This article explains the evolution of natural language processing, introduces transformer‑based large models such as BERT, GPT and T5, and details how words are represented through one‑hot vectors and dense word embeddings, illustrating their training and analogy capabilities.
In real‑time bidding advertising, a multi‑path recall framework quickly filters billions of ads using parallel non‑personalized and personalized strategies—such as hot‑item rules, collaborative‑filtering, skip‑gram vectors, and GraphSAGE embeddings—while respecting targeting constraints, before a ranking stage optimizes eCPM, with effectiveness measured offline and online and future extensions planned with large language models.
This article reviews how skip‑gram‑based graph embedding methods such as DeepWalk, LINE and node2vec can be interpreted as matrix factorization, explains the NetMF and NetSMF frameworks that use sparse matrix approximations and random SVD for large‑scale networks, and discusses extensions like GATNE and deep clustering approaches to address practical challenges in constructing and applying graph representations.
This article explains how 58.com applies various embedding methods—including ALS, Skip‑gram, and DeepWalk—to vectorize users and properties, improve similarity calculations, and enhance both recall and ranking stages of its real‑estate recommendation system, with detailed technical descriptions and evaluation results.
This article provides a comprehensive overview of embedding techniques, explaining their role in deep learning recommendation systems, detailing Word2Vec and its Skip‑gram model with negative sampling and hierarchical softmax, and extending the discussion to Item2Vec and dual‑tower architectures for item representation.
This article describes how Tujia applied embedding techniques, particularly a Skip‑Gram model, to build an item‑to‑item similarity recommender for low‑frequency, highly personalized homestay listings, detailing the data preparation, model architecture, training process, evaluation results, practical improvements, and future directions.
