Michael Alcorn, Sr. Software Engineer, Red Hat Inc.
summary
Shows concrete systems: trains word2vec on click streams as URL2vec (treating each day as a sentence) to surface contextually similar docs that TF‑IDF misses, and runs doc2vec over per-customer daily sessions to produce customer vectors used for clustering and learning‑to‑rank features. Implements Microsoft’s Deep Semantic Similarity Model to detect conceptually duplicate portal solutions and finds different, useful duplicates than bag‑of‑words. Prototypes include a color-name generator improved by pretrained embeddings and MLB pitcher/batter embeddings that cluster by style and support analogies (Mike Trout → Bryce Harper).
AI-written summary · about
chapters (6)
- 0:00Start
- 0:04Outline
- 0:18Background
- 5:12Naming Colors
- 11:19Duplicate Detection
- 16:51(batter pitcher)2vec
related talks
31mTaking AI/ML to the next level: Text!
25mAnusua Trivedi, Data Scientist at Texas Advanced Computing Center (TACC), UT Austin
28m"Embeddings: How Computers Learned to Read"
28mByron Galbraith, Chief Data Scientist, Talla at MLconf Seattle 2017
27mMachine Learning for Advanced AI Search
59mContextual search with vector search: exploring your options with open source tools
50mA fun and absurd introduction to Vector Databases
33mUsing the Amazing World