All updates
CompanyOct 2026

eomer keynote at Intelligence X 2026 @NUS

eomer presented a keynote on applied AI research and enterprise use cases at IntelligenceX 2026 at the National University of Singapore. The academic conference took place from 24 to 26 September at NUS University Town. NUS and the Society of Algorithmic Intelligence hosted the conference, which addressed AI, quantum technologies, and practical applications.

The eomer presentation connected research methods with the requirements of applied forecasts. A model must be assessed against the data that would have been available at the time of a decision. Forecast horizons, information cutoffs, and evaluation periods must remain consistent when researchers compare methods.

One part of the presentation addressed explicit reference models. These included persistence, seasonal naïve forecasts, climatology, and established time-series methods. A matched holdout allows researchers to determine whether a more complex model reduces forecast error relative to those reference models under the same conditions.

For probabilistic forecasts, the presentation used the continuous ranked probability score (CRPS) to assess the forecast distribution. A skill score compares CRPS with the score of a reference model on the same forecast cases. Lower CRPS indicates lower distributional error; a skill score then quantifies the change relative to the selected reference model.

The applied AI use cases connected this research perspective with enterprise decisions. Forecast error provides one criterion, while operational constraints and the cost of an action provide additional criteria. The presentation placed these questions within one evaluation process: define the decision, preserve the available information, compare reference models, and assess the proposed action with the same historical evidence.