Results 251 to 260 of about 6,811,260 (293)

On the Expressive Power of Spectral Invariant Graph Neural Networks

International Conference on Machine Learning
Incorporating spectral information to enhance Graph Neural Networks (GNNs) has shown promising results but raises a fundamental challenge due to the inherent ambiguity of eigenvectors.
Bohang Zhang, Lingxiao Zhao, Haggai Maron   +2 more
semanticscholar   +1 more source

The expressive power of clocks

1995
We investigate the expressive power of timing restrictions on labeled transition systems. In particular, we show how constraints on clock variables together with a uniform liveness condition—the divergence of time—can express Buchi, Muller, Streett, Rabin, and weak and strong fairness conditions on a given labeled transition system.
Henzinger, Thomas A., Kopke, Peter W., Wong-Toi, Howard   +2 more
openaire   +1 more source

Improving Expressive Power of Spectral Graph Neural Networks with Eigenvalue Correction

AAAI Conference on Artificial Intelligence
In recent years, spectral graph neural networks, characterized by polynomial filters, have garnered increasing attention and have achieved remarkable performance in tasks such as node classification. These models typically assume that eigenvalues for the
Kangkang Lu, Yanhua Yu, Hao Fei, Xuan Li, Zixuan Yang, Zirui Guo, Meiyu Liang, Mengran Yin, Tat-Seng Chua   +8 more
semanticscholar   +1 more source

On the Theoretical Expressive Power and the Design Space of Higher-Order Graph Transformers

International Conference on Artificial Intelligence and Statistics
Graph transformers have recently received significant attention in graph learning, partly due to their ability to capture more global interaction via self-attention. Nevertheless, while higher-order graph neural networks have been reasonably well studied,
Cai Zhou, Rose Yu, Yusu Wang
semanticscholar   +1 more source

Powers of Expression, Expressions of Power

1999
Abstract Can a speaker's words ever be faithfully reported? History, philosophy, ethnography, political theory, linguistics, and literary criticism all involve debates about discourse and representation. By drawing from Plato's theory of discourse, the lively analysis of speech presentation in this book provides a coherent and original ...
openaire   +1 more source

On the expressive power of CCS

1995
In the context of structured operational semantics, a useful measure of the expressive power of a process algebra \(\mathcal{P}\)is the class of operator specifications definable in \(\mathcal{P}\)up to a given equivalence. Our goal in this study is to characterize this class of operators for CCS up to strong and branching bisimulation.
Ashvin Dsouza, Bard Bloom
openaire   +1 more source

The expressive power of voting polynomials

Combinatorica, 1991
A new lower bound technique for Boolean circuits is presented where polynomials over the rationals are used to describe (viz., to approximate) Boolean functions. The degree of the polynomial is related to the accurately of the approximation. Techniques to handle symmetric functions are derived.
James Aspnes, Richard Beigel, Merrick L. Furst, Steven Rudich   +3 more
openaire   +2 more sources

GQL and SQL/PGQ: Theoretical Models and Expressive Power

Proceedings of the VLDB Endowment
SQL/PGQ and GQL are very recent international standards for querying property graphs: SQL/PGQ specifies how to query relational representations of property graphs in SQL, while GQL is a standalone language for graph databases.
A. Gheerbrant, Leonid Libkin, Liat Peterfreund, A. Rogova   +3 more
semanticscholar   +1 more source

A Theory on Deep Neural Network Based Vector-to-Vector Regression With an Illustration of Its Expressive Power in Speech Enhancement

IEEE/ACM Transactions on Audio Speech and Language Processing, 2019
This paper focuses on a theoretical analysis of deep neural network (DNN) based functional approximation. Leveraging upon two classical theorems on universal approximation, an artificial neural network (ANN) with a single hidden layer of neurons is used.
Jun Qi, Jun Du, Sabato Marco Siniscalchi, Chin-Hui Lee   +3 more
semanticscholar   +1 more source

On the expressive power of rewriting

1997
In this paper we address the open problem of classifying the expressive power of classes of rewriting systems. We introduce a framework to reason about the relative expressive power between classes of rewrite system, with respect to every property of interest P.
openaire   +1 more source