Research & insights

The work behind
mission-ready AI.

Explore Yrikka’s work in synthetic data, model adaptation, adversarial robustness, and world models.

Research in practice

Methods and demonstrations for the next mission.

Sensor adaptation

Expand the data available for a new sensor.

Explore RGB-to-infrared and RGB-to-SAR generation, with the research behind downstream model adaptation.

Explore the workflow

Synthetic infrared video

Extend scenario coverage through time.

Explore six counter-UAS scenes with aligned RGB and thermal video, plus interactive comparison controls.

Explore the counter-UAS demos

ATLAS mission simulation

Put autonomy into the operating environment.

Explore physics-based mission worlds, configurable conditions, and integration with your autonomy stack.

Explore ATLAS

Papers & technical writing

Original publications from the Yrikka team.

Research paper

10 November 2025

Inference-Time Scaling of Diffusion Models for Infrared Data Generation

Domain-adapted generation and verifier-guided sampling for infrared imagery. The study examines how additional inference-time computation can improve the quality of synthetic IR data.

Kai A. Horstmann, Maxim Clouser & Kia Khezeli

Research paper

3 September 2025 · Updated March 2026

CausalARC: Abstract Reasoning with Causal World Models

A testbed for learning causal world models from limited observations, interventions, and counterfactual examples—and testing whether that understanding generalizes.

Jacqueline Maasch, John Kalantari & Kia Khezeli