The Optimization and Artificial Intelligence in Latin America workshop brings together researchers and practitioners interested in the design, analysis, and application of optimization and machine learning with a special focus on the region’s challenges, data, and use cases. While both fields search for solutions and models, their internal dynamics—structural biases, behavioral transitions, exploration–exploitation, robustness, any-time performance, and complexity—remain insufficiently understood due to the nature of heuristic processes and experimental practices that often emphasize final outcomes only. We therefore welcome theoretical and empirical contributions that instrument executions—online or offline—to model population dynamics, identify desirable algorithmic traits, detect failure modes, and improve interpretability.
The program will combine technical presentations and two interactive sessions:
Panel: Challenges and opportunities in Latin America for research and development.
We value reproducibility (artifacts, scripts, baselines), transparency in reporting configurations, and honest discussion of limitations. Our goal is to map the current state of the art at the intersection of AI and optimization from and for Latin America, build bridges between regional groups and the international GECCO community, and foster sustained collaborations that accelerate scientific progress and its social and economic impact throughout the region.
This workshop is strategically aligned with the 7th IEEE International Conference on BioInspired Processing (BIP 2025), to be held in Costa Rica (UNA, Pérez Zeledón) on 3–5 December 2025. The BIP audience—bioinspired processing, nature-inspired computation, and AI applications—naturally converges with GECCO’s pillars (evolutionary computation, AI-driven optimization, evolutionary machine learning). We will leverage BIP to (i) announce this workshop’s CFP, (ii) invite authors with preliminary results to submit extended abstracts or full papers, and (iii) promote invited talks and a joint panel.
The temporal continuity with GECCO 2026 in San José (13–17 July 2026) supports maturation and deeper discussion of contributions and broadens participation through the local community. Additionally, the organizing committee will include co-chairs with prior workshop experience and active BIP committee members, ensuring the rigor of the review process.
Emerging applications in LATAM: Case studies in public health and hospitals, energy and smart grids, agribusiness, logistics and transportation, inclusive finance, smart cities, and environmental management; approaches robust to scarce/noisy data and domain shift.
Understanding, traceability, and benchmarking of AI–optimization pipelines: Analysis of internal dynamics and structural biases; exploration–exploitation balance and any-time measures; reproducibility (artifacts, scripts, baselines); benchmark suites with regional data; and transparent evaluation practices.
Safe AI, calibration, and uncertainty estimation (UQ): Methods to estimate, calibrate, and communicate model uncertainty, detect failures and out-of-distribution data, and quantify risk for safer decision-making.