Kangaroo: A Private and Amortized Inference Framework over WAN for Large-Scale Decision Tree Evaluation

2025-09-03

AI Summary

Existing privacy-preserving decision tree evaluation (PDTE) schemes suffer from rapidly escalating communication and computational overhead in wide-area networks (WANs) as model scale—measured by number of trees, nodes, and depth—increases. This work proposes the first efficient private inference framework for large-scale random forests, built upon packed homomorphic encryption (PHE). It integrates model hiding with adaptive encoding, secure feature selection, oblivious comparison, and path-evaluation protocols, augmented by latency-aware scheduling and same-modulus sharing optimization. Crucially, this design achieves amortized overhead scaling sublinearly with both tree count and node count—the first such result. Experiments on a WAN demonstrate inference over a random forest comprising 969 trees and 411,825 nodes, with average per-tree latency of ∼60 ms—3× to 59× faster than state-of-the-art methods—thereby substantially improving scalability and practical deployability.

Technology Category

• Machine Learning: Privacy
• Search and Optimization: Distributed Search
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems

Application Category

• Security and Privacy: Privacy-enhancing technologies
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

With the rapid adoption of Models-as-a-Service, concerns about data and model privacy have become increasingly critical. To solve these problems, various privacy-preserving inference schemes have been proposed. In particular, due to the efficiency and interpretability of decision trees, private decision tree evaluation (PDTE) has garnered significant attention. However, existing PDTE schemes suffer from significant limitations: their communication and computation costs scale with the number of trees, the number of nodes, or the tree depth, which makes them inefficient for large-scale models, especially over WAN networks. To address these issues, we propose Kangaroo, a private and amortized decision tree inference framework build upon packed homomorphic encryption. Specifically, we design a novel model hiding and encoding scheme, together with secure feature selection, oblivious comparison, and secure path evaluation protocols, enabling full amortization of the overhead as the number of nodes or trees scales. Furthermore, we enhance the performance and functionality of the framework through optimizations, including same-sharing-for-same-model, latency-aware, and adaptive encoding adjustment strategies. Kangaroo achieves a $14 imes$ to $59 imes$ performance improvement over state-of-the-art (SOTA) one-round interactive schemes in WAN environments. For large-scale decision tree inference tasks, it delivers a $3 imes$ to $44 imes$ speedup compared to existing schemes. Notably, Kangaroo enables the evaluation of a random forest with $969$ trees and $411825$ nodes in approximately $60$ ms per tree (amortized) under WAN environments.

Problem

• Private decision tree evaluation with high communication and computation costs
• Inefficient large-scale model inference over wide area networks
• Lack of amortization as number of nodes or trees increases

Innovation

• Packed homomorphic encryption for private decision tree evaluation
• Amortized inference framework reducing WAN communication overhead
• Secure protocols for model hiding and oblivious comparison