Pith. sign in

REVIEW 3 cited by

NodePiece: Compositional and Parameter-Efficient Representations of Large Knowledge Graphs

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2106.12144 v2 pith:TW6VJKNX submitted 2021-06-23 cs.CL cs.AIcs.LG

NodePiece: Compositional and Parameter-Efficient Representations of Large Knowledge Graphs

classification cs.CL cs.AIcs.LG
keywords embeddingnodepieceentitygraphvocabularyfewerfixed-sizegraphs
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Conventional representation learning algorithms for knowledge graphs (KG) map each entity to a unique embedding vector. Such a shallow lookup results in a linear growth of memory consumption for storing the embedding matrix and incurs high computational costs when working with real-world KGs. Drawing parallels with subword tokenization commonly used in NLP, we explore the landscape of more parameter-efficient node embedding strategies with possibly sublinear memory requirements. To this end, we propose NodePiece, an anchor-based approach to learn a fixed-size entity vocabulary. In NodePiece, a vocabulary of subword/sub-entity units is constructed from anchor nodes in a graph with known relation types. Given such a fixed-size vocabulary, it is possible to bootstrap an encoding and embedding for any entity, including those unseen during training. Experiments show that NodePiece performs competitively in node classification, link prediction, and relation prediction tasks while retaining less than 10% of explicit nodes in a graph as anchors and often having 10x fewer parameters. To this end, we show that a NodePiece-enabled model outperforms existing shallow models on a large OGB WikiKG 2 graph having 70x fewer parameters.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. InductWave: Inductive Multi-Hop Logical Query Answering on Knowledge Graphs

    cs.AI 2026-07 conditional novelty 6.0

    Combining graph wavelet embeddings with Neural Bellman-Ford message passing reduces the layers needed for inductive logical query answering on large knowledge graphs.

  2. A Hierarchical Quantized Tokenization Framework for Task-Adaptive Graph Representation Learning

    cs.IR 2025-10 unverdicted novelty 6.0

    A hierarchical quantized tokenization framework with task-adaptive routing and dual-view streams for improved graph representation learning on node classification and link prediction.

  3. A Hierarchical Quantized Tokenization Framework for Task-Adaptive Graph Representation Learning

    cs.IR 2025-10 conditional novelty 4.0

    QUIET is a hierarchical RVQ-based graph tokenizer with a learned level-weighting gate; it improves several benchmarks but not consistently against the strongest baselines.