Showing 1–2 of 2 results for author: A., P K
-
Graph Classification with GNNs: Optimisation, Representation and Inductive Bias
Authors:
P. Krishna Kumar a,
Harish G. Ramaswamy
Abstract:
Theoretical studies on the representation power of GNNs have been centered around understanding the equivalence of GNNs, using WL-Tests for detecting graph isomorphism. In this paper, we argue that such equivalence ignores the accompanying optimization issues and does not provide a holistic view of the GNN learning process. We illustrate these gaps between representation and optimization with exam…
▽ More
Theoretical studies on the representation power of GNNs have been centered around understanding the equivalence of GNNs, using WL-Tests for detecting graph isomorphism. In this paper, we argue that such equivalence ignores the accompanying optimization issues and does not provide a holistic view of the GNN learning process. We illustrate these gaps between representation and optimization with examples and experiments. We also explore the existence of an implicit inductive bias (e.g. fully connected networks prefer to learn low frequency functions in their input space) in GNNs, in the context of graph classification tasks. We further prove theoretically that the message-passing layers in the graph, have a tendency to search for either discriminative subgraphs, or a collection of discriminative nodes dispersed across the graph, depending on the different global pooling layers used. We empirically verify this bias through experiments over real-world and synthetic datasets. Finally, we show how our work can help in incorporating domain knowledge via attention based architectures, and can evince their capability to discriminate coherent subgraphs.
△ Less
Submitted 23 August, 2024; v1 submitted 17 August, 2024;
originally announced August 2024.
-
Multi Mode (Reflection and Transmission) Operated Dielectric Resonator based Displacement Sensor
Authors:
Premsai Regalla,
Praveen Kumar A. V
Abstract:
The authors propose a multi mode operated dielectric resonator based displacement sensor. A cylindrical dielectric resonator is coupled to two 50 ohm microstrip lines to enable the multi mode operating feature. The parameters like S11 in reflection mode and S21 in transmission mode are sensitive to DR displacements w.r.to transmission lines. The HFSS is carried out for numerical analysis and the f…
▽ More
The authors propose a multi mode operated dielectric resonator based displacement sensor. A cylindrical dielectric resonator is coupled to two 50 ohm microstrip lines to enable the multi mode operating feature. The parameters like S11 in reflection mode and S21 in transmission mode are sensitive to DR displacements w.r.to transmission lines. The HFSS is carried out for numerical analysis and the fabricated sensor prototype is realized with VNA measurement. Both the HFSS and VNA responses are in good agreement with each other for this multi mode operation. This reveals the selection of mode freedom to the choice of operation.
△ Less
Submitted 23 February, 2023;
originally announced February 2023.
