Demand forecasting is the problem of predicting the amount of goods or services demanded by customers during some future time range: a critical application for many businesses. Retailers base in-stock management decisions like ordering and storage, as well as supply chain management, on demand forecasts. Energy utility companies use forecasting for scheduling operations, investment planning and price bidding. The data revolution creates new opportunities to improve forecast accuracy and granularity, given that heterogenous data sources can be integrated.

The focus of the workshop is on demand forecasting by means of data-driven techniques, with a specific emphasis on retail, energy, and transportation industries. We aim at bringing demand forecasting problems to the attention of the machine learning community, where many data-driven success stories have originated. We hope to identify the most important challenges from a business point of view, and to start a focussed discussion on how to formalize and solve them by means of machine learning techniques, or on which tools are missing and require additional research efforts.

The workshop will bring together researchers from industry and academia in order to recognize synergies between the fields. Challenges in demand forecasting include:

  • Dealing with uncertainty in demand forecasting
  • Hierarchical and granular level forecasting
  • Methods for combining forecasts and their effectiveness
  • Leveraging data from social media to improve forecasts
  • Using demand data for product and customer segmentation
  • Demand forecasting of short life span products/services
  • Forecasting the demand and price of new products/services
  • Quantifying the impact of marketing campaigns on the demand
  • Intermittent demand forecasting

Among techniques studied in machine learning, the following could have an impact on data-driven demand forecasting:

  • Deep Learning / Neural Networks
  • Markovian state space models
  • Gaussian processes
  • Kernel Methods
  • Approximate Bayesian Inference
  • Ensemble Methods
  • Functional Data Analysis
  • Transfer and Multi-task Learning
  • Causal Inference
We are particularly concerned with identifying methodologies that allow to tackle large scale problems. A further goal of the workshop is to identify a collection of high quality publicly available datasets that can be used as benchmarks for machine learning and demand forecast researchers.

Invited Speakers
  • Nicolas Chapados, ApSTAT Technologies
  • Gregory Duncan, Amazon and University of Washington
  • Shie Mannor, Technion
  • Jean Michel Poggi, Univ. Paris Descartes and Univ. Paris-Sud Orsay, LMO
  • Brian Seaman, Walmart Labs
  • Gilles Stoltz, CNRS
  • Rafal Weron, Wroclaw University of Technology
  • Felix Wick, Blue Yonder

  • Francesco Dinuzzo, IBM Research
  • Mathieu Sinn, IBM Research
  • Yannig Goude, EDF R&D
  • Matthias Seeger, Amazon

  • IBM Research
  • EDF R&D
  • Amazon

Venue and Registration

This workshop is co-located with the 32nd International Conference on Machine Learning. For information about the venue, please visit the ICML 2015 website.

All participants need to register. Information about registration and fees can be found here .