the version of the system under discussion, individual  threads are effectively equivalent regarding the features of chance,  curiosity, and sagacity. The only thing to distinguish them from one  another is their value. Accordingly, Referring to the moderate-at-best likelihood  measure,  even those threads which maximise value cannot be regarded as particularly serendipitous. \subsubsection{Qualitative assessment}  diff --git a/bibliography/biblio.bib b/bibliography/biblio.bib  index 3e2f124..2644d10 100644  --- a/bibliography/biblio.bib  +++ b/bibliography/biblio.bib 
...
year={1980},  publisher={ACS Publications}  }  @inproceedings{jordanous2015four,  title={Four {PPPP}erspectives on {C}omputational {C}reativity},  author={Jordanous, Anna},  booktitle={Procs. of the AISB Symposium on Computational Creativity},  year={2015}  }