Introduction
To provide advanced search capabilities for the AI revolution, Tencent Cloud Elasticsearch Service officially launched version 8.8.1! This release features the Elasticsearch Relevance Engine™ (ESRE™), a powerful AI-enhanced search engine that brings a new cutting-edge experience to search and analytics.
1. Native vector search engine
Efficiently create, store, and search dense vectors using Elasticsearch as a vector database. Its features include:
1) Provide a graph index to achieve efficient nearest neighbor search through HNSW.
Figure 1
2) Supports end-to-end vector generation, vector indexing, and vector similarity comparison, without requiring additional platforms for vector inference.
Figure II
3) Integrated search experience: multi-channel recall mixed scoring, Faceting aggregation analysis capabilities, role-based access control.
Figure three
4) Store vectors as bytes instead of floating points. At the same time, methods such as principal component analysis (PCA) are applied to support the reduction of vector dimensions and save storage space.
Figure four
5) Continuously optimized vector index and query performance.
Figure five
2. New Hybrid Sorting Algorithm
Reciprocal Rank Fusion (RRF) hybrid ranking algorithm supports full-text search and vector search, allowing developers to better optimize the Al search engine and enable it to achieve combined semantic and keyword queries.
Figure six
3. Customize the Transformer model
Developers can manage and use their own transformer models in Elastic to complete various natural language processing tasks to suit specific business scenarios.
Figure seven
4. Optimized search model trained by Elastic
Use the out-of-the-box Learned Sparse Encoder machine learning model trained by Elastic to optimize search, which can provide better relevance and semantic search in various fields.
Figure eight
5. Integrate with the third-party Transformer model to extract intuitive summaries
Integrate with large language models (such as OpenAl's GPT-3 and 4) via API to extract summaries from Elasticsearch data sources.
Figure 9
Fully apply various Natural Language Processing (NLP) tasks and models
ESRE™ has powerful natural language processing capabilities and can handle various NLP tasks and models, making search results more in line with the semantics of natural language.
Figure ten
Integrate with third-party tools to build complex data pipelines and generative AI applications. Integrate with third-party tools such as LangChain to help users build complex data pipelines and generative AI applications.
Figure Eleven
With this release, we hope to help users easily switch to a cutting-edge search and analysis experience powered by GAI. The powerful functions of ESRE™ will provide your business with more accurate, smarter and more efficient search services, helping you achieve greater success in the highly competitive market.
6. How to use it?
Log in to the Tencent Cloud ES console, select the version as 8.8.1, and complete the cluster creation.