Szymon Miłkoś
Not only academic website
I inquire into
scientific discovery
to help others innovate
Here will be a nice photo, soon
Research Questions
How do breakthroughs and innovations really happen? How do scientists create true theories, and how do we – as individuals, organizations and societies – solve our most complex problems? How can we study and improve scientific discovery practice itself?
My Approach
As a metascientist and methodologist, I take a transdisciplinary approach, building on foundations in philosophy, cognitive science, and public management. I bridge academic research, societal issues and business implementation, having led project reaching millions and collaborated with organizations on innovative processes and AI products.
Impact
My work provides frameworks, tools, and initiatives for studying and improving scientific discovery practice. The world urgently needs better science, and improved discovery methods can help us address vital challenges in climate, technology, healthcare, and beyond.
The future of science lies not just in what we discover, but how we do it, and why
Science advances humanity, yet we have limited understanding of how scientific breakthroughs actually occur. By studying and improving the discovery process itself, we can create a reinforcing loop: better understanding leads to improved practices, which enable more effective research methods, accelerating further insights into discovery. This virtuous cycle, combined with the digital transformation of scientific practice, has the potential to dramatically enhance humanity's capabilities to address our most significant challenges.
Key Directions
- Advance the theory and methodology of scientific discovery
- Create tools for studying and empowering breakthroughs
- Cultivate a community dedicated to the science of discovery
- Apply these insights to solve humanity's most pressing challenges
Future Projects
Scientists for Science
A pioneering cohort of researchers who volunteer to have their discovery processes studied broadly, deeply, and longitudinally. This project aims to create a first-of-its-kind database of scientific discovery processes in action.(initial conceptualisation, seeking collaborators)
Discovery Platform
A digital environment where scientists conduct research while simultaneously receiving tools to enhance their discovery process. The platform provides real-time support while generating valuable data on discovery patterns, creating a virtuous cycle of improvement. This platform would integrate AI assistance with metascience insights to accelerate breakthrough thinking.(advanced conceptualisation, seeking collaborators)
This is an ambitious vision, and collaboration is key
Join the future of scientific discovery.
Current Work:
I am completing my PhD on scientific practice with causal discovery algorithms under the supervision of Marcin Miłkowski at the Polish Academy of Sciences. I collaborate with a pioneering metascientific group of psychologists of science.
After 10 years of research, I have developed a new theory that explains what happens at the forefront of science, when we don't know what to do. The methodology of meta-creative problem-solving can help understand — and guide — how discovery practices and norms originate and evolve at the edge of knowledge. If you're facing a bottleneck problem at the frontier of discovery or innovation, let's talk how this approach might help.
Current Projects (7 of 7)
PhD Project
Advancing the methodology of scientific discovery through an integrated study of Causal Discovery Methods, Meta-Creative Problem-Solving, and Cognitive Metascience.
This dissertation aims to understand and advance the methodology of scientific discovery. It uses Causal Discovery Algorithms as a lens to study and improve scientific practice, introduces Meta-Creative Problem-Solving as a novel theory of how scientists overcome ignorance at the frontiers of discovery, and pioneers a cognitive metascience investigation to improve theoretical practices in science. The three main parts of this work are being developed as standalone papers (see related projects).
Keywords:
Widespread Adoption of Causal Discovery Algorithms Through Theoretical Practice Checklist
This study aims for an impactful understanding of scientific discovery practice using Causal Discovery Algorithms (CDAs). Despite their potential, widespread CDA adoption is hindered by conceptual confusion and problems merging domain-specific theories with the causal framework, leading to uncertain trustworthiness of results. This paper develops a fundamental causal assumption to provide external validity criteria and a central methodological definition to synthesize debates on CDA outcome validity. Building on this, a 'Theoretical Practice Checklist' is proposed. This checklist allows for estimating confidence in CDA results, enriching peer-review, and fostering the iterative integration of theory-driven scientists with data-driven methods.
Keywords:
Overcoming Ignorance: Meta-Creative Problem-Solving as Theory and Methodology of Scientific Discovery
Introducing a new approach to study and improve discovery practice and norms practices as they origin and develop.
Cognitive Metascience: Studying and Improving Scientific Theoretical Practices
This project introduces a Cognitive Metascience approach to systematically study and improve theoretical practices within scientific research, with an focus on Causal Discovery Algorithms applications. It leverages insights from Meta-Creative Problem-Solving to develop a neurosymbolic workflow for assessing the explicitness of theoretical traces, carefulness of practice, and quality of theoretical grounding. The aim is to provide validated guidelines and tools that enhance the quality, robustness, and impact of scientific endeavors by making theoretical work more transparent and effective. All modules for the workflow tool have been coded and are currently being assembled.
Keywords:
Causal Mapping App
A collaborative tool for creating prior causal graphs, essential for applying formal causal frameworks
This application facilitates the collaborative construction of initial causal graphs. Users can define variables and invite participants to add suspected causal relationships, specifying direction, strength, polarity, and their reasoning. The app also prompts users to assess their confidence in the absence of links they didn't specify. Finally, individual inputs are aggregated to generate a summary graph, highlighting areas of consensus and disagreement among researchers. This tool aims to streamline the crucial first step in formal causal modeling and discovery. If this tool could be useful for your current needs, please get in touch.
Keywords:
Burnout in Science: Causes and Intervention
This project, led by Dominik Dianovics, aims to understand the prevalence and multifaceted nature of burnout in the scientific community. A pilot survey study has been conducted, and we are currently developing the preregistration for the main study. My role involves providing methodological support and conducting causal analysis
Keywords:
The Emotional Landscape of Scientific Practice
Led by Balazs Aczel, this study explores emotions and their causes in the research practice. I helped in conceptualizing the project, designing survey questions, and will be conducting natural language processing and causal analysis on the collected data. A pilot study has been completed, and we are now moving into the analysis and discussion phase to refine our approach for a larger study.
Keywords:
