Staged Hybridisation for Visual Quantum Reinforcement Learning via Knowledge Distillation
2026-06-29 • Machine Learning
Machine Learning
AI summaryⓘ
The authors explore an easier way to train quantum reinforcement learning agents that see images. Instead of teaching the quantum system directly from raw pixels, they first train a regular classical model and then copy its knowledge into smaller quantum or classical parts using a technique called knowledge distillation. This approach helps the quantum parts learn visual control tasks more reliably than training from scratch. They tested this method on game-like tasks, showing that quantum components can perform almost as well as the classical teacher when trained this way.
Quantum Reinforcement LearningKnowledge DistillationVariational Quantum CircuitsVisual ControlReinforcement LearningFeature EncodingCartPoleAcrobotAngle EncodingAmplitude Encoding
Authors
Javier Lazaro, Juan-Ignacio Vazquez, Pablo Garcia-Bringas
Abstract
Visual environments are a demanding setting for quantum reinforcement learning (QRL): high-dimensional observations, unstable RL optimisation, and constrained variational quantum circuits (VQCs) are difficult to train jointly. This paper studies knowledge distillation (KD) as a staged hybridisation strategy for visual QRL. Instead of training a hybrid visual agent end-to-end from pixels, we first train a classical visual teacher, freeze its encoder as a feature interface, and distil the teacher's policy behaviour into compact downstream heads. These heads can be classical or VQC-based, enabling small quantum-compatible students to be evaluated under the same frozen representation as compact classical controls. We evaluate the pipeline on CartPole Pixels and Acrobot Pixels. The results show that staged KD enables shallow VQC heads to acquire non-trivial visual-control behaviour in settings where direct pixel-based training would be substantially more difficult. Angle-encoded VQC heads retain near-teacher performance, while amplitude-encoded heads push compactness to an extreme regime, at the cost of greater fragility, stronger budget sensitivity, and higher simulation time. Overall, staged KD reframes visual QRL as a compact-head learning problem, opening a practical route for training small quantum-compatible policies outside the standard end-to-end RL loop.