https://tgraphx.com Technical articles, tutorials, and research notes about TGraphX — tensor-aware GNN for PyTorch. en Sat, 06 Jun 2026 09:33:19 +0000 https://tgraphx.com/articles/evolutionary-optimization-graph-structures-tgraphx/ https://tgraphx.com/articles/evolutionary-optimization-graph-structures-tgraphx/ Sat, 06 Jun 2026 09:30:00 +0000 A practical guide to TGraphX's evolutionary optimization subsystem — GA, simulated annealing, and NSGA-II for searching over graph structures. Tutorial https://tgraphx.com/articles/tgraphx-vs-pykeen/ https://tgraphx.com/articles/tgraphx-vs-pykeen/ Fri, 05 Jun 2026 09:30:00 +0000 A balanced comparison of TGraphX's KG subsystem and PyKEEN, the established library for knowledge graph embedding benchmarks. When each fits. Comparison https://tgraphx.com/articles/tgraphx-benchmark-disclaimers/ https://tgraphx.com/articles/tgraphx-benchmark-disclaimers/ Thu, 04 Jun 2026 09:30:00 +0000 A look at TGraphX's benchmark artifact system, what gets recorded automatically, and how it supports honest benchmark reporting. Article https://tgraphx.com/articles/neighbor-sampling-tgraphx-large-graphs/ https://tgraphx.com/articles/neighbor-sampling-tgraphx-large-graphs/ Wed, 03 Jun 2026 09:30:00 +0000 A practical guide to mini-batch graph training in TGraphX using NeighborLoader, GraphSAINT, and ClusterLoader — the three sampling strategies that scale… Tutorial https://tgraphx.com/articles/neuroschemax-schema-aware-neural-architecture/ https://tgraphx.com/articles/neuroschemax-schema-aware-neural-architecture/ Tue, 02 Jun 2026 09:30:00 +0000 Why schema awareness matters in deep learning over structured data, and the design direction that schema-aware architectures take. Package https://tgraphx.com/articles/graph-reinforcement-learning-tgraphx/ https://tgraphx.com/articles/graph-reinforcement-learning-tgraphx/ Mon, 01 Jun 2026 09:30:00 +0000 A practical introduction to TGraphX's graph reinforcement learning subsystem — environments, algorithms, and what to expect from a research-grade module… Tutorial https://tgraphx.com/articles/flat-vector-node-features-insufficient-gnn/ https://tgraphx.com/articles/flat-vector-node-features-insufficient-gnn/ Sun, 31 May 2026 09:30:00 +0000 A decision guide: when flat node feature vectors work well, when they break, and how to tell which category your task falls into. Article https://tgraphx.com/articles/graph-learning-research-integrity-benchmarks/ https://tgraphx.com/articles/graph-learning-research-integrity-benchmarks/ Sat, 30 May 2026 09:30:00 +0000 GNN benchmark numbers often look more impressive than they are. This note discusses the common evaluation shortcuts that inflate results and how to read… Research Note https://tgraphx.com/articles/knowledge-graph-embedding-tensor-features-tgraphx/ https://tgraphx.com/articles/knowledge-graph-embedding-tensor-features-tgraphx/ Fri, 29 May 2026 09:30:00 +0000 A practical tutorial on TGraphX's knowledge graph subsystem: TransE, DistMult, ComplEx, and RotatE, with optional tensor-valued entity features. Tutorial https://tgraphx.com/articles/llm-friendly-graph-apis-tgraphx/ https://tgraphx.com/articles/llm-friendly-graph-apis-tgraphx/ Thu, 28 May 2026 09:30:00 +0000 AI coding tools produce better graph code when the API is explicit, the error messages are actionable, and the canonical surface is small. Article https://tgraphx.com/articles/shape-aware-validation-tgraphx/ https://tgraphx.com/articles/shape-aware-validation-tgraphx/ Wed, 27 May 2026 09:30:00 +0000 Most GNN bugs are shape bugs in disguise. TGraphX provides validation utilities that catch them early, before they show up as obscure runtime errors after… Article https://tgraphx.com/articles/tgraphx-vs-pytorch-geometric-v2/ https://tgraphx.com/articles/tgraphx-vs-pytorch-geometric-v2/ Tue, 26 May 2026 09:30:00 +0000 A decision-oriented comparison aimed at researchers choosing a graph learning framework for a specific project. Comparison https://tgraphx.com/articles/gnn-research-reproducibility-tgraphx/ https://tgraphx.com/articles/gnn-research-reproducibility-tgraphx/ Mon, 25 May 2026 09:30:00 +0000 Reproducing published GNN results is harder than it should be. This note walks through the hidden sources of non-determinism and how explicit tooling helps. Research Note https://tgraphx.com/articles/tensor-valued-nodes-deep-tutorial-tgraphx/ https://tgraphx.com/articles/tensor-valued-nodes-deep-tutorial-tgraphx/ Sun, 24 May 2026 09:30:00 +0000 A step-by-step deeper tutorial: validate a tensor graph, build a custom training loop with NeighborLoader, save artifacts, and reproduce results. Tutorial https://tgraphx.com/articles/tensor-valued-nodes-graph-neural-networks/ https://tgraphx.com/articles/tensor-valued-nodes-graph-neural-networks/ Sat, 23 May 2026 09:30:00 +0000 Most GNN frameworks assume every node is a flat vector. That assumption breaks for image patches, volumetric blocks, and sequences. Article https://tgraphx.com/articles/what-is-tgx-graph/ https://tgraphx.com/articles/what-is-tgx-graph/ Fri, 22 May 2026 09:30:00 +0000 TGraphX is a tensor-native graph learning framework for PyTorch. This introduction explains what a TGX graph is, why tensor-valued node features matter, and… blog_post https://tgraphx.com/articles/getting-started-tensor-valued-nodes-tgraphx/ https://tgraphx.com/articles/getting-started-tensor-valued-nodes-tgraphx/ Fri, 22 May 2026 09:30:00 +0000 A step-by-step tutorial for representing graph data with tensor-valued node features in TGraphX — including validation, a first training run, and common… tutorial https://tgraphx.com/articles/tgraphx-vs-pytorch-geometric/ https://tgraphx.com/articles/tgraphx-vs-pytorch-geometric/ Fri, 22 May 2026 09:30:00 +0000 A balanced comparison of TGraphX and PyTorch Geometric. Both run on PyTorch, but they serve different use cases. comparison_page