Alexandros Mittos

Career Profile

I am Data Scientist and a Researcher passionate about Machine Learning, Social Computing, and Privacy. My prior work, among others, includes leveraging Machine Learning to accurately identify online hate speech and I have also conducted several large-scale studies on social networks using various NLP tools.

I have experience both with supervised and unsupervised learning (e.g., Scikit-Learn, Tensorflow, PyTorch) as well as various Natural Language Processing technologies and tools (e.g., BERT, NLTK).

In the past, I served as a Marie Skłodowska-Curie Fellow at University College London from where I received my Ph.D. and I was also a Research Assistant at Queen Mary University of London. I hold an M.Sc. in Information Security and a M.Eng. in Software Engineering.

Experience

Research Assistant

Aug 2019 - Aug 2020

Queen Mary University of London

• Used various Machine Learning tools to accurately identify online hate speech (Random Forest, SVM, BERT, Logistic Regression).
• Studied how different features and datasets affect the performance of hate speech classifiers.

Machine Learning Natural Language Processing Supervised Learning Hate Speech Detection Classification Feature Engineering Data Handling

Marie Skłodowska-Curie Fellow

Aug 2016 - Jul 2019

University College London

• Awarded a Horizon 2020 Marie Skłodowska-Curie Fellowship for the Privacy & Usability Innovative Training Network.
• Studied how direct-to-consumer genetic testing is being used by far-right groups to promote racist ideologies. To do so, I conducted several large-scale studies on social networks using various Natural Language Processing, Computer Vision, and Machine Learning tools.
  Media Coverage The Times, StatNews
• Critically evaluated and systematized the research produced by the genome privacy community in the context of privacy-enhancing technologies geared for testing, storing, and sharing genomic data.

Machine Learning Natural Language Processing Unsupervised Learning Computer Vision Social Network Analysis Sentiment Analysis Privacy Security

Visiting Researcher

Sep 2017 - Dec 2017

Bonn University

• Designed and ran a survey on the public perceptions of direct-to-consumer genetic testing.

Qualitative Research Survey Design

Visiting Researcher

Apr 2017 - May 2017

Data Protection Authority of Bavaria for the Private Sector (BayLDA)

• Studied how health and genetic data is being handled under GDPR.

GDPR

Skills & Proficiency

Expertise

A summary of my main expertise.

Machine Learning Classification Supervised Learning Unsupervised Learning Natural Language Processing Feature Engineering Data Handling Data Visualization NoSQL Privacy Security

Tools

A non-exhaustive list of tools that I am proficient with.

 Programming Languages: Python Java
 Libraries: Scikit-Learn Matplotlib Pandas Tensorflow PyTorch HuggingFace NLTK
 Methodologies: Agile Scrum
 Databases: mongoDB MySQL
 Tools: Git SVN Docker Bash

Transferable Skills

Skills that I have gained through experience.

Project Management Public Speaking Academic Writing Analytical Skills Technical Communication Flexibility & Versatility Critical Thinking

Selected Publications

• Mittos Alexandros, Savvas Zannettou, Jeremy Blackburn, and Emiliano De Cristofaro. 'And We Will Fight For Our Race!' A Measurement Study of Genetic Testing Conversations on Reddit and 4chan. In Fourteenth International AAAI Conference on Web and Social Media, 2020 (ICWSM).
  Media Coverage The Times, StatNews
  Acceptance Rate: 21%
  


• Mittos Alexandros, Bradley Malin, and Emiliano De Cristofaro. Systematizing Genome Privacy Research: A Privacy-Enhancing Technologies Perspective. Proceedings on Privacy Enhancing Technologies, 2019 (PoPETs).
  Acceptance Rate: 22%
  


• Mittos Alexandros, Savvas Zannettou, Jeremy Blackburn, and Emiliano De Cristofaro. Analyzing Genetic Testing Discourse on the Web Through the Lens of Twitter, Reddit, and 4chan. ACM Transactions on the Web, 2020 (TWEB).
  

Open-Source Projects

Online Harms: A Meta-Tool for Abusive Speech Detection: A machine learning meta tool for detecting abusive speech.
Machine Learning Abusive Speech Detection Feature Engineering

Oblivious Transfer Extensions: A concrete implementation of the IKPN03 protocol written in Java, using the SCAPI interface.
Java Cryptography

Daedalus: An academic schedule manager made for the department of Information and Communications Systems Engineering at the University of the Aegean.
Full Stack Development

Awards

Marie Skłodowska-Curie Fellow Scholarship

2016 - 2019

University College London

Received a Horizon 2020 Marie Skłodowska-Curie fellow scholarship for 3 years to investigate the societal challenges stemming from the rise of personal genomic testing. Acceptance Rate: 6%

University of the Aegean Scholarship

2015 - 2016

University of the Aegean

Received a scholarship for my M.Sc. studies at the University of the Aegean.