Yiannos A. Stathopoulos
I am a researcher at the University of Cambridge, Department of Computer Science and Technology working on the ALEXANDRIA project. My work on the ALEXANDRIA project revolves around usersupport tools and methods, based on AI and machine learning, for the interactive theorem prover (ITP) Isabelle.
I have the privillage of collaborating with Dr Alexis Litvine and Dr Oliver Dunn on the THOTH project. THOTH is an initiative within Cambridge University to apply AI and machine learning in the humanities and social sciences. Presently, the team is working on automatic extraction of cell data from tables in historical documents.
My background is in Information Retrieval (i.e., search), Artificial Intelligence, Machine Learning, Computer Vision and Natural Language Processing (NLP)
I have previously worked on extracting and parsing mathematical expressions directly from PDF documents. As part of this work, supervised by Dr Brian Harrington, I have:
 Built tools that extract text box, text line and character data from PDFs and produce machinereadable XML representations of page data. I modified pdftotext to align data in PDF documents with rasterised pages in PDF documents.
 Built Mathalyzer, an interactive machine learning annotation and exploratory tool for automaticall extracting and parsing mathematical expressions from PDF documents (see below).
 Obtained extensive experience working with computer vision algorithms and the OpenCV library (in C++ and Python).
My PhD was on Mathematical Information Retrieval (MIR) of research mathematics under the supervision of Dr. Simone Teufel.
Education
 BSc in Computer Science, First class  University of Nottingham
 MSc in Statistics  University of Nottingham
 MSc in Computer Science  University of Oxford
 PhD in Computer Science  University of Cambridge
Publications
2020

SErAPIS: A ConceptOriented Search Engine for the Isabelle Libraries Based on Natural Language
Yiannos A. Stathopoulos, Angeliki KoutsoukouArgyraki, Lawrence Paulson
Isabelle Workshop 2020. March, 2020
2019

Developing a ConceptOriented Search Engine for Isabelle Based on Natural Language: Technical Challenges
Yiannos A. Stathopoulos, Angeliki KoutsoukouArgyraki, Lawrence Paulson
Artificial Intelligence and Theorem Proving (AITP) 2020. December 2019
2018

Variable Typing: Assigning Meaning to Variables in Mathematical Text
Yiannos A. Stathopoulos, Simon Baker, Marek Rei and Simone Teufel
In Proceedings of the 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL 2018) New Orleans, United States, 2018
2016

Mathematical Information Retrieval Based on Type Embeddings and Query Expansion
Yiannos A. Stathopoulos and Simone Teufel
In Proceedings of the 26th International Conference on Computational Linguistics (Coling 2016). Osaka, Japan, 2016.
2015

Retrieval of researchlevel mathematical information needs: A Test Collection and Technical Terminology Experiment
Yiannos A. Stathopoulos and Simone Teufel
In Proceedings of the Short Papers of the 53rd Annual Meeting of the Association for Computational Linguistics (ACL 2015). Beijing, China, 2015.
2011

OMEX: Software for Mining Mathematical Expression Semantics from Scientific Documents.
Yiannos A. Stathopoulos and Brian Harrington
In Proceedings of the 2011 IEEE Fifth International Conference on Semantic Computing (ICSC '11). IEEE Computer Society, Washington, DC, USA, 209210. DOI=10.1109/ICSC.2011.65 http://dx.doi.org/10.1109/ICSC.2011.65
Code and Data Downloads
 Download the Cambridge University MathIR Test Collection (for retrieval of researchlevel mathematics) described in
"Retrieval of researchlevel mathematical information needs: A Test Collection and Technical Terminology Experiment"  Download the Cambridge Dictionary of Mathematical Types (CDMT) seed type dictionary (10601 phrases), goldstandard data set for type detection from "Mathematical Information Retrieval Based on Type Embeddings and Query Expansion" and extended type dictionary (1.23m phrases) from "Variable Typing: Assigning Meaning to Variables in Mathematical Text"
 Download the Variable Typing Data Set for assigning meaning to mathematical variables using Machine Learning
Cool things I've built
This is a partial list of cool stuff I've built.

Mathalyzer  an interactive tool for analysing mathematical formuale in PDF documents. Written in C++ and GTK+, this tool employs the PresentationAbstractionControl (PAC) pattern to synchronise multiple data elements in a unified presentation. The idea behind Mathalyzer is to produce a tool that combines elements of Acrobat, Photoshop and SPSS.
 Spine  A small C++ library, forked from the subsystems of Mathalyzer, that implements PresentationAbstractionControl (PAC) message passing with GTK+ controls. This library is used to synchronise the datamodel of GUI apps, with various independent GUI elements implemented in GTK+.
 Interval and range trees  A small C++ library of interval and range trees for optimising the Mathalyzer canvas. My implementation of interval and range trees is built on top of Redblack trees. Upon rotation, the RB tree implementation raises a rotation event. Event handlers at higher levels are responsible for applying transformations that reestablish the invariants of the interval and range trees.
 OMEX  Software that detects and extracts mathematical expressions from PDF. The pipeline is the subject of my paper with Dr. Brian Harrington. Mathalyzer was built to extend aspects of this pipeline with machine learning.
 MapReduce in C++  I built a small C++ implementation of Google's MapReduce. The implementation is designed to abstract parallelisation of tasks using Mappers, Grouppers and Reducers on multicore systems.