Zachariah ZhanginNLPlanetTraining State of the Art Semantic Search ModelsTLDR — The key innovation which has allowed for huge advances in semantic search has primarily been improvements in training methodology…8 min read·Dec 4, 2022----
Zachariah ZhangColBERT — A Late Interaction Model For Semantic SearchTLDR — Single vector embedding search model’s are efficient for search but they create an information bottleneck which limits performance…7 min read·Nov 17, 2022--2--2
Zachariah ZhangCombining Embedding and Keyword Based Search for Improved PerformanceTLDR — Ensembling keyword and embedding models for search is one of the quickest and easiest ways to improve search performance over the…5 min read·Nov 15, 2022----
Zachariah ZhangBoot Strapping Data for Natural Language Interfaces from ScratchTLDR — Semantic parsing is a task in which we map from a natural language utterance to a logical form program that can be executed. This…11 min read·Nov 7, 2022----
Zachariah ZhangTraining Cross Lingual Semantic Search Models with Monolingual DataTLDR — Here we look at approaches for training cross lingual search models using only english training data. Zero shot transfer is the…6 min read·Oct 26, 2022----
Zachariah ZhangEfficient Extractive Question Answering on CPU using QUIPTLDR — Extractive question answering is an important task for providing a good user experience in many applications. The popular…7 min read·Oct 24, 2022----
Zachariah ZhanginBetter ProgrammingUsing Large Language Models for Data LabelingLet the AI annotate data for you8 min read·Oct 21, 2022----
Zachariah ZhangSparse Representations of Text for SearchTLDR — While dense embeddings are still the most common approach for search, sparse representations of text can achieve competitive…8 min read·Oct 18, 2022----
Zachariah ZhangZero-Shot Named Entity Recognition Using Question AnsweringTLDR8 min read·Oct 16, 2022--2--2