MathUI
Live Mathematics
Authorea
Showcase
Flot
D3JS
iPython
LaTeXML
Pandoc
http://creativecommons.org/licenses/by/4.0/
A case for transparency in science, exemplified via a showcase of active mathematical features on Authorea.com
2015-07-13
class: center, middle # Live Mathematics on Authorea ## MathUI 2015 .footnote[**Deyan Ginev**, Alberto Pepe, Nate Jenkins] <div style="position:absolute; top: 12px; right: 20px;"> <div style="position:relative;"> <a rel="license" href="http://creativecommons.org/licenses/by/4.0/" title="This work is licensed under a Creative Commons Attribution 4.0 International License"> <img alt="Creative Commons License" style="border-width:0" width="80" src="http://mirrors.creativecommons.org/presskit/buttons/80x15/png/by.png" /> </a> </div> </div> --- # Agenda 1. Intro to Authorea 2. A Case for Transparency in Science 3. Live Mathematics 4. The Authorea Publication Life-cycle 5. A Wishlist for the Future --- # Authorea * "Collaborative typewriter for academia" * Science publishing startup, based in New York * Seed funding round closed on 09/2014 * "Disruptive" vision --- ## Authorea — Vision Statement > We empower scholars with the leading platform to **read**, **write**, and **discuss** research findings. > > Authorea is collaborative, rapid, efficient, open and transparent. > > The way scientific research was always meant to be. --- ## Problems * Outdated tools * collaboration is difficult * Heterogeneous data * disconnect between paper and experiments * Hard-to-verify results * Distributed, international, large volume experiments --- ## Authorea Goals * Accelerate the academic process * Improve the quality of academic submissions --- ## Authorea Features — Authoring * Collaborative web editing * CERN paper with over 200 authors * MarkDown, LaTeX and RichText as authoring languages * Write once, export to 70+ styles with 1 click * Version control — track changes, undo, smart archiving * Integrated in-article citation and reference management --- ## Authorea Features — Ecosystem * Get notified of edits, comments and new articles by co-authors and colleagues * Easy import of existing articles * Fork and remix any public article * Free for Open Science * GitHub integration for offline editing --- ## Authorea Features — Delights * Academic word count * Institution pages * Manage affiliations and connections * Article group chat --- class: center # Why Open Science? Let's take a closer look... --- ## Evolution of Co-authorship <img id="number-of-authors" width="80%" src="/assets/number-of-authors-0a71df9af402a171a8a6990218ced3c09e46eca16e75e0c3e9361e4a1fdc1ccd.png" /> Number of authors per paper by decade in the field of medicine.<br> Source: (Wren 2007) --- ## Retractions and Impact <span> <img id="retractions" width="60%" src="/assets/retractions-F1-d3113bea13a08ebd9ae2c07b702a4de8cbccd19e22ee063b077ebefceedcc53e.jpg" /> </span> <span style="font-style: italic;">Correlation between impact factor and retraction index. The 2010 journal impact factor (37) is plotted against the retraction index as a measure of the frequency of retracted articles from 2001 to 2010."</span> Source: (Fang 2011) --- ## Retractions in Mathematics * Retraction Watch initiative * http://retractionwatch.com/category/by-subject/physical-sciences-retractions/math-retractions/ * Math retractions are largely plagiarism-based --- ## A Case for Transparency * **Reproducibility** * Full verification requires a full reproduction of the experiment * **Reusability** * "Standing on the shoulders of giants" * Can't remix a black box * **Accessibility** * Language, dissemination, technology, licensing * **Availability** * Paywalls * Archival - Standard, Open Science components --- # Live Mathematics > Interactivity enhances learning and reviewing. Openness ensures the entire ecosystem benefits. * Textual mathematics (Pandoc, LaTeXML) * Interactive plots (Flot, Bokeh) * Vizualizing data (D3JS) * Embedded Computation (iPython/Jupyter) --- ## Textual Mathematics * LaTeX - Lingua franca for writing math equations * Embeddable in MarkDown (via Pandoc) * Embeddable in LaTeX (via Pandoc and LaTeXML) * Served to the user via MathJaX: * Renders in all browsers * Fills in equation numbers * Allows for labels and references \begin{equation} {\varphi}(x, \mu, \sigma) = \frac{1}{\sqrt{2\pi\sigma^2} } e^{ -\frac{(x-\mu)^2}{2\sigma^2} } \end{equation} * Native MathML remains a dream --- ### A Note on LaTeX Backends <a id="refresh-latexml-pandoc" href="javascript:void()"> <img id="latexml-pandoc" width="100%" src="/assets/latexml-pandoc-af5827a895d2fcbc7f998c31642cc53a14c52a064a9eea74a0f26e80d03f3087.gif" /> </a> --- ## Interactive Plots * Static print-oriented plots struggle in high-dimensional spaces * Parameter fine-tuning can introduce accidental bias * Interactive plots allow for exploring the full mathematical object * NIST's DLMF uses WebGL for Complex Function Surfaces * Authorea has an example integration with Flot and Bokeh --- ## Interactive Plots — Example <center><iframe style="border: none; margin-top:-15px;" id="flot-iframe" src="/assets/flot_normal_distro.html" width="700" height='450'></iframe></center> --- ## Vizualizing Data * Going beyond plots, interact with an entire dataset * Use cases: Sampling, Evaluating models, Advanced Data Visualization <center><iframe style="border: none;" id="d3js-iframe" src="/assets/normal_distribution_d3js.html" width="500" height='500'></iframe></center> --- ## Visualizing Data — Goal <img id="active" width="80%" src="/assets/visualizing-data-active-cf70ba350e2b930d62151874c5fc620e7a64bb961fbcbd3938095fdb053e4908.gif" /> --- ## Embedded Computation * Go beyond pseudocode: "Live algorithms" * Allows for open, as well as user/reviewer-contributed tests and benchmarks * Crucial: standardized computational frameworks (e.g. iPython/Jupyter) * [Authorea Example](https://www.authorea.com/41811#article-paragraph-figures__slash__pdf__slash__pdf__dot__png) --- # Authorea Publication Life-cycle * Author a paper * Obtain an academic identifier * Publish and share * Archive and Remix --- ## Authoring Life-cycle * Mimicks existing publications * Rought draft * Working draft * Preprint * Peer-reviewed * Postprint * Versioned tags * Living papers? If desired. --- ## Identifiers and Notability * Largely "Future Work" * Papers are intended as self-contained objects * "Notable" and/or "sizeable" components deserve separate publication * Identifiers should point to concrete versions * Plans for "active citations" and Authorea-minted DOIs --- ## Sharing Life-cycle * Pre-print and post-print feedback and dissemination * Adjustable privacy settings of comments --- ## Technology Life-cycle * Shelf-life of programming components tends to be in the order of decades * Is archival doomed? * No, because: 1. Shelf-life affects everything * Math notation and language change never stop evolving * Change happens "at the margins of cultural memory". 2. Science continuously reinvents itself * Important experiments are regularly repeated * Important tools and frameworks get reimplented in new languages --- class: center # Future Work Interactive, not yet interconnected --- ## UI Roadmap * Authoring still "proof-of-concept". Need: * Simple UX and widgets for creating active components * LaTeX is open-ended, hence equation markup - dirty. Simplify? * Reduce the learning curve * Fewer authoring languages * Less overall complexity --- ## Technology Wishlist * Live Mathematics * Common data language (JSON?) * Common APIs (community established!) * Accessibility * Universal native MathML support in browsers --- class: center ## Thank you! **Questions?** <video preload="auto" autoplay="autoplay" muted="muted" webkit-playsinline="" width='90%'> <source src="/assets/minions.webm" type="video/webm"> </video> --- # Bibliography * [Fang 2011] F. C. Fang and A. Casadevall. Retracted Science and the Retraction Index. Infection and Immunity, 79(10):3855–3859, aug 2011. * [Wren 2007] Jonathan D. Wren, Katarzyna Z. Kozak, Kathryn R. Johnson, Sara J. Deakyne, Lisa M. Schilling, Robert P. Dellavalle. The write position. A survey of perceived contributions to papers based on byline position and number of authors. EMBO Rep 8, 988–991 (2007).