El Mehdi Rochd

Training and Experience:

• Post-doctoral student under a Franco-Indian contract (CEFIPRA), in the Social Networks and Connected Objects Chair at the Institut Mines-Télécom Business School.
• Expert in Data Science / Semantic Web: Development and enrichment of film industry ontologies using NLP and NLU models for a company specializing in the development of commercial and financial management service for creative industry.
• PhD in computer science (datamining / machine learning) from the University of Aix-Marseille, Very honorable mention. Subject: Probabilistic models for online consumers analysis: recommendation and personalization.

Methodologies:

• Topic Models : knowledge of state-of-the-art models and development of new models (based on the EM algorithm).
• Graphical models: passing from sequences to hierarchies using conditional random fields.
• Semantic web: OWL, RDF, RDFS, SPARQL.
• Word embedding : combining semantic representation of words with ontologies to develop relevant retrieval information systems.

Programming Languages and databases:

• Programming languages : Python, VBA, Java, C.
• Databases : MySQL, PostgreSQL.
• BI decision-making tools : BI OLAP Cube tools, MDX, JPivot, Mondrian, Pentaho.

Publications:

• El Mehdi Rochd, Christine Balagué. A predictive system based on the inclusion of conditional random fields features within topic models. In Marketing Science: the marketing journal of INFORMS, 2021 (Soumis).
• Christine Balagué, El Mehdi Rochd. Towards using responsible artificial intelligence in product recommender systems in marketing. In 2019 INFORMS Marketing Science Conference.
• El Mehdi Rochd, Mohamed Quafafou. A Top-N Recommender Model with Partially Predefined Structure. In The IEEE International Conference on e-Business Engineering (ICEBE, 2014).
• El Mehdi Rochd, Mohamed Quafafou. A Nonparametric Mixture Model for Personalizing Web Search. In The International Symposium on Intelligent Data Analysis (IDA, 2014).
• El Mehdi Rochd, Mohamed Quafafou, Éric Cholet. A Topic Model-based Personalization Over Time. In The Second Workshop on User Engagement Optimization at KDD (UEO-KDD, 2014).
• El Mehdi Rochd, Mohamed Quafafou, Mustapha Aznag. Encoding Local Correspondence in Topic Models. In The IEEE International Conference on Tools with Artificial Intelligence (ICTAI, 2013).
• Mustapha Aznag, Mohamed Quafafou, El Mehdi Rochd, Zahi Jarir. Probabilistic Topic Models for Web Services Clustering and Discovery. In The European Conference on Service-Oriented and Cloud Computing (ESOCC, 2013).
• El Mehdi Rochd. Annotation automatique d’images à l’aide d’un modèle probabiliste à variables latentes et à structure dynamique. In 10 èmes Journées des doctorants du LSIS (JDL, 2013).