Multivariate time series forecasting of transactional data - BC-891

Genre de projet: Recherche
Discipline(s) souhaitée(s): Génie - informatique / électrique, Génie, Informatique, Sciences mathématiques
Entreprise: Mastercard AI Garage
Durée du projet: 6 mois à 1 an
Date souhaitée de début: Dès que possible
Langue exigée: Anglais
Emplacement(s): BC, Canada
Nombre de postes: 1
Niveau de scolarité désiré: CollègeÉtudes de premier cycle/baccalauréatMaîtriseDoctoratRecherche postdoctoraleNouvelle diplômée/nouveau diplômé
Ouvert aux candidatures de personnes inscrites à un établissement à l’extérieur du Canada: Yes

Au sujet de l’entreprise: 

We work to connect and power an inclusive, digital economy that benefits everyone, everywhere by making transactions safe, simple, smart and accessible. Using secure data and networks, partnerships and passion, our innovations and solutions help individuals, financial institutions, governments and businesses realize their greatest potential. Our decency quotient, or DQ, drives our culture and everything we do inside and outside of our company. We cultivate a culture of inclusion for all employees that respects their individual strengths, views, and experiences. We believe that our differences enable us to be a better team – one that makes better decisions, drives innovation and delivers better business results. At AI Garage, we use state-of-the-art AI techniques to solve some of the most important problems in the financial world.

Veuillez décrire le projet.: 

Forecasting transactional attributes for multiple transactions is crucial to many use-cases at Mastercard like revenue and GDV forecasting, TransactionGPT, etc. In the past, we have explored generative data augmentation for multivariate time series on a short and long-term basis [paper link]. However, traditional forecasting methods have limitations in handling complex patterns and long-term dependencies in time series data. Transformers, which have shown remarkable success in natural language processing tasks, have the potential to address these challenges and improve forecasting accuracy. This proposal outlines a project to develop a multivariate long-term time series forecasting model using transformers.

The primary objectives of this project are as follows:

• Data Preprocessing: Create a robust data preprocessing pipeline to handle multivariate time series data, including feature engineering and normalization.
• Model Development: Develop a transformer-based neural network architecture tailored for multivariate time series forecasting.
• Evaluation Metrics: Define appropriate evaluation metrics to measure the forecasting model's performance, considering both short-term and long-term predictions.
• Hyperparameter Tuning: Optimize hyperparameters to improve the model's forecasting accuracy and efficiency.
• Scalability: Ensure the model can scale to handle large and complex time series datasets used at Mastercard.

Expertise ou compétences exigées: 

• Good theoretical and practical familiarity of Deep Learn Models
• Good understanding of Machine Learning theory
• Decent understanding of probability theory and statistics
• Good experience with Python, Pytorch or Tensorflow