Abstract:Graph Neural Networks (GNNs) are becoming increasingly popular for vision-based applications due to their intrinsic capacity in modeling structural and contextual relations between various parts of an image frame. On another front, the rising popularity of deep vision-based applications at the edge has been facilitated by the recent advancements in heterogeneous multi-processor Systems on Chips (MPSoCs) that enable inference under real-time, stringent execution requirements. By extension, GNNs employed for vision-based applications must adhere to the same execution requirements. Yet contrary to typical deep neural networks, the irregular flow of graph learning operations poses a challenge to running GNNs on such heterogeneous MPSoC platforms. In this paper, we propose a novel unified design-mapping approach for efficient processing of vision GNN workloads on heterogeneous MPSoC platforms. Particularly, we develop MaGNAS, a mapping-aware Graph Neural Architecture Search framework. MaGNAS proposes a GNN architectural design space coupled with prospective mapping options on a heterogeneous SoC to identify model architectures that maximize on-device resource efficiency. To achieve this, MaGNAS employs a two-tier evolutionary search to identify optimal GNNs and mapping pairings that yield the best performance trade-offs. Through designing a supernet derived from the recent Vision GNN (ViG) architecture, we conducted experiments on four (04) state-of-the-art vision datasets using both ( i ) a real hardware SoC platform (NVIDIA Xavier AGX) and ( ii ) a performance/cost model simulator for DNN accelerators. Our experimental results demonstrate that MaGNAS is able to provide 1.57 × latency speedup and is 3.38 × more energy-efficient for several vision datasets executed on the Xavier MPSoC vs. the GPU-only deployment while sustaining an average 0.11% accuracy reduction from the baseline.

GAME: Gaussian Mixture Model Mapping and Navigation Engine on Embedded FPGA

GPU Accelerated Real-Time Traversability Mapping.

A Fully-Pipelined Hardware Design for Gaussian Mixture Models

GMMap: Memory-Efficient Continuous Occupancy Map Using Gaussian Mixture Model

A Fully-Pipelined Expectation-Maximization Engine For Gaussian Mixture Models

MR-GMMapping: Communication Efficient Multi-Robot Mapping System via Gaussian Mixture Model

An FPGA-based Real-Time Simultaneous Localization and Mapping System.

Implementation of real-time Gaussian mixture models based on FPGA

MaGNAS: A Mapping-Aware Graph Neural Architecture Search Framework for Heterogeneous MPSoC Deployment

Efficient video object segmentation based on Gaussian mixture model and Markov random field

Trajectory-Based Mobile Game Bots Detection with Gaussian Mixture Model.

MR-GMMExplore: Multi-Robot Exploration System in Unknown Environments based on Gaussian Mixture Model.

Three-level GPU Accelerated Gaussian Mixture Model for Background Subtraction.

PGMF-VINS: Perpendicular-Based 3D Gaussian–Uniform Mixture Filter

ICGMM: CXL-enabled Memory Expansion with Intelligent Caching Using Gaussian Mixture Model

MM3DGS SLAM: Multi-modal 3D Gaussian Splatting for SLAM Using Vision, Depth, and Inertial Measurements

RGBD GS-ICP SLAM

MCMG Simulator: A Unified Simulation Framework for CPU and Graphic GPU

An FPGA Based Energy Efficient DS-SLAM Accelerator for Mobile Robots in Dynamic Environment

GEVO: Memory-Efficient Monocular Visual Odometry Using Gaussians