Menü
Matteo Ottaviani

Dr. Matteo Ottaviani

Abteilung Forschungssystem und Wissenschaftsdynamik
wissenschaftlicher Mitarbeiter
  • 030 2064177-46
  • 030 2064177-99
  • Orcid
Projekte

Liste der Projekte

Leider konnte für diese Suchkombination kein Ergebnis gefunden werden
Kompetenznetzwerk Bibliometrie
Publikationen

Liste der Publikationen

Leider konnte für diese Suchkombination kein Ergebnis gefunden werden

On the performativity of SDG classifications in large bibliometric databases.

Ottaviani, M., & Stahlschmidt, S. (2024).
On the performativity of SDG classifications in large bibliometric databases. ArXiv (online first). https://doi.org/10.48550/arXiv.2405.03007
Abstract

This work proposes using the feature of large language models (LLMs) to learn about the "data bias" injected by diverse SDG classifications into bibliometric data by exploring five SDGs. We build a LLM that is fine-tuned in parallel by the diverse SDG classifications inscribed into the databases' SDG classifications. Our results show high sensitivity in model architecture, classified publications, fine-tuning process and natural language generation. The wide arbitrariness at different levels raises concerns about using LLM in research practice.

How A/B testing changes the dynamics of information spreading on a social network.

Ottaviani, M., Herzog, S., Nickl, P. L., & Lorenz-Spreen, P. (2024).
How A/B testing changes the dynamics of information spreading on a social network. ArXiv (online first). https://doi.org/10.48550/arXiv.2405.01165

Market structure or agent rationality: How efficiency trades with belief updating?

Giachini, D., Mousavi, S., & Ottaviani, M. (2024).
Market structure or agent rationality: How efficiency trades with belief updating? Research Square. https://doi.org/10.21203/rs.3.rs-3963671/v1
Abstract

For the Bayesian learner, updating priors based on observation is the path that improves the beliefs and this, in turn, affects the efficiency of prices. Or so was believed until agents with no learning advantage, that is zero-intelligence, produced near-optimal levels of efficiency. We design a market environment characterized by model misspecification in which agents update their initial probabilistic guesses with increasing sophistication in incorporating observations. In this designed environment, the market penalizes divergence from the truth, measured in relative entropy. We show that a non-linear U-shaped relation brings together individual rationality in learning and quality of pricing (i.e. informative efficiency).

Modern Tools for Agent-Based Model Sensitivity Analysis.

Ottaviani, M. (2024).
Modern Tools for Agent-Based Model Sensitivity Analysis. Social Science Research Network. https://dx.doi.org/10.2139/ssrn.4689174
Abstract

Agent-based models successfully allow to computationally set up socio-economic environments populated by heterogeneous agents interacting with one another and their environment based on internal behaviours, social norms, and data learned from past observations. Nowadays, the large availability of data on many dimensions considerably enriches the dynamic description. However, parameter space exploration, calibration, and sensitivity analysis (SA) have become computationally prohibitive. I propose a novel, general, multi-purpose, and trustworthy procedure (i.e., Monte Carlo-Once-At-a-Time); either it is an advanced local SA method itself or it serves as a reliability check before building the surrogate of an ABM for performing global SA.

Publikationen aus DFG-geförderten Projekten.

Meier, A., Mittermaier, B., Möller, T., Ottaviani, M., Scheidt, B., & Stahlschmidt, S. (2023).
Publikationen aus DFG-geförderten Projekten. Praxis und Nutzbarkeit von Funding Acknowledgements. Bonn: DFG.
Abstract

Die vorliegende Studie entstand im Auftrag der Deutschen Forschungsgemeinschaft (DFG) und wurde durch das Deutsche Zentrum für Hochschul- und Wissenschaftsforschung (DZHW) und das Forschungszentrum Jülich (FZJ) gemeinschaftlich durchgeführt. Ausgehend von den DFG-Vorgaben wurde der Indikator Funding-Acknowledgement-Güte (FA-Güte) entwickelt. Anhand von drei Güteklassen (regelkonforme Nennung, nicht regelkonforme Nennung und keine Nennung) misst er für jede Publikation, inwiefern die Geförderten den DFG-Vorgaben nachgekommen sind.

Market selection and learning under model misspecification.

Bottazzi, G., Giachini, D., & Ottaviani, M. (2023).
Market selection and learning under model misspecification. Journal of Economic Dynamics and Control2023. https://www.sciencedirect.com/science/article/pii/S0165188923001458 (Abgerufen am: 12.09.2023).
Abstract

The paper explores market selection in an Arrow-Debreu economy with complete markets where agents learn over misspecified models. Standard Bayesian learning loses formal justification, and biased learning processes may provide a selection advantage. The ecology of traders in the market affects selection dynamics and long-run asset valuation. Model misspecification makes it difficult to rank learning behaviors based on survival prospects. Prediction averaging has an advantage when the true data generating process belongs to the same family of models, but this advantage disappears when the true model belongs to a more general class. Rules guaranteeing survival exploit imitative mechanisms.

Vorträge & Tagungen

Liste der Vorträge & Tagungen

Leider konnte für diese Suchkombination kein Ergebnis gefunden werden

Enabling Reliable Sensitivity Analysis for Agent-Based Modeling in Economics and Finance.

Ottaviani, M. (2023, Oktober).
Enabling Reliable Sensitivity Analysis for Agent-Based Modeling in Economics and Finance. Vortrag im Rahmen der Conference on Complex Systems 2023 (CCS2023), Salvador, Brazil.
Abstract

Agent-based models (ABM) allow for computationally setting up socio-economic environments based on internal behaviors and data. However, parameter space exploration, calibration, and sensitivity analysis become computationally prohibitive. A novel procedure, Monte Carlo-Once-At-a-Time, is proposed for building a cheaper proxy for reliable sensitivity analysis. This approach allows for the creation of trustworthy proxies, saving minimal computational effort.

How A/B testing changes the dynamics of information spreading on a social network.

Ottaviani, M. (2023, Oktober).
How A/B testing changes the dynamics of information spreading on a social network. Vortrag im Rahmen der Conference on Complex Systems 2023 (CCS2023), Salvador, Brazil.
Abstract

This paper explores the impact of A/B testing on user decision-making and collective behavior. It uses a case study dataset of news headlines to analyze linguistic features that attract clicks and amplify them through A/B testing. The study reveals that A/B testing increases the homogeneity of information spread, reducing exploration and amplifying exploitation of successful features. This could help develop intervention and tools to promote user autonomy and reduce unwanted outcomes.

On the performativity of SDG classifications in large bibliometric databases.

Ottaviani, M., & Stahlschmidt, S. (2023, Oktober).
On the performativity of SDG classifications in large bibliometric databases. Poster auf der Konferenz Nordic Workshop on Bibliometrics and Research Policy, Gothenburg, Sweden.
Abstract

Large bibliometric databases, such as Web of Science, Scopus and Dimensions, are performative and affect the visibility of scientific outputs and impact measurement. However, these databases do not match the UN's SDGs classifications. This work proposes using large language models (LLM) to learn about data bias injected by diverse SDG classifications. The results show high sensitivity to diverse classifications in model architecture, classified publications, fine-tuning process and natural language generation. The wide arbitrariness raises ethical concerns. Qualitative text analyses can help inform science policy on diverse SDG classifications.

Entwicklung eines Indikators zur Güte der DFG-Funding Acknowledgments.

Ottaviani, M., & Möller, T. (2022, Dezember).
Entwicklung eines Indikators zur Güte der DFG-Funding Acknowledgments. Vortrag im Rahmen des Arbeitstreffens des Projekts zur Praxis und Nutzung von DFG-Funding Acknowledgements, Deutsche Forschungsgemeinschaft (DFG).

Zwischenergebnisse zur Praxis und Nutzbarkeit von Funding Acknowledgements für die Zuordnung von Publikationen zu DFG-Projekten.

Scheidt, B., Möller, T., Ottaviani, M., Meyer, A., & Mittermaier, B. (2022, August).
Zwischenergebnisse zur Praxis und Nutzbarkeit von Funding Acknowledgements für die Zuordnung von Publikationen zu DFG-Projekten. Vortrag im Rahmen des des Zwischen-Workshops, Deutsche Forschungsgemeinschaft.