Master Thesis, 30 hp: Valid and Reachable Counterfactual Explanations for Reinforcement-Learning Policies in Pursuit–Evasion
Teknik, data och digitalt · Data, AI och analys · Maskininlärning · Artificiell intelligens · Mjukvaruutveckling
I korthet
Saab is seeking a Master Thesis student for a 30 hp project investigating valid and reachable counterfactual explanations for Reinforcement Learning policies in pursuit-evasion scenarios. The role involves implementing and comparing methods for generating counterfactuals, developing a visualization UI, and contributing to research on understanding autonomous decision-making.
Ansvarsområden
- Investigate to what extent do minimal counterfactual state changes provide faithful, feasible, and stable explanations of an RL agent’s decisions in a sequential pursuit-evasion environment.
- Implement a baseline method that identifies minimal state changes required to alter an agent’s action.
- Compare the baseline method with a dynamics-constrained method that generates only feasible and reachable counterfactuals.
- Quantitatively evaluate counterfactual explanations for action validity, reachability, stability, intervention faithfulness, and trajectory-level consequences.
- Develop a simple user interface to visualize the counterfactuals.
- Evaluate the user interface through user studies.
- Engage with the Planning, Optimization and Reinforcement Learning unit’s RL evaluation and debugging team.
- The resulting XRL tool will be added to the toolbox and made available to the unit’s evaluation and debugging capabilities of advanced AI techniques.
Krav
- Student eager to apply theoretical knowledge and fresh perspectives to real-world challenges.
- Interest in reinforcement learning, eXplainable AI and interface design using visualization techniques.
- In the final stage of civilingenjör programmes – Computer Science and Engineering, Systems and Controls, Autonomous Systems, Robotics and AI and Machine Learning profile or equivalent.
- About to begin a 30 hp thesis project.
- Experience with PyTorch, TensorFlow or JAX.
- Good understanding of MDPs, policies, value functions and deep RL.
- Ability to work with an existing research codebase using Git.
- Basic knowledge in statistics, e.g., confidence intervals, multiple random seeds, hypothesis testing and interpreting variance.
- Interested in analysing why an agent behaves as it does.
- Must pass a security vetting.
Önskade kvalifikationer
- Academic rigor and enthusiasm for cutting-edge technology.
- Unique and insightful lens to approach problems.
Förmåner
- Opportunity to collaborate with experienced engineers and specialists.
- Gain invaluable practical experience.
- Make a tangible contribution to our growth and development.
- Support and guidance needed to translate theoretical knowledge into practical solutions.
- Work with cutting edge technology, creativity and an innovative environment.
- Contribution to keeping people and society safe.
- Opportunities to explore a wealth of possibilities, take on challenges, create smart inventions, and grow beyond.
Färdigheter
PyTorchTensorFlowJAXGit
#reinforcement learning#XAI#master thesis#counterfactual explanations#pursuit-evasion
Om Saab
As a leading defence and security company, we offer solutions that range from the depths of the oceans to high in the sky, on land and in cyberspace, to keep people and society safe. Empowered by our 22,000 talented people, we constantly push the boundaries of technology to create a safer, more sustainable and more equitable world. We design, manufacture and maintain advanced systems in aeronautics, weapons, command and control, sensors and underwater systems. We are headquartered in Sweden and have major operations all over the world and are part of the domestic defence capability of several…
Företag
SaabPublicerade jobb
för 2 dagar sedan
Upphör
om 1 vecka
Anställningstyp
Praktik
Arbetsform
På plats
Erfarenhetsnivå
Junior
Platser
Linköping, Sweden
Kvalifikation
Kandidatexamen
Sökande
Ansök tidigt
Liknande jobb
Stockholms stad
Farsta, Sweden +1 flerSök extra hos oss på hemtjänsten i Farsta
Tillfällig Ansök tidigt Publicerad för 4 dagar sedan
ICA
Solna, SwedenVi söker fler stjärnor till vårt team på ICA Maxi Solna!
Deltid Ansök tidigt Publicerad för 4 dagar sedan
Stockholms stad
Stockholm, SwedenSamordnare i sociala system, Järva stadsdelsförvaltning
Heltid Ansök tidigt Publicerad för 4 dagar sedan
Stockholms stad
Bromma, Sweden +1 flerNorra Ängby skola söker vikarierande lågstadielärare
Tillfällig Ansök tidigt Publicerad för 4 dagar sedan
Stockholms stad
Stockholm, SwedenSocialsekreterare, Coping Power Progran (CPP)
Tillfällig Ansök tidigt Publicerad för 4 dagar sedan
Stockholms stad
Bromma, SwedenNorra Ängby skola söker lågstadielärare
Heltid Ansök tidigt Publicerad för 4 dagar sedan