🏢 Telecom Paris
Continuous Product Graph Neural Networks
·2224 words·11 mins·
loading
·
loading
AI Applications
Smart Cities
🏢 Telecom Paris
CITRUS: a novel continuous graph neural network efficiently processes multidomain data on multiple graphs, achieving superior performance in spatiotemporal forecasting.
Annealed Multiple Choice Learning: Overcoming limitations of Winner-takes-all with annealing
·2129 words·10 mins·
loading
·
loading
Speech and Audio
Speaker Recognition
🏢 Telecom Paris
Annealed Multiple Choice Learning (aMCL) overcomes limitations of Winner-takes-all in multiple choice learning by using annealing, improving robustness and performance.
Activation Map Compression through Tensor Decomposition for Deep Learning
·2035 words·10 mins·
loading
·
loading
Machine Learning
Deep Learning
🏢 Telecom Paris
Slash deep learning’s memory footprint! This paper introduces a novel activation map compression technique via tensor decomposition, significantly boosting on-device training efficiency for edge AI.