fixed merge conflict

89436783 · Ross Duncan · cbfe0b7a · 255c57da · 89436783 · 89436783
Commit 89436783 authored 6 years ago by Ross Duncan
--- a/NEWPROPOSAL/FULLPROP.tex
+++ b/NEWPROPOSAL/FULLPROP.tex
@@ -163,7 +163,7 @@ Describe the specific objectives of the project, which should be clear, measurab
  This consists of techniques for a compiler to translate high-level specifications of quantum programs, to operations on a variety of hardware platforms, automatically managing resources and architectural constraints in doing so.
  This allows the software stack to be developed and organised in a modular fashion for multiple platforms, and then compiled in an intelligently managed way to make the most of quantum hardware resources.

-\TODOb{Summary/context should contain a clear statement that NOW zx does better (i.e. "outperforms") than anything else for circuit simplification using PyZX.  This should be explained in even more detail elsewhere.  This is the strongest case for getting the money!}
+\TODOb{Summary/context should contain a clear statement that NOW zx does better (i.e. "outperforms") than anything else for circuit simplification using PyZX.  This should be explained in even more detail elsewhere.  Aleks maybe?}
 \paragraph{Context:}
 \label{sec:context}  

@@ -186,7 +186,7 @@ techniques which may be used by a compiler to translate high-
    \vspace*{-10mm}%
 %    \texttt{\color{red!70!black} [Placeholder?]} \\[-5ex]
    \hspace*{-4mm}%
-    \cgraph[0.7375]{dzxc-arch-diagram.pdf}%
+    \cgraph[0.7375]{dzxc-arch-diagram3.pdf}%
    \vspace*{-9mm}  
    \end{wrapfigure}
 level quantum programs to low-level operations on quantum hardware platforms.
@@ -712,18 +712,6 @@ First, we will identify re-writing processes among the automated theorem proving
 This will allow us to identify candidate subroutines that require nonclassical resources to be carried out.
 Such subroutines then will be used to develop procedures for characterising if a \zx-represented algorithm demonstrates speed-up. 

-\KILL{\begin{color}{red!70!black}
-\texttt{[Delete the following once it has been properly incorporated into the WP description in \S3.2]}
-
-Deliverables: 
-\begin{itemize}
-  \item M12: Preliminary assessment of the comparative study of the axiomatizations of paradigms of quantum computation.
-  \item M15: \zx representation and explanation of the result that promotes magic states as a resource of quantum computation in the state injection paradigm.
-  \item M18: Preliminary assessment of nonclassicality of re-writing processes.
-  \item M24: A \zx formulation of contextuality (Kochen--Specker and/or generalised Spekken's type). 
-\end{itemize}
-\end{color}}
-
 \subsubsection{Machine-independent optimisation}
 \label{sec:repr-reas-azx}

@@ -890,10 +878,10 @@ Modelling error correction in \newt{the annotated \zxcalculus} will thus enable

 %doing cool stuff with cool stuff

-The retargetable \newt{\dzxc} system will make it easy to support new quantum
-devices, thus making the latest developments in quantum technology available to all academic and industrial users, and maximising the return on investment in quantum computing.
+\newt{The purpose of the \dzxc system is to form the basis of a retargetable compiler, making it easy to support new quantum devices.
+This will help to make the latest developments in quantum technology available to all academic and industrial users,  maximising the return on investment in quantum computing.}
 Our consortium includes an industrial partner (\newt{Cambridge Quantum Computing}) to help ensure the industrial relevance of our work.
-We also have experimental figures on the advisory panel.
+We also have \newt{experimentalists} on the advisory panel.
 With \newt{the \dzxc system}, high-level quantum languages and protocols can be designed without needing to know the underlying hardware.
 This will streamline the production of quantum software, opening it up to individuals and companies with limited prior knowledge of quantum computing. 
 Quantum hardware will also be more accessible, both in academia and industry. Individual developers will not need to know the entire architecture, as different elements can be \newt{adapted to automatically during compilation}.
@@ -919,15 +907,9 @@ This will accelerate the widespread commercial and academic development and expl
 %The inherently interdisciplinary nature of this project will bring a larger community to bear solving problems during the lifetime of the project. The development of the \azx language and compiler stacks themselves will also open up further challenges as the project progresses (and beyond) that are accessible to broader communities. 
 \newt{The \dzxc system} will connect the entire range of knowledge involved in building quantum technologies, from experimental and theoretical Physics, through to quantum computing theory, and on to formal methods of Computer Science.
 All of these are needed to develop the deep compilation system, \newt{so developing the \dzxc system} is a fundamentally interdisciplinary task.
-This \newt{system} opens the prospect of an acceleration in the development of quantum algorithms in a way which can then \newt{easily be adapted} to many different hardware platforms.
-For example, algorithm and protocol designers will not need to interface directly with quantum technologies: the \newt{\dzxc system} does all the compilation and optimisation necessary. This will allow the
-integration of quantum computing into mainstream Computer Science, and so the easy importing of tools (for example, techniques for optimisation or verification) that have been developed over many years. 
-By aiding the development of intuitively accessible programming
-languages, \newt{the \dzxc system} will also make quantum technologies
-accessible to a broader range of users and developers. End-users
-outside quantum physics and computer science will be able to build
-protocols for use in their own field that do not require them to
-understand the physical action of the hardware.  
+\newt{By aiding the development of intuitively accessible programming languages, the \dzxc system will also make quantum technologies accessible to a broader range of users and developers. 
+For example, algorithm and protocol designers will not need to interface directly with quantum technologies in order to test the effectiveness of their work, lowering the bar for development in the field.
+End-users outside of quantum physics and computer science will also be able to build protocols for use in their own field that do not require them to understand the physical action of the hardware.}

 The \dzxc system will not be a closed system: our commitment to open
 APIs (See \ref{task:trans1} and \ref{task:backendapi}) and our open
@@ -986,7 +968,8 @@ publication venues are conferences with published proceedings.)
  (QIP), \emph{Theory of Quantum Computation} (TQC), and \emph{Quantum
    Physics and Logic} (QPL). 
 \item Mainstream computer science venues: the \emph{Journal of the 
-    ACM} (JACM), \emph{Logic in Computer
+    ACM} (JACM),
+    the \emph{International Conference on Computer-Aided Design} (ICCAD), \emph{Logic in Computer
    Science} (LiCS), \emph{Principles of Programming Languages}
  (POPL), \emph{Automata, Logic and Programming} (ICALP), \emph{Tools
    and Algorithms for the Construction and Analysis of Systems}
@@ -1026,14 +1009,10 @@ post-doc will spend approximately 25\% of their time on site at the
 CQC offices in Cambridge.


-\TODOb{Niel/Dom please confirm NQIT role.}  \bR With NQIT, we will
-provide a programming framework for the networked quantum computer
-developed as part of that project. As part of this we will present the
-project results at the semi-annual NQIT Industry Forum events and the
-UK Quantum Technologies annual showcase.  To further this aim, our
-postdoc in Oxford will spend 10--20\% of their time working closely
-with the NQIT project.  In both cases, our work can be exploited
-directly by end-users. \e
+\newt{With NQIT, we will provide a programming framework for the networked quantum computer developed as part of that project (and the quantum computing project which follows afterwards in Phase~II of the UK Quantum Technologies Programme), seek to collaborate with their architectures team, and present the project results at the UK Quantum Technologies annual showcase.
+
+In both cases, our work can be exploited
+directly by end-users.} %% copied from Niel's document so I assume this is up-to-date

 \TODOb{This para to be cut.}
 \bR In addition, we have also recruited a board of advisers (see below)
@@ -1148,9 +1127,9 @@ early outputs of each WP will be used in later outputs of other WPs.
  
  \caption{Approximate timings and durations of tasks (months)}
 \label{fig:gantt}
-\end{figure}
+\end{figure}}

-The project is a single integrated whole, so there are many linkages
+\oldt{The project is a single integrated whole, so there are many linkages
 between the work packages; these are displayed in
 Figure~\ref{fig:pert}. As discussed in
 \S~\ref{sec:manag-struct-milest}, only some of these linkages are true
@@ -1158,6 +1137,7 @@ dependencies, where later tasks rely on results of earlier ones. On the other ha

 Our work plan consists of a balance of short tasks with concrete software deliverables (e.g. \ref{task:circuit-model}) and longer term, more ambitious and open-ended tasks (e.g. \ref{task:algorithms} and \ref{task:opt-machine}) which can offer significant, but less predictable, step-changes in the state of the art.

+
 The tasks of \ref{wp:frontend} are
 long and thin. That is, they are intended to work in parallel with the other WPs, with new features being integrated as they are developed.

@@ -1173,9 +1153,9 @@ The more challenging machine models of \ref{task:delft-model} and
 %  delayed until sufficient theory has been
 % developed to attempt them.

-\ref{wp:usefulstuff} requires integrating and generalising many of the
+\ref{wp:usefulstuff} requires integrating and generalising many of the 
 ideas of \ref{wp:backends} and \ref{wp:theory}, so it is mostly
-scheduled toward the end of the project.
+scheduled toward the end of the project.}

 \begin{figure}[h]
  \centering
@@ -1191,7 +1171,6 @@ However, because of the integrated nature of the project, and the high
 degree of past collaboration among the consortium members, most tasks
 receive attention from the personnel of several sites.  This degree of
 collaboration is a strong point of this project.
-}

 \newpage

@@ -1312,7 +1291,6 @@ Devise test-suite of concrete instances of circuits and
 \end{WPdeliverables}
 \end{WP}

-
 \newpage

 %%%
@@ -1320,36 +1298,36 @@ Devise test-suite of concrete instances of circuits and
 %%%
 \begin{WP}{Representation, reasoning, and resources in \zx}{1M}{36M}{wp:backends}
 \WPleaderPOL
-\WPeffort{\newt{12}}{\newt{12}}{\newt{14}}{\newt{4}}{\newt{30}}{\newt{6}}
+\WPeffort{{12}}{{12}}{{14}}{{4}}{{42}}{{6}}
 \begin{WPaim}
 %We build the theoretical foundations for \zx as an intermediate representation. This includes extending the capabilities of \zx to represent mixed states, qudit states, and control flows. We use \zx axiomatisations and automated theorem provers to extract out post-classical computing resources, which will be used both for further optimisation work, and for characterisation of quantum algorithmic speed-up.
 We build the theoretical foundations for \zx as an intermediate representation. This includes extending the capabilities of \zx to represent qudit states with a fixed $d$, arbitrary finite-dimensional quantum states, and control flows. We explore the structure of W-type tensors with interaction with \zx generators of GHZ-type. We use \zx axiomatisations and automated theorem provers to extract out post-classical computing resources, which will be used both for further optimisation work, and for characterisation of quantum algorithmic speed-up.
 \end{WPaim}
 \begin{WPtasks} 
    \WPtask[\label{task:axioms}]{Beyond qubits and stabilisers 
-   \newt{   (M1--M14; Responsible: \partnerref{partner:oxford}; Involved: \partnerref{partner:loria},\partnerref{partner:gdansk})}}{%
+   {   (M1--M14; Responsible: \partnerref{partner:oxford}; Involved: \partnerref{partner:loria},\partnerref{partner:gdansk})}}{%
    %We will exploit further the recent completeness results to give representations for mixed state qubit quantum theory. We will
    %extend the \textsc{zx} tensor formalism from the qubit domain to higher dimensions.
    We will extend the completeness results of the \textsc{zx}-calculus from the qubit domain to higher dimensions, to have complete qudit \textsc{zx}-calculus. Furthermore, we will combine all the qudit \textsc{zx}-calculus into a single framework so that we can deal with the whole finite-dimensional quantum theory in a \textsc{zx} style. In addition, we will exploit techniques from the \textsc{zw}-calculus to understand the deep structure of  W-type tensors.        
    % and exploit the translation from \textsc{zx}-    to \textsc{zw}-calculus.
    }
  \WPtask[\label{task:betterboxes}]{Control in \zx
-    \newt{  \ (M1--M18; Responsible: \partnerref{partner:gdansk}; Involved: \partnerref{partner:grenoble},\partnerref{partner:loria},\partnerref{partner:oxford})}}{%
+    {  \ (M1--M18; Responsible: \partnerref{partner:gdansk}; Involved: \partnerref{partner:grenoble},\partnerref{partner:loria},\partnerref{partner:oxford})}}{%
   % Support simple control flow at the level of \azx, making it a more suitable target for compiling from a high-level language. In particular, add support for repetition and recursive definitions of diagrams, e.g. for expressing and transforming regular families of circuits. 
   We will use parametric \zx terms to support simple control flow at the level of the \dzxc system, making it a more suitable target for compiling from a high-level language. In particular, we will add support for repetition and recursive definitions of diagrams, e.g. for expressing and transforming regular families of circuits. 
  }
-      \WPtask[\label{task:resources}]{Resources and axioms
-   \newt{ (M1--M36; Responsible: \partnerref{partner:gdansk}; Involved: \partnerref{partner:grenoble},\partnerref{partner:loria},\partnerref{partner:oxford})}}{%
+      \WPtask[\label{task:resources}]{Resources and axioms 
+   { (M1--M36; Responsible: \partnerref{partner:gdansk}; Involved: \partnerref{partner:grenoble},\partnerref{partner:loria},\partnerref{partner:oxford})}}{%
 We will exploit the three axiom sets for Clifford, Clifford+T, and universal qubit QM,
 to identify and distill specific resources that are necessary to quantum speed-up. In particular, to focus on finding multiple resource elements (rather than simply magic states), and to characterise post-classical composition as a resource.
 This includes developing \zx representations of contextuality, as a possible post-classical resource.
    }
          \WPtask[\label{task:resourcesagain}]{Computational resources
-  \newt{  (M12--M36; Responsible: \partnerref{partner:gdansk}; Involved: \partnerref{partner:grenoble},\partnerref{partner:oxford},\partnerref{partner:CQC},\partnerref{partner:radboud})}}{%
+  {  (M12--M36; Responsible: \partnerref{partner:gdansk}; Involved: \partnerref{partner:grenoble},\partnerref{partner:oxford},\partnerref{partner:CQC},\partnerref{partner:radboud})}}{%
 We will use the existing graph re-writing and automated theorem proving tools of Quantomatic and PyZX to determine parts of the re-writing process that are difficult to compute classically. This will then be used to extract candidate subroutines for sources of quantum speed-up. Along with the previous task, these will be used to develop procedures for characterising if a \zx-represented algorithm demonstrates speed-up or not.
    }
 
-\end{WPtasks}
+\end{WPtasks} 
 \begin{WPdeliverables}
  \WPdeliverable{M9}{Preliminary assessment of the comparative study of the axiomatizations of paradigms of quantum computation} 
   \WPdeliverable{M14}{Completeness  of qudit \zx calculus} 
@@ -1369,7 +1347,7 @@ We will use the existing graph re-writing and automated theorem proving tools of
 %%%
 \begin{WP}{Machine-independent optimisation}{M1}{M36}{wp:theory}
  \WPleaderOXF
-  \WPeffort{\newt{12}}{\newt{9}}{\newt{30}}{\newt{12}}{\newt{12}}{\newt{12}}
+  \WPeffort{{12}}{{9}}{{30}}{{12}}{{12}}{{12}}
  \begin{WPaim}
 We develop practical logical and algorithmic techniques for transforming  ``abstract'' \zx terms produced from a high-level program in ways
    which will be required by any practical compiler, and reasoning about their properties. Examples include:
@@ -1379,7 +1357,7 @@ We develop practical logical and algorithmic techniques for transforming  ``abst
  \end{WPaim}
  \begin{WPtasks}
    \WPtask[\label{task:algorithms}]{Reduction strategies, algorithms,
-      and complexity (M1--M24; \newt{Responsible: \partnerref{partner:radboud}; Involved: \partnerref{partner:loria}, \partnerref{partner:oxford}, \partnerref{partner:CQC}})}{%
+      and complexity (M1--M24; {Responsible: \partnerref{partner:radboud}; Involved: \partnerref{partner:loria}, \partnerref{partner:oxford}, \partnerref{partner:CQC}})}{%
      Develop new strategies for simplifying \textsc{zx}-style tensor
      networks and reducing to (pseudo) normal forms, with the help of
      automated techniques such as Knuth-Bendix completion and
@@ -1388,14 +1366,14 @@ We develop practical logical and algorithmic techniques for transforming  ``abst
    }

    \WPtask[\label{task:annotate1}]{Topological and causal constraints 
-      (M1--M18; \newt{Responsible: \partnerref{partner:oxford}; Involved: \partnerref{partner:loria}, \partnerref{partner:CQC}, \partnerref{partner:radboud}})}{%
+      (M1--M18; {Responsible: \partnerref{partner:oxford}; Involved: \partnerref{partner:loria}, \partnerref{partner:CQC}, \partnerref{partner:radboud}})}{%
      Extend \dzxc language and tools to express and enforce: (1) topological
      constaints, such as nearest-neighbour connectivity of qubits and
      (2) causal/temporal constraints, such as sequential ordering of
      measurements and classically-controlled operations.
    }
    \WPtask[\label{task:annotate2}]{Quantitative Properties (M13--M24; 
-      \newt{Responsible: \partnerref{partner:CQC}; Involved: \partnerref{partner:oxford}, \partnerref{partner:CQC}, \partnerref{partner:radboud}})}{%
+      {Responsible: \partnerref{partner:CQC}; Involved: \partnerref{partner:oxford}, \partnerref{partner:CQC}, \partnerref{partner:radboud}})}{%
      Extend \dzxc language and tools to account for several kinds of
      numerical annotations, e.g.~timing data related to performing
      operations, gate fidelities, channel fidelities, and decoherence
@@ -1405,7 +1383,7 @@ We develop practical logical and algorithmic techniques for transforming  ``abst
      quantities from local to global properties.  
    }
        \WPtask[\label{task:basic-opt}]{Generic optimisations of ZX-terms 
-      (M12--M24; \newt{Responsible: \partnerref{partner:oxford};   Involved: \partnerref{partner:loria}, \partnerref{partner:CQC}, \partnerref{partner:radboud}})}{% 
+      (M12--M24; {Responsible: \partnerref{partner:oxford};   Involved: \partnerref{partner:loria}, \partnerref{partner:CQC}, \partnerref{partner:radboud}})}{% 
      Use the results of task~\ref{task:algorithms} to develop
      procedures to optimise \zx-terms, in a way which is applicable
      for families of circuits (e.g.~Clifford, Clifford+T, CNOT+T,
@@ -1417,7 +1395,7 @@ We develop practical logical and algorithmic techniques for transforming  ``abst
      }
 %%
      \WPtask[\label{task:ECC}]{Application of Error-Correction
-        (M1--M24; \newt{Responsible: \partnerref{partner:oxford}; Involved: \partnerref{partner:grenoble}, \partnerref{partner:gdansk}})}{%
+        (M1--M24; {Responsible: \partnerref{partner:oxford}; Involved: \partnerref{partner:grenoble}, \partnerref{partner:gdansk}})}{%
        Develop algorithms which rewrite abstract tensor networks to
        equivalent tensors in codeword space of a chosen
        error-correcting code.  This may be combined with additional
@@ -1456,12 +1434,12 @@ Perdrix, Valiron, Carette.}
 %%%
 %%%%%%%WP 4
 %%%
-\begin{WP}{Machine-dependent optimisation}{1M}{36M}{wp:usefulstuff}
+\begin{WP}{Machine-dependent optimisation}{13M}{36M}{wp:usefulstuff}
  \WPleaderGREN
-  \WPeffort{12}{9}{12}{6}{\newt{6}}{0}
+  \WPeffort{12}{9}{12}{6}{6}{0}
  \begin{WPaim}
 We import machine-dependent specifications to \zx terms, and use this to optimise algorithms further for specific hardware constraints. We focus on the silicon spin qubits developing in Grenoble, the ion traps developed in Oxford, and the superconducting devices accessible through CQC and partnership with IBM. This is the culmination of all previous work packages, and feeds back into them. The final result will be \ldots.
-Also machine-dependent error correction here?
+Also machine-dependent error correction?
  \end{WPaim}
  \begin{WPtasks}
 \WPtask[\label{task:qdot-model}]{Grenoble silicon spin qubits (M13--M36  Responsible: \partnerref{partner:grenoble};
@@ -1479,6 +1457,7 @@ Also machine-dependent error correction here?
  In collaboration with the Oxford ion trap group and the NQIT team, we will design an output module which generates code for a realistic model of
  ion trap quantum computers, including qubit losses and leakage, gate
  timings, and circuit layout. Output language to be defined in collaboration with hardware experts at Oxford.}
+  
  \WPtask[\label{task:runnable}]{Formatting for target systems 
    (M15--M30; Responsible: \partnerref{partner:loria}; Involved: \partnerref{partner:grenoble},\partnerref{partner:oxford},\partnerref{partner:gdansk})}{%
    Develop algorithms which, given a collection of constraints
@@ -1504,9 +1483,10 @@ Also machine-dependent error correction here?
  \WPdeliverable{M30}{General purpose layout engine}
  \WPdeliverable{M36}{Optimising \dzxc, suitable for compiling to Grenoble and/or Oxford architecture}    
  \WPdeliverable[\label{del:backendapi}]{M36}{API for back-end
-    modules, including specification language for architectures.}    
+    modules, including specification language for architectures.}
  \end{WPdeliverables}
 \end{WP}
+\TODOb{Should D4.5 be the same as D1.6?}

 \REM{\emph{Leader:} Kissinger.
 \emph{Others:} Abramsky, de Beaudrap, Duncan, Jeandel, Perdrix, 
@@ -1518,7 +1498,7 @@ Staton, Carette.}

 \begin{WP}{Administation and Communications}{M1}{M36}{wp:admin} 
  \WPleaderGREN
-  \WPeffort{3}{3}{2}{1}{\newt{4}}{0}
+  \WPeffort{3}{3}{2}{1}{4}{0}
  \begin{WPaim}
    This work package collects general administrative activities and
    the organisation of the project meetings.  All meetings will be
@@ -1573,13 +1553,13 @@ Staton, Carette.}
    & \ref{wp:usefulstuff} 
    & \ref{wp:admin} 
    & \textbf{TOTAL} \\\hline
-1. Grenoble   &  12 & 12 & 12 & 12 & 3  & 51 \\\hline
+1. Grenoble   &  12 & 2 & 12 & 12 & 3  & 41 \\\hline
 2. LORIA      & 20 & 12 & 9  & 9  & 3  & 53 \\\hline
 3. Oxford     & 32 & 14 & 30 & 12 & 2  & 90 \\\hline
 4. CQC        &  12 & 4  & 12 &  6 & 1  & 28 \\\hline
-5. Gdansk     &  12& 30 & 12 & 6 & 4  & 71 \\\hline
+5. Gdansk     &  12& 42 & 12 & 6 & 4  & 76 \\\hline
 6. Nijmegen   &  3 & 6  & 12 & 0 & 2  & 23 \\\hline
-\textbf{TOTAL}& 91 & 78 & 87 & 45 & 11 & 317   \\\hline 
+\textbf{TOTAL}& 91 & 80 & 87 & 45 & 11 & 317   \\\hline 
  \end{tabular}
 \end{center}}

@@ -1622,9 +1602,9 @@ Each work package will be lead by a responsible PI who will coordinate
 research activity between sites to ensure that deliverables are met,
 achieve WP-specific objectives, and organise collaboration meetings as
 needed.  
-\textbf{\ref{wp:frontend}}: \newt{S. Perdrix (LORIA)}, 
+\textbf{\ref{wp:frontend}}: S. Perdrix (LORIA), 
 \textbf{\ref{wp:backends}}: A. B. Sainz (Gdansk),
-\textbf{\ref{wp:theory}}: \newt{B. Coecke (Oxford)},
+\textbf{\ref{wp:theory}}: B. Coecke (Oxford),
 \textbf{\ref{wp:usefulstuff}}: D. Horsman (Grenoble),
 \textbf{\ref{wp:admin}}: D. Horsman (Grenoble).

@@ -1666,19 +1646,24 @@ reduce risk and improve communication across the consortium.
    \textbf{Title} \\\hline
 %
 \ms \label{ms:qasmqasm}&
-9 &
+12 &
+\ref{wp:frontend},\ref{wp:backends},  &
+Deep-\zx based circuit optimiser software that outperforms all other procedures.\\\hline
+%
+\ms \label{ms:qasmqasm}&
+14 &
 \ref{wp:frontend},\ref{wp:backends},  &
-Minimal QASM$\rightarrow$QASM circuit optimiser\\\hline
+Compiler module for qudit and mixed state \\\hline
 %
 \ms \label{ms:quippermbqc}&
 18&
 All&
-Quipper$\to$MBQC compilation pipeline\\\hline
+\zx $\to$MBQC and circuit compilation pipeline\\\hline
 %
 \ms \label{ms:simbackend}&
 24&
 All&
-Simulator back-end with parametric \azx support\\\hline
+ \\\hline
 %
 \ms \label{ms:delftbackend}&
 30&
@@ -1688,7 +1673,7 @@ All&
 \ms \label{ms:nqitbackend}&
 36&
 All&
-Integrate NQIT back end\\\hline
+Complete deep-\zx compiler stack with open APIs. \\\hline
  \end{tabular}}
 \end{center}

@@ -2326,7 +2311,7 @@ industrial partner, a Senior Scientist and a Research Software Developer will co
  At Grenoble, Richard East will contribute
  approx 9 months to the project; this is funded from other sources.
  At Gdansk, a PhD student will contribute
-  approx 12 months to the project; this is funded from other sources. \newt{At Oxford, at least tow PhD students will contribute to the project, as well as several MSc students}.
+  approx 12 months to the project; this is funded from other sources. \newt{At Oxford, at least two PhD students will contribute to the project, as well as several MSc students}.
 \item \emph{Engineers}: At CQC, a Senior Scientist and a Research Software Developer will contribute, funded from other sources.
 \end{itemize}


--- a/NEWPROPOSAL/images/dzxc-arch-diagram.key
+++ b/NEWPROPOSAL/images/dzxc-arch-diagram.key
--- a/NEWPROPOSAL/images/dzxc-arch-diagram2.pdf
+++ b/NEWPROPOSAL/images/dzxc-arch-diagram2.pdf
--- a/NEWPROPOSAL/images/dzxc-arch-diagram3.pdf
+++ b/NEWPROPOSAL/images/dzxc-arch-diagram3.pdf