Amine
Dadoun

Data Scientist & Researcher

Intro

ID


I'm Amine Dadoun, an industrial PhD student in Data Science at Eurecom in collaboration with Amadeus IT.

My research focuses on recommender systems applied to the travel domain, including Deep Learning based Recommender Systems, Knowledge Graph and Word Embeddings for Recommender Systems, and Session Based Recommender Systems.

My resume can be found here .

Experience

Experience

06/2018 -- Present
Doctoral Researcher. Amadeus IT Group, Nice.

  • Part of the Merchandising Division of the Airline IT entity in Amadeus, the main objective of the PhD is to study the impact of recommender systems on the offer construction and retailing of airline products. Several recommender systems are developed to tackle airline-specific recommendation use cases through the traveler journey.
  • Main Technologies: Azure ML, Python, Jupyter, Spark, SQL, Tensorflow, Pytorch, RDF, Knowledge Graphs.

11/2017 -- 05/2018
Data Scientist. Amadeus IT Group, Nice.

  • Consulting Data Science projects for some European airlines are conducted in two topics namely Price Optimization through A/B test techniques and Recommender Systems.
  • Main Technologies: Python, Spark, SQL, Sklearn, Java, Tableau.

03/2017 -- 09/2017
Data Scientist Intern. Amadeus IT Group, Nice.

  • Develop Deep learning-based recommender systems for ancillary recommendation in the traveler booking flow and integrate the system within Amadeus Merchandizing IT System.
  • Main Technologies: Python, Keras, Pandas, Java

Publications

Publications


2021

  • Dadoun, A., Troncy, R., Defoin-Platel, M., \& Ayala Solano, G. (2021). Predicting your next trip: A knowledge graph-based multi-task learning approach for travel destination recommendation–submitted. In Recsys ’21: Fourteenth acm conference on recommender systems, Amsterdam: Association for Computing Machinery.
  • Dadoun, A., Troncy, R., Defoin-Platel, M., Petitti, R., & Ayala Solano, G. (2021). Optimizing email marketing campaigns in the airline industry using knowledge graph embeddings .IEEE DataEngineering Bulletin Journal. KMEcommerce’21 Workshop, held in conjunction with WWW’21.
    .
  • Dadoun, A., Defoin-Platel, M., Fiig, T., Landra, C., & Troncy, R. (2021). How recommender systems can transform airline offer construction and retailing. Journal of Revenue and Pricing Management. doi:10.1057/s41272-021-00313-2.


    • 2020

  • Dadoun, A., & Troncy, R. (2020). Many-to-one recurrent neural network for session-based recommendation . arXiv:2008.11136

  • Abbas, N., Alghamdi, K., Dadoun, A., Domingue, J., Dumontier, M., Emonet, V., . . . Xu, W. (2020). Knowledge graphs evolution and preservation – a technical report from isws 2019 . arXiv:2012.11936.
  • Dadoun, A., Harrando, I., Lisena, P., Reboud, A., & Troncy, R. (2020). Two stages approach for tweet engagement prediction . arXiv:2008.10419 [cs.LG]3.
    .


    • 2019

  • Amine Dadoun, Raphael Troncy, Olivier Ratier, Riccardo Petitti. Location Embeddings for Next Trip Recommendation , LocWeb Workshop at WWW2019, San Francisco, USA, 2019.

    • Code

    Open source repos

    hand covering the sun
    • KGMTL4Rec: A python implementation of KGMTL4Rec model proposed in Predicting your next trip: A Knowledge Graph-Based Multi-task learning Approach for Travel Destination Recommendation. Submitted to the International 2021 RecSys Conference (RecSys'21).

    • TKE4Rec : is a python implementation of TKE framework proposed in Optimizing Email Marketing Campaigns the Airline Industry using Knowledge Graph Embeddings. Submitted to the International Workshop on Product Knowledge Graph for Ecommerce (PKG4Ecommerce'21).

    • Tweet Engagement Prediction: This repository contains source codes and experiments by the D2KLab in the context of the RecSys Challenge 2020. The Objective of the challenge is to predict users' engagements with a Tweet. Submitted to the International RecSys'20 Challenge WS (RecSys WS'20)

    • Deep Knowledge Factorization Machines (DKFM) provides top-N travel destination recommendations to travelers based on their history but also the travel context. Submitted to the International Web Conference (Locweb'19)

    • Many to One RNN for Hotel recommendation in Trivago platform The framework provides personalized ranking of hotels/accommodations for Trivago users. Submitted to the International RecSys'19 Challenge WS (RecSys WS'19)

    Contacts

    Location

    Nice, France

    Email
    amine.dadoun@gmail.com
    Top