Learnable knowledge transfer improves robotic vision efficiency
Can Knowledge Transfer Parameters Be Learned? LePoKet for Efficient Robotic Vision
Computer Vision and Pattern RecognitionMachine Learning
Summary
Robots need to see and understand their surroundings quickly but often don’t have powerful computers to do this. The authors found a new way to help small robot vision systems learn better from big, smart models by letting the system figure out how to pass on knowledge. This method, called LePoKet, works better than older ways on image recognition and motion detection tasks. It means robots can see more accurately and efficiently without needing extra computing power.
What this means in practice
- •For robotics engineers: Integrate LePoKet to improve smaller robot vision systems’ accuracy and efficiency under hardware constraints without adding extra training losses.
- •For computer vision developers: Use the learnable parameter transfer method to enhance compact optical-flow models for better motion estimation on datasets like Sintel and KITTI.
Authors
Yanick C. Tchenko, Felix Mohr, Hicham Hadj-Abdelkader, Hedi Tabia
Abstract
Efficient perception is central to robotic systems operating under constrained computation, memory, and latency budgets. Knowledge transfer from larger pretrained models offers a practical route to stronger compact perception networks, but existing approaches commonly rely on fixed distillation objectives or manually designed interaction mechanisms. Building on Hereditary Knowledge Transfer (HKT), we propose LePoKet (Learnable Parameter Optimization for Knowledge Transfer), a structural transfer framework that embeds knowledge inheritance directly into the forward computation. LePoKet introduces a block-wise Extract-Transform-Mix interface whose interaction parameters are optimized jointly with the child network through a Learnable Genetic Attention (LGA) operator, without auxiliary distillation losses or temperature scaling. We first characterize the mechanism on CIFAR-10 and CIFAR-100 using ResNet parent-child pairs, obtaining relative error reductions of 24.57% and 25.1%, respectively, over standard child training. We then evaluate LePoKet for dense motion estimation by integrating it into a compact RAFT-based optical-flow model trained only on FlyingChairs and FlyingThings3D. LePoKet improves the compact RAFT baseline from 2.21 to 1.92 EPE on Sintel Clean, from 3.35 to 3.01 on Sintel Final, and from 7.51 to 6.39 on KITTI. A direct comparison with HKT further shows that LePoKet improves CIFAR-10 accuracy from 92.40% to 93.40% while achieving the best Sintel Final and KITTI errors among the evaluated compact transfer variants, with comparable performance on Sintel Clean. These results demonstrate that learnable structural transfer generalizes across recognition and motion perception tasks and provides a promising approach for efficient robotic vision.