HOMER

Call for collaborators

Human Oriented autoMated machinE leaRning

About the project

Where: MI2DataLab at Warsaw University of Technology. Our GitHub.

With whom: Przemysław Biecek Google Scholar, GitHub.

What:

Machine learning models are used everywhere. Predictive modelling fundamentally changed data- driven disciplines like health-care, biology, finance, legal, military, security, transportation, and many more. The increasing availability of large annotated data sources combined with recent developments in machine learning models leads to the next industrial revolution.

BUT: Predictive models are handcrafted by data scientists in a tedious and laborious process. Most of the time spent on data exploration and model training is a set of try-and-error experiments. Models become more and more complex to catch even vague signals. Techniques like boosting, bagging or neural networks result in models with thousands or millions of parameters. Lack of understanding of complex models and poor automation results in problems with replicability and quality of models.

The main goal of this project is to develop new methods for human-oriented model exploration, interpretable model audits and automated model assembly. The newly appointed research team will create a grammar for human - model interaction.

Post-doc

Required:

  • PhD in Computer Science, Mathematics, Statistics or similar. (Different background? We encourage cross-domain research short visits).
  • Experience in Scientific Programming (R and/or Python and/or C++)
  • Good track of Scientific Records

Scope of work:

  • Automated model exploration
  • Interpretable measures for model performance
  • Meta/transfer learning in automated model development
  • Automated model validation
  • Experiments with XAI for deep learning

Offer:

  • Excellent atmosphere for work in a young and very active ML lab
  • Conferences and training budget
  • Short visits in cooperating abroad group
  • Access to CPU / GPU clusters
  • Full-time job at Warsaw University of Technology
  • 10 000 PLN gross per month (plus results-driven extras)
  • contract for 6 months (short visit) / 12 months or 24 months (long visit)

PhD student (2 positions)

Required:

  • MSc in Computer Science, Mathematics, Statistics or similar. (Different background? We encourage cross-domain research short visits).
  • Experience in Scientific Programming (R and/or Python and/or C++)
  • Strong motivation

Scope of work:

  • Automated model exploration
  • Conferences and training budget
  • Concept drift detection
  • Reinforcement learning for Human Centered AI
  • Knowledge extraction from representations
  • Experiments with XAI for deep learning

Offer:

  • Excellent atmosphere for work in a young and very active ML lab
  • Conferences and training budget
  • Short visits in cooperating abroad group
  • Access to CPU / GPU clusters
  • PhD scholarship
  • 3 500 PLN net per month (plus results-driven extras)
  • contract for 6 months (short visit) / 3.5 years (full PhD)

Scientific programmers (2 positions)

Required:

  • Experience in Scientific Programming (R and/or Python)
  • Strong motivation

Scope of work:

  • Development of scientific software
  • Maintaining of existing projects
  • Experiments with XAI for deep learning

Offer:

  • Excellent atmosphere for work with experienced researchers
  • Contract for a specific job
  • In a part-time settings (1 day per week) 1 500 PLN gross per month
  • short (3 months) and long (up to 3 years) contracts

How to apply?

Send

  • your resume
  • your best publication
  • link to your best software project
  • other relevant information

to Przemyslaw.Biecek at gmail.