#representation learning

IMGNet introduces a face verification model that replaces cosine similarity with sliding window sign pattern matching, achieving 96.27% accuracy on LFW with a 10.58 MB model trained on CASIA-WebFace. This novel approach could enable more efficient and compact face verification systems, and the finding that sign pattern matching improves ArcFace embeddings without retraining suggests a fundamental property of well-trained face embeddings. The model uses a SW Block replacing standard convolution with multi-scale relational operations, an IMG Sign MSE Loss defined purely over sign pattern agreement, and a voting system combining three metrics. Applied to ArcFace (buffalol) without retraining, IMG Sign Score achieves 99.58% on LFW, only 0.24% below ArcFace+Cosine.

A new knowledge graph system called BaryGraph treats every relationship as a first-class embedded document called a BaryEdge, rather than an edge between nodes. It also introduces recursive MetaBary triads to discover structural bridges between distant concepts. This approach addresses a key limitation of flat vector search, which treats relationships as mere proximity between points, missing cross-domain connections. By embedding relationships themselves, BaryGraph can surface analogies and bridges that standard RAG systems cannot detect. The system uses nomic-embed-text for embeddings, runs on MongoDB Community with mongot, and processes the entire English Wiktionary (6.6M documents) locally. It is available as an open-source project with a live MCP server for probing.