Reducing Attention Complexity in Graph Transformers through Subgraph Partitioning
| dc.contributor.author | Choubey, Ranjan Kumar | |
| dc.date.accessioned | 2025-07-22T09:43:15Z | |
| dc.date.available | 2025-07-22T09:43:15Z | |
| dc.date.issued | 2025-06 | |
| dc.description | Dissertation under the supervision of Dr. Swagatam Das | en_US |
| dc.description.abstract | This dissertation addresses the challenge of scaling Graph Transformers by proposing a subgraph-based strategy to reduce attention complexity. The proposed framework preserves representational power while making attention computation tractable for largescale graphs. The method begins by partitioning the input graph into K subgraphs using the METIS algorithm. Each subgraph is encoded using a combination of local structural features from a Graph Convolutional Network (GCN) and global positional cues from Laplacian Positional Embeddings (LPEs). These embeddings are fused via a trainable projection function to form subgraph tokens. A supergraph is constructed to model interactions among subgraphs, allowing attention to be applied over a K × K matrix instead of the full n × n space, thereby reducing complexity from O(n2) to O(K2). Finally, a component-aware prediction strategy maps subgraph-level predictions to individual nodes using learned weights and regularization. Empirical evaluations demonstrate that the framework delivers higher accuracy, improved convergence, and scalability across diverse benchmark datasets. | en_US |
| dc.identifier.citation | 59p. | en_US |
| dc.identifier.uri | http://hdl.handle.net/10263/7593 | |
| dc.language.iso | en | en_US |
| dc.publisher | Indian Statistical Institute, Kolkata | en_US |
| dc.relation.ispartofseries | MTech(CS) Dissertation;23-16 | |
| dc.subject | Graph Transformer | en_US |
| dc.subject | Subgraph Attention | en_US |
| dc.subject | Laplacian Positional Embeddings | en_US |
| dc.subject | Component-Aware Propagation | en_US |
| dc.subject | Scalable Graph Learning | en_US |
| dc.title | Reducing Attention Complexity in Graph Transformers through Subgraph Partitioning | en_US |
| dc.type | Other | en_US |
Files
Original bundle
1 - 2 of 2
No Thumbnail Available
- Name:
- Plagiarism_Report.pdf
- Size:
- 1.99 MB
- Format:
- Adobe Portable Document Format
- Description:
- Plagiarism_report
No Thumbnail Available
- Name:
- Ranjan_CS2316_Dissertation.pdf
- Size:
- 2.45 MB
- Format:
- Adobe Portable Document Format
- Description:
- Dissertations - M Tech (CS)
License bundle
1 - 1 of 1
No Thumbnail Available
- Name:
- license.txt
- Size:
- 1.71 KB
- Format:
- Item-specific license agreed upon to submission
- Description:
