Contrastive learning improves intention prediction across environments
CARE: Contrastive Representation Learning with Environment-Invariance Regularization for Behavioural Intention Prediction
Machine Learning
Summary
Predicting what someone intends to do based on their movement paths is tricky because different surroundings can change how they act. The authors developed a method called CARE that helps computers learn from these movement patterns so they focus on behavior itself, not the specific environment. It does this by comparing parts of the data in ways that highlight important signals and avoid environment-specific details. Their approach led to better accuracy and reliability when tested in different places than it was trained on.
What this means in practice
- •For robotics engineers: Improve robot predictions of human intentions by using representations less affected by environment changes, enhancing safe interaction.
- •For autonomous vehicle developers: Enhance vehicle decision-making by better understanding pedestrian and driver intentions across different locations using environment-robust representations.
Authors
Shang Gao, Zhixin Huang, Franz Götz-Hahn, Konrad Doll, Bernhard Sick
Abstract
Trajectory-based behavioural intention prediction faces environmental variation, and representations learned with a pooled contrastive objective may retain environment-specific regularities. We present CARE, a self-supervised trajectory-learning framework that couples temporally asymmetric contrastive views with an invariant-risk-minimisation (IRM)-style gradient penalty on environment-wise contrastive risks. Each input window yields an anchor, a partially masked positive view, and a temporally perturbed negative view. Masked reconstruction and observability-aware evidence consistency complement this objective by promoting temporal information retention under partial input. On an IMPTC-derived four-class intention task, CARE achieves 0.840 accuracy and 0.023 expected calibration error (ECE) under IID fine-tuning, and 0.806 accuracy with a Brier score of 0.308 on the reported held-out environment. Under frozen-encoder linear probing, reconstruction provides the largest individual held-out accuracy gain, whereas the complete objective yields the lowest held-out ECE (0.049) among tested variants. These results support the combined objective as a competitive representation-learning approach under the evaluated shift, while the IRM-style regulariser is not taken as evidence of formal invariance.