AllSet hypergraph node classifier — hg

Implements the AllSet framework of Chien et al. (2022) as one complete node-to-edge-to-node propagation layer. model = "deepsets" uses Eq. 7, an MLP before and after each multiset sum. model = "transformer" uses Eq. 8: pooling by multihead attention from a learned query, followed by residual MLPs and layer normalization. Both directions have independent learned multiset functions and are permutation invariant.

Usage

hg_allset(
  hg,
  labels,
  features = "incidence",
  model = c("deepsets", "transformer"),
  hidden = 128L,
  heads = 4L,
  epochs = 200L,
  lr = 0.001,
  weight_decay = 5e-04,
  dropout = 0.5,
  validation = 0.1,
  seed = 1L,
  verbose = FALSE
)

Arguments

hg: A text_hypergraph() (or any hypernets net_hg), dense or sparse.
labels: The known labels: a named character vector (names are node identifiers, values class labels) or a tidy data.frame with a node column and a label, cluster or predicted column. At least two classes.
features: Vertex feature matrix, one row per node in hg node order (rownames must match the node names), e.g. sbert embeddings. The default "incidence" uses the hypergraph's own weighted incidence rows (tf-idf bag-of-words features for a nodes = "doc" text hypergraph).
model: "deepsets" for AllDeepSets or "transformer" for AllSetTransformer.
hidden: Hidden width (default 128).
heads: Number of attention heads for AllSetTransformer. hidden must be divisible by heads.
epochs: Training epochs (default 200).
lr, weight_decay: Adam learning rate and L2 penalty.
dropout: Hidden-layer dropout.
validation: Fraction of labeled nodes held out to select an epoch.
seed: Reproducible R and torch seed.
verbose: Message loss every 20 epochs.

Value

A prediction data.frame in the same format as hg_neural(), with training history and the AllSet model name attached.

References

Chien, E., Pan, C., Peng, J., & Milenkovic, O. (2022). You are AllSet: A multiset function framework for hypergraph neural networks. ICLR 2022.