Output-Level Regularization Eliminates the Seed Lottery in Single-GPU VLA Fine-Tuning

The paper frames the work as vlaModern Robot LearningVision-Language-Action model (VLA)A model that takes images and language as input and outputs robot actions.. Start here because it defines what success means and which assumptions the rest of the method inherits.

Fine-tuningModern Robot LearningFine-tuningTaking a pretrained model and adapting it to a specific robot or task. vision-language-actionModern Robot LearningVision-Language-Action model (VLA)A model that takes images and language as input and outputs robot actions. models on a single GPU is unreliable—the same code produces 91-94% success 12 times, then silently fails to 65% on random seed 13. This paper shows output-level regularizationData, Distributions & Training IssuesRegularizationMethods used to reduce overfitting. (VICReg, dropout, or halved learning rate) eliminates this catastrophic failure modeData, Distributions & Training IssuesFailure modeA common way the system breaks or gets the task wrong. entirely, letting you reliably deploy VLAs without seed roulette. When reading the method section, identify the inputs, the learned or engineered representation, and the actionCore ConceptsActionA command the robot sends to its motors, controller, or low-level system. or prediction produced by the system.

For robotics work, the data story is part of the method: check whether the system depends on teleoperationImitation & Reinforcement LearningTeleoperation (teleop)A human remotely controlling the robot, often to collect demonstrations., simulationSimulation & Sim-to-RealSimulationA virtual environment where robots can be trained or tested., internet video, human labels, or robotCore ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. rollouts.

The paper should be judged through its evaluationSimulation & Sim-to-RealEvaluationMeasuring how well a robot system performs. protocol: what data is used, what robotCore ConceptsRobotA physical system with sensors and actuators that can observe the world and take actions. or simulator is tested, and which baselineEvaluation & ResearchBaselineA reference method used for comparison. comparisons support the claim. Look for the gap between the headline result and the deploymentSimulation & Sim-to-RealDeploymentPutting the trained system on a real robot. setting you would actually care about.