The ENTC Chair is open to supervising semester projects by EPFL students in entrepreneurship, innovation and strategic management. In addition to the availabilities listed below, we are engaged in several initiatives that leverage machine learning and natural language processing to advance management science. You are welcome to reach out to discuss some of these opportunities or to propose your own topic.
#1. Tracking Novel Research Throughout the Publication Process: A Text-Embedding Approach
This semester project is part of an ongoing collaboration with the International Center for the Study of Research (ICSR Lab), which provides access to anonymized manuscript metadata and peer review outcomes from more than 2,000 Elsevier journals [1]. The goal is to investigate scientific conservatism in the peer review process—i.e., systematic tendencies to reject manuscripts that address novel topics, despite their societal importance and potential for long-term impact.
Objectives
- Review recent advances in text embeddings and SOTA models for semantic text similarity tasks.
- Detect the novelty of submitted manuscripts based on anonymized manuscript metadata.
- Develop a framework to track rejected papers by matching them with their eventual publications [2].
Requirements
- Familiarity with natural language processing libraries and text embedding techniques
- An interest in the “backstage” of science, and how research is evaluated and published
Details
- Open to current EPFL master students
- Possibility of supervising multiple students
- For inquiries, contact Davide Bavato (davide.bavato@epfl.ch)
[1] Bavato, D., Ayoubi, C., Gruber, M., Lee, C., & Buchanan, N. T. (2023). Study Protocol: The Role of Disagreement in Scientific Conservatism. Available at SSRN 4569834.
[2] Tkaczyk, D. (2023, December 07). Discovering Relationships Between Preprints and Journal Articles. https://doi.org/10.64000/dpcc9-k4564