Institute for
Inverse Problem
Research

From observations to models.

An independent virtual institute connecting researchers working on inverse problems across mathematics, statistics, scientific machine learning and AI.

The inversion act Observations — data, signals and measurements — enter an inference engine of algorithms, models and prior knowledge, which returns models, parameters, and hidden states and predictions. OBSERVATIONS MODELS & PREDICTIONS Data, signals, measurements INFERENCE ENGINE Algorithms, models, prior knowledge Models Parameters Hidden states & predictions

Parameter inference

Estimating unknown quantities from observed data.

Model identification

Selecting the most plausible model class.

Hybrid inversion

Combining mechanistic, statistical, and neural models.

Neuro-symbolic inversion for zooplankton systems

Integrating ecological mechanisms and data to recover hidden dynamics.

LLM-RAG assisted calibration of large simulation software

Natural language interfaces and retrieval to guide calibration and analysis.

Research community

A small founding group supported by a network of associated researchers, in a distributed virtual institute.

2Founders
4Associated
2Active projects
Open Library →