Zobrazeno 1 - 10
of 77
pro vyhledávání: '"Chamberlain, Benjamin"'
We propose a new class of deep reinforcement learning (RL) algorithms that model latent representations in hyperbolic space. Sequential decision-making requires reasoning about the possible future consequences of current behavior. Consequently, captu
Externí odkaz:
http://arxiv.org/abs/2210.01542
Autor:
Rusch, T. Konstantin, Chamberlain, Benjamin P., Mahoney, Michael W., Bronstein, Michael M., Mishra, Siddhartha
We present Gradient Gating (G$^2$), a novel framework for improving the performance of Graph Neural Networks (GNNs). Our framework is based on gating the output of GNN layers with a mechanism for multi-rate flow of message passing information across
Externí odkaz:
http://arxiv.org/abs/2210.00513
Autor:
Chamberlain, Benjamin Paul, Shirobokov, Sergey, Rossi, Emanuele, Frasca, Fabrizio, Markovich, Thomas, Hammerla, Nils, Bronstein, Michael M., Hansmire, Max
Publikováno v:
The Eleventh International Conference on Learning Representations 2023 (oral - top 5%)
Many Graph Neural Networks (GNNs) perform poorly compared to simple heuristics on Link Prediction (LP) tasks. This is due to limitations in expressive power such as the inability to count triangles (the backbone of most LP heuristics) and because the
Externí odkaz:
http://arxiv.org/abs/2209.15486
Autor:
Di Giovanni, Francesco, Rowbottom, James, Chamberlain, Benjamin P., Markovich, Thomas, Bronstein, Michael M.
Graph Neural Networks (GNNs) typically operate by message-passing, where the state of a node is updated based on the information received from its neighbours. Most message-passing models act as graph convolutions, where features are mixed by a shared
Externí odkaz:
http://arxiv.org/abs/2206.10991
Autor:
Bodnar, Cristian, Di Giovanni, Francesco, Chamberlain, Benjamin Paul, Liò, Pietro, Bronstein, Michael M.
Cellular sheaves equip graphs with a "geometrical" structure by assigning vector spaces and linear maps to nodes and edges. Graph Neural Networks (GNNs) implicitly assume a graph with a trivial underlying sheaf. This choice is reflected in the struct
Externí odkaz:
http://arxiv.org/abs/2202.04579
Autor:
Rusch, T. Konstantin, Chamberlain, Benjamin P., Rowbottom, James, Mishra, Siddhartha, Bronstein, Michael M.
We propose Graph-Coupled Oscillator Networks (GraphCON), a novel framework for deep learning on graphs. It is based on discretizations of a second-order system of ordinary differential equations (ODEs), which model a network of nonlinear controlled a
Externí odkaz:
http://arxiv.org/abs/2202.02296
Autor:
Topping, Jake, Di Giovanni, Francesco, Chamberlain, Benjamin Paul, Dong, Xiaowen, Bronstein, Michael M.
Most graph neural networks (GNNs) use the message passing paradigm, in which node features are propagated on the input graph. Recent works pointed to the distortion of information flowing from distant nodes as a factor limiting the efficiency of mess
Externí odkaz:
http://arxiv.org/abs/2111.14522
Autor:
Rossi, Emanuele, Kenlay, Henry, Gorinova, Maria I., Chamberlain, Benjamin Paul, Dong, Xiaowen, Bronstein, Michael
While Graph Neural Networks (GNNs) have recently become the de facto standard for modeling relational data, they impose a strong assumption on the availability of the node or edge features of the graph. In many real-world applications, however, featu
Externí odkaz:
http://arxiv.org/abs/2111.12128
Autor:
Chamberlain, Benjamin Paul, Rowbottom, James, Eynard, Davide, Di Giovanni, Francesco, Dong, Xiaowen, Bronstein, Michael M
We propose a novel class of graph neural networks based on the discretised Beltrami flow, a non-Euclidean diffusion PDE. In our model, node features are supplemented with positional encodings derived from the graph topology and jointly evolved by the
Externí odkaz:
http://arxiv.org/abs/2110.09443
Autor:
Chamberlain, Benjamin Paul, Rowbottom, James, Gorinova, Maria, Webb, Stefan, Rossi, Emanuele, Bronstein, Michael M.
We present Graph Neural Diffusion (GRAND) that approaches deep learning on graphs as a continuous diffusion process and treats Graph Neural Networks (GNNs) as discretisations of an underlying PDE. In our model, the layer structure and topology corres
Externí odkaz:
http://arxiv.org/abs/2106.10934