Maximilian Mozes

I'm a final year undergraduate Computer Science student at the Technical University of Munich (TUM) with a minor in Mathematics. I am currently writing my thesis on deep learning-based concepts at the intersection of natural language processing and computer vision in the framework of a joint collaboration between TUM's Computer Vision Group and the University of Munich's Center for Information and Language Processing.

I recently worked as visiting research scholar at the Language and Information Technologies Group of the University of Michigan's Artificial Intelligence Lab, where my focus was on natural language processing research. In particular, I was involved in multiple projects dealing with the concept of semantic plausibility in human language. My advisor was Rada Mihalcea.

Apart from that, I worked as research intern in the Department of Psychology at the University of Amsterdam under supervision by Bennett Kleinberg and Bruno Verschuere. After a great internship I am now working in collaboration with them on conceptualising and realising automated cognition-based approaches for verbal deception detection by using concepts from the field of computational linguistics, statistical data analyses and supervised learning algorithms.

Email  /  GitHub  /  Twitter  /  Google Scholar  /  LinkedIn

Identifying the sentiment styles of YouTube's vloggers
Bennett Kleinberg, Maximilian Mozes, Isabelle van der Vegt.
To appear in the proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), Brussels, Belgium, 2018.
paper / dataset
Using Named Entities for Computer-Automated Verbal Deception Detection
Bennett Kleinberg, Maximilian Mozes, Arnoud Arntz, Bruno Verschuere.
The Journal of Forensic Sciences, 63, 3, p. 714 - 723, 2017.
paper / code

We propose named entity recognition (NER; i.e., the automatic identification and extraction of information from text) to model three established theoretical principles: (i) truth tellers provide accounts that are richer in detail, (ii) contain more contextual references (specific persons, locations, and times), and (iii) deceivers tend to withhold potentially checkable information. We test whether NER captures these theoretical concepts and can automatically identify truthful versus deceptive hotel reviews. Named entities discriminated truthful from deceptive hotel reviews above chance level, and outperformed the lexicon approach and sentence specificity.

Web-based text anonymization with Node.js: Introducing NETANOS (Named entity-based Text Anonymization for Open Science)
{Bennett Kleinberg, Maximilian Mozes}.
The Journal of Open Source Software, 2, 14, 2017.
paper / code

NETANOS (Named Entity-based Text ANonymization for Open Science) is a natural language processing software that anonymizes texts by identifying and replacing named entities. The key feature of NETANOS is that the anonymization preserves critical context that allows for secondary linguistic analyses on anonymized texts.

NETANOS is also available as a Node.js package on npm.

NETANOS - Named entity-based Text Anonymization for Open Science
Bennett Kleinberg, Maximilian Mozes, Yaloe van der Toolen.
Preprint, 2017.
preprint / code

We propose and empirically evaluate NETANOS: named entity-based text anonymization for open science. NETANOS is an open-source context-preserving anonymization system that identifies and modifies named entities (e.g. persons, locations, times, dates). The aim is to assist researchers in sharing their raw text data. NETANOS anonymizes critical, contextual information through a stepwise named entity recognition (NER) implementation: it identifies contextual information (e.g. "Munich") and then replaces them with a context-preserving category label (e.g. "Location_1").

Technical reports
Sampling from High-Dimensional Probability Distributions: An Introduction to Markov Chain Monte Carlo Methods
Maximilian Mozes.
Seminar paper - TUM Seminar "Computational Aspects of Machine Learning" (winter semester 2017).

An Introduction to Statistical and Probabilistic Linear Models
Maximilian Mozes.
Seminar paper - TUM Proseminar "Data Mining" (summer semester 2017).


Linguistic temporal trajectory analysis - A dynamic approach to text data
Bennett Kleinberg, Maximilian Mozes, Isabelle van der Vegt.
2nd European Symposium on Societal Challenges in Computational Social Science: Bias and Discrimination, December 2018, Cologne, Germany.


Chat-based information elicitation in verbal deception detection
Bennett Kleinberg, Maximilian Mozes, Yaloe van der Toolen.
Department of Psychology, University of Amsterdam, September 2016 , Amsterdam, Netherlands.

In the framework of my internship at the Department of Clinical Psychology at the University of Amsterdam (UvA) I have built a Node.js/ chat environment tailored to elicit information in chat-based interviews. The chat environment has been applied in research studies related to the field of verbal deception detection. I presented the environment's functionality to participants in our mini-workshop on chat-based information elicitation.

Teaching activities

Tutor/teaching assistant: Analysis for Computer Science
Technical University of Munich, Winter term 2018/19.

Organizing tutoring sessions in "Analysis for Computer Science" for undergraduate students in Informatics/Computer Science.

2018 - Munich, Germany.

Forked from jonbarron_website.