Skylads Research Lab

Research Topics

Research topics informed by the media buying industry needs, within 6 main research areas:

  • Strong Artificial Intelligence

  • Weak Artificial Intelligence

  • Big Data Science

  • Natural Language Processing (NLP)

  • Human-Machine Cooperation

  • Distributed Computing

Publications

Leading to published papers and conference participations (2018):

  • A New Layer of Optimization for Real-Time Bidding Campaigns (accepted in journal IDA: Intelligent Data Analysis international journal)

  • Fast and Accurate Temporal Data Classification using Nearest Weighted Centroid (published and presented in research conference   ICCAIRO: International Conference on Control, Artificial Intelligence, Robotics and Optimization – Prague, Czech Republic in June 2018)

  • Ensemble Learning using Frequent Itemset Mining for Anomaly Detection (research conference: 8th International Conference on  Artificial Intelligence, Soft Computing and Applications (AIAA-2018) – Melbourne, Australia in November 2018)

  • Click-Through Rate (CTR) Prediction based on Feature Engineering

  • Digital Media Buying Campaigns Structural Optimization with Automatic Partitioning

Skylads Research Lab Team Members

 
 
Saeid Soheily-Khah.jpeg

Saeid Soheily-Khah

Saeid graduated in computer engineering, and received master degree in artificial intelligence & robotics in 2005. He then received his second master in information analysis and management from Skarbek university in Warsaw.In 2013, he joined to the LIG (Laboratoire d’Informatique de Grenoble) at Universit ́e Grenoble Alpes as a doctoral researcher. He successfully defended his dissertation and got his Ph.D in Oct 2016. In Nov 2016, he joined to the IRISA/Expression at Universit Bretagne Sud as a postdoctoral researcher. Lastly, in Oct 2017, he joined Skylads as a research scientist. His research interests are machine learning, data mining, cyber security system, anomaly detection, digital advertising and artificial intelligence.

 
Yiming Wu.jpeg

Yiming Wu

Yiming received his B.S.E.E. degree from Northwestern Polytechnical Uni- versity, Xian, China, 2011. He received his Ph.D. degree in Electrical Engineering from University of Technology of BelfortMontbliard, Belfort, France, 2016. He joined Skylads as a research scientist in 2018, and his research has addressed topics on machine learning, artificial intelligence and digital advertising.

 
 

Research Work presented at Conferences

 
 

Skylads Research Lab Team is regularly invited to participate in academic conferences to present their state-of-the-art research work on Digital Marketing Artificial Intelligence.

Below is a list of recent conferences in which the Skylads Research Lab work has been presented.

 
ICCAIRO AI Conference.jpg
LOGO cps_published_logo_blue_black.jpg

Fast and Accurate Temporal Data Classification using Nearest Weighted CentroiD

  • Conference name: International Conference on Control, Artificial Intelligence, Robotics and Optimization (ICCAIRO 2018)

  • Date and location: May 21-23, 2018, Prague, Czech Republic

  • Status: accepted + published

  • Author(s): s. soheily-khah

  • Abstract: propose a fast accurate nearest weighted centroid classifier for temporal data, to speed up the nearest neighbor classification algorithm, and specially to make it applicable for huge datasets. Since the computational complexity of comparing the temporal data is too high, and these comparisons pose some problems for large data, to obtain fast accurate results by reducing the data size, classification based on nearest centroid could be one solution.

  • Publication url: http://www.iccairo.org/

 
aiaai2018image.jpg
logo AIAA 2018.png

Ensemble learning using frequent itemset mining for anomaly detection

  • Conference name: 8th international conference on Artificial Intelligence, soft computing and Applications (AIAA 2018)

  • Date and location: November 24-25, 2018, Melbourne, Australia

  • Status: accepted + published

  • Author(s): s. soheily-khah and y. wu

  • Abstract: propose a hybrid supervised learning of anomaly detection using frequent itemset mining and random forest with an ensemble probabilistic voting method. The main contribution is to boost base (weak) learners to strong learners by ensemble learning, which can make very accurate classifiers.

  • Publication url: https://cndc2018.org/aiaa/papers.html

 
southbank SAI 2019 image.jpg
logo SAI 2019.png

An Enhanced Ad Event-Prediction Method Based on Feature Engineering

  • Conference name: 8th international conference on Soft computing, Artificial Intelligence and applications (SAI 2019)

  • Date and location: June 29-30, 2019, Copenhagen, Denmark

  • Status: accepted + to be presented & published

  • Author(s): s. soheily-khah and y. wu

  • Abstract: introduce an enhanced method for ad event prediction (i.e. clicks, conversions) by proposing a new efficient feature engineering approach. To do so, we propose two novel adjusted statistical measures for feature selection and provide an enhanced framework for ad event-prediction by analyzing the huge amount of historical data.

  • Publication url: https://icaita2019.org/sai/