# Interactive Optimization and Learning

The IOL research lab is dedicated to exploring the intersection of mathematical optimization and machine learning, with a focus on developing innovative techniques for learning and optimization. By integrating these two fields, we aim to create new approaches to solving complex problems that leverage the strengths of both optimization and machine learning.

Our group is located both at the Mathematical Optimization research group at the Technische Universität Berlin (TUB) and in the AI in Society, Science, and Technology (AIS²T) department at the Zuse Institute Berlin (ZIB). We are also part of the Berlin mathematics research center MATH+ as well as the Berlin Mathematical School (BMS).

## Publication highlights

- Mundinger, K., Zimmer, M., and Pokutta, S. (2024).
*Neural Parameter Regression for Explicit Representations of PDE Solution Operators*. [arXiv]## [BibTeX]

- Pauls, J., Zimmer, M., Kelly, U. M., Schwartz, M., Saatchi, S., Ciais, P., Pokutta, S., Brandt, M., and Gieseke, F. (2024). Estimating Canopy Height at Scale.
*Proceedings of International Conference on Machine Learning*. [arXiv] [code]## [BibTeX]

- Mundinger, K., Pokutta, S., Spiegel, C., and Zimmer, M. (2024). Extending the Continuum of Six-Colorings.
*Geombinatorics Quarterly*,*XXXIV*. [URL] [arXiv]## [BibTeX]

- Parczyk, O., Pokutta, S., Spiegel, C., and Szabó, T. (2024). New Ramsey Multiplicity Bounds and Search Heuristics.
*Foundations of Computational Mathematics*. DOI: 10.1007/s10208-024-09675-6 [arXiv] [code]## [BibTeX]

- Braun, G., Carderera, A., Combettes, C., Hassani, H., Karbasi, A., Mokhtari, A., and Pokutta, S. (2022).
*Conditional Gradient Methods*. [arXiv]## [BibTeX]

## Software repositories

All our publically accessible software repositories are available on GitHub. We have a list of actively maintained repositories:

- FrankeWolfe.jl, a toolbox for Frank-Wolfe and conditional gradients algorithms
- Boscia.jl, a package for Branch-and-Bound on top of Frank-Wolfe methods
- BellPolytopes.jl, a package that addresses the membership problem for local polytopes

We are also actively involved in the development of the SCIP Optimization Suite at ZIB and its interfaces to other programming languages:

- SCIP, one of the fastest academically developed solvers for mixed integer programming (MIP) and mixed integer nonlinear programming (MINLP)
- SoPlex, an optimization package for solving linear programming problems.
- PaPILO parallel presolve routines for (mixed integer) linear programming problems
- PySCIPOpt, a Python interface for SCIP
- SCIP.jl, a Julia interface for SCIP
- JSCIPOpt, a Java interface for SCIP
- russcip, a Rust interface for SCIP