INT-27031 INTERNSHIP VISUAL ANALYSIS OF FEATURE INTERACTIONS IN TREE ENSEMBLES

The Luxembourg Institute of Science and Technology (LIST) is a Research and Technology Organization (RTO) active in the fields of materials, environment and IT. By transforming scientific knowledge into technologies, smart data and tools, LIST empowers citizens in their choices, public authorities in their decisions and businesses in their strategies.

Do you want to know more about LIST? Check our website: https://www.list.lu/

Your LIST benefits

• An organization with a passion for impact and strong RDI partnerships in Luxembourg and Europe that works on responsible and independent research projects
• Sustainable by design, empowering our belief that we play an essential role in paving the way to a green society
• Innovative infrastructures and exceptional labs occupying more than 5,000 square metres, including innovations in all that we do
• An environment encouraging curiosity, innovation and entrepreneurship in all areas
• Multicultural and international work environment with more than 50 nationalities represented in our workforce
• Diverse and inclusive work environment empowering our people to fulfil their personal and professional ambitions
• Gender-friendly environment with multiple actions to attract, develop and retain women in science
• 32 days’ paid annual leave, 11 public holidays

The monthly gross internship allowance is 1,400 euros for a full-time equivalent (40 hours per week).

Internship contract | Belval | 6 Months

Are you passionate about research? So are we! Come and join us

How will you contribute?

Interpreting random forest models is a critical challenge in applied machine learning. Recent work has shown that the structure of a random forest can be encoded as a feature graph - a network where nodes represent input variables and edges capture co-occurrence patterns along decision paths. While this representation is promising, no systematic study has examined how such graphs should be visualized to support the analytical tasks of data scientists and domain experts. This internship conducts a structured design study addressing this gap. Starting from a working implementation of the feature graph construction pipeline, the student will conduct a task analysis with domain experts to identify the key analytical questions these graphs must answer. Based on this analysis, the student will design, implement, and comparatively evaluate multiple visual representations—including node-link diagrams with centrality-based encodings, adjacency matrices, and layered representations by tree depth. A lightweight user study with data science practitioners will be conducted to assess the effectiveness and usability of the proposed designs. The results will directly inform the design of a full visual analytics system and will be submitted as a short paper to a top visualization venue (EuroVis or IEEE VIS).

You will be mainly in charge of:

• Task analysis methodology and structured design study protocol
• Implementation of a feature graph construction pipeline from scikit-learn Random Forest models
• Design and prototyping of interactive graph visualizations (D3.js or equivalent)
• Comparative evaluation of visual representations across datasets of varying dimensionality
• Lightweight user study design and qualitative analysis
• Scientific writing targeting a short paper in a top visualization conference

You will be closely mentored by, and work alongside, researchers active in the visualization (VIS) research community, giving you direct access to state-of-the-art knowledge and hands-on training in visual analytics and machine learning interpretability.

Is your profile described below? Are you our future trainee? Apply now!

Education

• Final year in Master of Science in Computer Science, Engineering, Data Science or equivalent.
• Proficient in a modern programming language such as Python, Java or similar.
• Proficient in Frontend development languages (Javascript, HTML, CSS).
• Familiarity with a data visualization library, D3js is a plus.
• Familiar with machine learning methods, especially Decision Trees or Random Forests.
• Ability to understand scientific and technical documents and reimplement their ideas.

Language skills

• Fluency in English (and French), both oral and written. Other relevant languages are an asset.