A weekend experiment where I attempted to use [GGML](https://github.com/ggerganov/ggml/) quantized tensors to run a state-of-the-art code completion model on commodity hardware including laptops, desktops, ARM machines like Macbooks and even Raspberry Pis. As the GGML library matures, I'm adding support for things like Nvidia GPU support too.
A scientific paper indexing system that uses machine learning to enrich papers in order to make them more easy to search and filter. Originally written in Python 2 with xml-rpc worker processes and recently updated to use Python 3 and [dramatiq](https://dramatiq.io/) for concurrency.
An NLP pipeline for processing and enriching scientific papers with sentence-level information about their core scientific concepts (CoreSCs). This is a Python 3 implementation of Prof Maria Liakata's 2010 paper. We provide a free web service for low volume requests and a simple to use docker configuration for those who want to run the software over a larger number of papers.
A tool for annotating co-references of entities that occur in linked news paper article/scientific paper pairings. Some 'sharp' code but this was my first venture into 'full stack' using ReactJS on the frontend and Flask on the backend. The repository also contains the final corpus which we made available as part of our EACL21 publication.
A small command-line tool I wrote for monitoring my time spent on projects - it has API integration with the popular SaaS timesheet tool [Harvest](https://www.getharvest.com/)
<li>Maufe, M., <strong>Ravenscroft, J.</strong>, Procter, R., & Liakata, M. (2022, December). <ahref="">A Pipeline for Generating, Annotating and Employing Synthetic Data for Real World Question Answering.</a> In Proceedings of the The 2022 Conference on Empirical Methods in Natural Language Processing: System Demonstrations (pp. 80-97).
<li><strong>Ravenscroft J.</strong>, Cattan A., Clare A., Dagan I., Liakata M. <ahref="https://aclanthology.org/2021.eacl-main.21.pdf">CD<sup>2</sup>CR: Co-reference Resolution Across Documents and Domains</a> In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume (pp. 270-280).</li>
<li>Steffens A,, Campello A., <strong>Ravenscroft J.</strong>, Clark A., Hagras H. <ahref="http://ceur-ws.org/Vol-2380/paper_151.pdf">Deep Segmentation: using Deep Convolutional Networks for Coral Reef pixel-wise Parsing</a><em>technical notes</em></li>
<li><strong>Ravenscroft, J.</strong>, Clare, A., & Liakata, M. <ahref="http://aclweb.org/anthology/P18-4004">HarriGT: Linking news articles to scientific literature.</a> In Proceedings of ACL 2018, System Demonstrations (pp. 19-24).</li>
<li><strong>Ravenscroft, J.</strong>, Clare, A., Duma, D., Liakata, M. <ahref="https://doi.org/10.1371/journal.pone.0173152">Measuring scientific impact beyond academia: An assessment of existing impact metrics and proposed improvements</a> PloS one, 12(3), e0173152.</li>
<p><strong>Ravenscroft J.</strong>, Oellrich A., Saha S., & Liakata M. <ahref="http://www.lrec-conf.org/proceedings/lrec2016/summaries/928.html">Multi-label Annotation in Scientific Articles - The Multi-label Cancer Risk Assessment Corpus</a></p>
</li>
<li>
<p>Duma D., Liakata M., Clare A., <strong>Ravenscroft J.</strong>, & Klein E. <ahref="http://www.lrec-conf.org/proceedings/lrec2016/pdf/676_Paper.pdf">Applying Core Scientific Concepts to Context-Based Citation Recommendation</a></p>
</li>
<li>
<p>Duma D., Liakata M., Clare A., <strong>Ravenscroft J.</strong>, & Klein E. <ahref="https://doi.org/10.1045/september2016-duma">Rhetorical Classification of Anchor Text for Citation Recommendation</a></p>
</li>
</ul>
### 2013
<ul>
<li><strong>Ravenscroft, J.</strong>, Liakata, M., & Clare, A. <ahref="https://doi.org/10.1007/978-3-319-02621-3_26">Partridge: An Effective System for the Automatic Classification of the Types of Academic Papers</a></li>
