Dr. Salar Mohtaj

I'm a Senior Research Scientist at German Research Centre for Artificial Intelligence (DFKI) at the Speech & Language Technology group in Berlin, where my work focuses on evaluating large language models for quality and factual accuracy. I got my PhD in Computer Science at TU Berlin, Advised by Prof. Sebastian Möller. I spent eight-plus years working across fake news detection, hate speech detection, plagiarism detection and argument mining.

• Dec. 2019 — Present

  Senior Research Scientist

  DFKI, Speech & Language Technology group · Berlin

• Jan 2024 - March 2024

  Visiting Researcher

  Universitat Politècnica de València

• Mar – Dec 2019

  Research Scientist

  Technische Universität Berlin, Quality & Usability Lab

• April 2014 — Feb 2019

  NLP Specialist

  ICT Research Center, ACECR

• Sep. 2016 — June 2018

  Guest Lecturer, Data Mining

  University of Science and Culture

• 2020 — 2024

  PhD, Computer Engineering

  Technische Universität Berlin, Germany · advised by Prof. Sebastian Möller

• 2011 - 2014

  MSc, Information Technology

  Amirkabir University of Technology, Iran

• 2006 - 2011

  BSc, Information Technology

  Shahrood University of Technology, Iran

• 2024

  From Construction to Application: Advancing Argument Mining with the Large-Scale KIALOPRIME Dataset

  COMMA 2024Sahitaj, Ruiz-Dolz, Sahitaj, Nizamoglu, Schmitt, Mohtaj, Möller

• 2024

  Augmented Political Leaning Detection: Leveraging Parliamentary Speeches for Classifying News Articles

  CPSS @ KONVENSJakob, Wenzel, Mohtaj, Schmitt

• 2022

  On the Importance of Word Embedding in Automated Harmful Information Detection

  TSD 2022Mohtaj, Möller · Springer LNCS

• 2022

  Overview of the GermEval 2022 Shared Task on Text Complexity Assessment of German Text

  GermEval @ KONVENSMohtaj, Naderi, Möller

• 2022

  The Impact of Pre-processing on the Performance of Automated Fake News Detection

  CLEF 2022Mohtaj, Möller

• 2019

  Using External Knowledge Bases and Coreference Resolution for Detecting Check-Worthy Statements

  CLEF 2019Mohtaj, Himmelsbach, Woloszyn, Möller

• 2019

  Automated Text Readability Assessment for German Language: A Quality of Experience Approach

  QoMEX 2019Naderi, Mohtaj, Karan, Möller

• 2018

  Parsivar: A Language Processing Toolkit for Persian

  LREC 2018Mohtaj, Roshanfekr, Zafarian, Asghari

• 2016

  Algorithms and Corpora for Persian Plagiarism Detection: Overview of PAN at FIRE 2016

  FIRE 2016Asghari, Mohtaj, Fatemi, Faili, Rosso, Potthast

Cited by 400+ · full list on Google Scholar

• HASOC — Hate Speech & Offensive Content Detection

  GitHub

  Feature-extraction and character-based LSTM models built for the HASOC 2020/2021 shared tasks, identifying hate speech and offensive content across English, German, and Hindi social media text.

• PerPaDa — Persian Paraphrase Dataset

  GitHub

  A 2,446-instance Persian paraphrase corpus collected via implicit crowdsourcing from a plagiarism-detection system's user submissions, published at LREC 2022.

• SemEval 2022 Task 8 — Multilingual News Similarity

  GitHub

  A shared-task submission scoring how similar pairs of news articles are across 18 language combinations — cross-lingual groundwork for later fact- and news-evaluation work.

• Intelligent Crawler — Open NLP Data Collection

  GitHub

  A configurable web crawler built with the TU Berlin NLP group to collect open text corpora for downstream NLP research and dataset construction.

• Natural Language Processing - Lecture @ TU Berlin
• Natural Language Processing - Online Course @ KI-Campus
• Data Mining - University of Sience and Culture