technique

Graph Neural Network Encoder

techniqueactivegraph-neural-network-encoder-e5fab569·1 events·first seen 25d ago

Aliases: Graph Neural Network Encoder

Co-occurring entities

large language models BERT-F1 Baiyu Chen AnyMo Inertial Measurement Unit (IMU)Human Activity Recognition (HAR)

More like this (12)

Graph Neural Networks Understanding Truncated Positional Encodings for Graph Neural Networks Sparse Autoencoders Feature Auto-Encoder NNCF (Neural Network Compression Framework)Influcoder: Distilling Decoders' Gradient Influence Rankings into an Encoder for Data Attribution encoder-only language models Graph Neural Network leveraging Higher-order Class Label Connectivity for Heterophilous Graphs VAE Encoder Neural Operator Graph Transformer transformer-based neural renderer

Recent events (1)

5arXiv · cs.CL·25d ago·source ↗

AnyMo: Geometry-Aware Setup-Agnostic Framework for Wearable IMU Human Motion Understanding

AnyMo is a geometry-aware framework that addresses the setup-dependence problem in wearable IMU-based human motion modeling by using physics-grounded simulation over dense body-surface placements to generate synthetic training signals. It pre-trains a graph encoder from synthetic placement views and masked partial observations, then tokenizes multi-position IMU data into full-body motion tokens aligned with an LLM for motion-language understanding. Evaluated across zero-shot activity recognition (14 unseen datasets), cross-modal retrieval, and motion captioning, AnyMo improves average Accuracy/F1 by ~11.7%/11.6%, zero-shot retrieval MRR by 15.9–28.6%, and captioning BERT-F1 by 18.8%. The work positions itself as a generalist model for wearable motion understanding transferable across devices and sensing configurations.

Agent and Tool Ecosystem Multimodal Progress large language models BERT-F1 Baiyu Chen +4 more