MonoSparse-CAM: Efficient Tree Model Processing via Monotonicity and Sparsity in CAMs

2024-07-12

AI Summary

Tree models deployed on content-addressable memory (CAM) hardware suffer from low energy efficiency due to the lack of circuit-level exploitation of their structural properties. Method: This paper proposes a structure-aware hardware-software co-optimization framework that jointly exploits monotonicity and sparsity inherent in tree models. It enables decision-path pruning, monotonic feature grouping, and sparse rule encoding within CAM, establishing both compression and skip mechanisms. Crucially, the design explicitly maps tree model structure onto circuit-level CAM architecture to maximize hardware resource utilization. Results: Experiments demonstrate a 28.56× energy reduction over baseline CPU execution and an 18.51× improvement over the state-of-the-art CAM accelerator, alongside ≥1.68× higher throughput. This work establishes a new paradigm for energy-efficient hardware deployment of interpretable machine learning models.

Technology Category

• Machine Learning: Hardware-aware ML
• Computer Vision: Learning & Optimization for CV
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

While the tree-based machine learning (TBML) models exhibit superior performance compared to neural networks on tabular data and hold promise for energy-efficient acceleration using aCAM arrays, their ideal deployment on hardware with explicit exploitation of TBML structure and aCAM circuitry remains a challenging task. In this work, we present MonoSparse-CAM, a new CAM-based optimization technique that exploits TBML sparsity and monotonicity in CAM circuitry to further advance processing performance. Our results indicate that MonoSparse-CAM reduces energy consumption by upto to 28.56x compared to raw processing and by 18.51x compared to state-of-the-art techniques, while improving the efficiency of computation by at least 1.68x.

Problem

• Efficient Hardware Implementation
• Tree-based Machine Learning Models
• Energy-saving and Data Processing Acceleration

Innovation

• MonoSparse-CAM
• Energy Efficiency
• Computational Speedup