Voyager: a new open-source library for fast nearest-neighbor searches. Voyager uses the HNSW algorithm, outperforming its previous library, Annoy. 

 Haziqa Sajid.jpeg

 Haziqa Sajid, Freelance Technical Writer

 Haziqa Sajid

large_Getting Started with Voyager.png

medium_Getting Started with Voyager.png

small_Getting Started with Voyager.png

thumbnail_Getting Started with Voyager.png

Getting Started with Voyager.png

Getting Started with Voyager: Spotify's Nearest-Neighbor Search Library

ALIGN model is designed to learn visual and language representations from noisy image-alt-text pairs.

large_Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision-20241023-070736.png

medium_Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision-20241023-070736.png

small_Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision-20241023-070736.png

thumbnail_Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision-20241023-070736.png

Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision-20241023-070736.png

ALIGN (A Large-scale ImaGe and Noisy-text embedding) model is designed to learn visual and language representations from noisy image-alt-text pairs.

ALIGN Model Explained

ALIGN Explained: Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision

Mamba is a new architecture for sequence modeling, designed to offer an alternative to the Transformer models commonly used in machine learning. 

large_Mamba_ Linear-Time Sequence Modeling with Selective State Spaces-20241009-120116.png

medium_Mamba_ Linear-Time Sequence Modeling with Selective State Spaces-20241009-120116.png

small_Mamba_ Linear-Time Sequence Modeling with Selective State Spaces-20241009-120116.png

thumbnail_Mamba_ Linear-Time Sequence Modeling with Selective State Spaces-20241009-120116.png

Mamba_ Linear-Time Sequence Modeling with Selective State Spaces-20241009-120116.png

Mamba: A Potential Transformer Replacement

The Potential Transformer Replacement: Mamba

Mixture of Experts (MoE): a neural network architecture to improve model efficiency and scalability by selecting specialized experts for different tasks.


large_What is Mixture of Experts (MoE)_ (1).png

medium_What is Mixture of Experts (MoE)_ (1).png

small_What is Mixture of Experts (MoE)_ (1).png

thumbnail_What is Mixture of Experts (MoE)_ (1).png

What is Mixture of Experts (MoE)_ (1).png

What is Mixture of Experts (MoE)? How it Works and Use Cases

What is Mixture of Experts (MoE)? 

Object detection is a computer vision technique that uses neural networks to classify and locate objects, such as humans, buildings, or cars, in images or video. 

large_What is Object Detection_ A Comprehensive Guide.png

medium_What is Object Detection_ A Comprehensive Guide.png

small_What is Object Detection_ A Comprehensive Guide.png

thumbnail_What is Object Detection_ A Comprehensive Guide.png

What is Object Detection_ A Comprehensive Guide.png

What is Object Detection? 

What is Object Detection? A Comprehensive Guide

DistilBERT was introduced as a smaller, faster, and distilled version of BERT. It maintains 97% of BERT's language understanding capabilities while being 40% small and 60% faster. 

large_DistilBERT, a distilled version of BERT_ smaller, faster, cheaper and lighter.png

medium_DistilBERT, a distilled version of BERT_ smaller, faster, cheaper and lighter.png

small_DistilBERT, a distilled version of BERT_ smaller, faster, cheaper and lighter.png

thumbnail_DistilBERT, a distilled version of BERT_ smaller, faster, cheaper and lighter.png

DistilBERT, a distilled version of BERT_ smaller, faster, cheaper and lighter.png

DistilBERT maintains 97% of BERT's language understanding capabilities while being 40% small and 60% faster.


Distilbert: A Smaller, Faster, and Distilled BERT  

DistilBERT: A Distilled Version of BERT

Building state-of-the-art embedding models for high-quality RAG systems needs careful attention to pretraining, fine-tuning, and scalability. Zilliz Cloud and Milvus help manage embeddings at scale and create more intelligent and responsive neural search systems.



large_Harnessing Embedding Models for AI-Powered Search (3).png

medium_Harnessing Embedding Models for AI-Powered Search (3).png

small_Harnessing Embedding Models for AI-Powered Search (3).png

thumbnail_Harnessing Embedding Models for AI-Powered Search (3).png

Harnessing Embedding Models for AI-Powered Search (3).png

Building state-of-the-art embedding models for high-quality RAG systems needs careful attention to pretraining, fine-tuning, and scalability. Zilliz Cloud and Milvus help manage embeddings at scale and create more intelligent and responsive neural search systems.


Harnessing Embedding Models for AI-Powered Search

A knowledge graph is a data structure representing information as a network of entities and their relationships.

large_What is a Knowledge Graph_.png

medium_What is a Knowledge Graph_.png

small_What is a Knowledge Graph_.png

thumbnail_What is a Knowledge Graph_.png

What is a Knowledge Graph_.png

 What is a Knowledge Graph? 

What is a Knowledge Graph (KG)? 

RoBERTa (A Robustly Optimized BERT Pretraining Approach) is an improved version of BERT designed to address its limitations. 


large_Aug 28 —— RoBERTa_ An Optimized Method for Pretraining Self-supervised NLP Systems.png

medium_Aug 28 —— RoBERTa_ An Optimized Method for Pretraining Self-supervised NLP Systems.png

small_Aug 28 —— RoBERTa_ An Optimized Method for Pretraining Self-supervised NLP Systems.png

thumbnail_Aug 28 —— RoBERTa_ An Optimized Method for Pretraining Self-supervised NLP Systems.png

Aug 28 —— RoBERTa_ An Optimized Method for Pretraining Self-supervised NLP Systems.png

RoBERTa (A Robustly Optimized BERT Pretraining Approach) is an improved version of BERT designed to address its limitations. 

What is RoBERTa? 

RoBERTa: An Optimized Method for Pretraining Self-supervised NLP Systems

Despite advances in LLMs like ChatGPT, search still matters. Combining GenAI with search and vector databases enhances search accuracy and experience.

large_Search Still Matters_ Enhancing Information Retrieval with Generative AI and Vector Databases.png

medium_Search Still Matters_ Enhancing Information Retrieval with Generative AI and Vector Databases.png

small_Search Still Matters_ Enhancing Information Retrieval with Generative AI and Vector Databases.png

thumbnail_Search Still Matters_ Enhancing Information Retrieval with Generative AI and Vector Databases.png

Search Still Matters_ Enhancing Information Retrieval with Generative AI and Vector Databases.png

Search and Information Retrieval in the Era of Generative AI 

Search Still Matters: Enhancing Information Retrieval with Generative AI and Vector Databases

Information retrieval (IR) is the process of efficiently retrieving relevant information from large collections of unstructured or semi-structured data. 

large_What is Information Retrieval_.png

medium_What is Information Retrieval_.png

small_What is Information Retrieval_.png

thumbnail_What is Information Retrieval_.png

What is Information Retrieval_.png

What is Information Retrieval? A Comprehensive Guide. 

What is Information Retrieval?

Vision Transformers (ViTs) are neural network models that use transformers to perform computer vision tasks like object detection and image classification. 

large_Aug 02 --A Beginner's Guide to Understanding Vision Transformer (ViT).png

medium_Aug 02 --A Beginner's Guide to Understanding Vision Transformer (ViT).png

small_Aug 02 --A Beginner's Guide to Understanding Vision Transformer (ViT).png

thumbnail_Aug 02 --A Beginner's Guide to Understanding Vision Transformer (ViT).png

Aug 02 --A Beginner's Guide to Understanding Vision Transformer (ViT).png

What are Vision Transformers (ViT)?

A Beginner's Guide to Understanding Vision Transformers (ViT)

HyDE (Hypothetical Document Embeddings) is a retrieval method that uses "fake" documents to improve the answers of LLM and RAG. 

large_July.31 ——Improving Information Retrieval and RAG with HyDE.png

medium_July.31 ——Improving Information Retrieval and RAG with HyDE.png

small_July.31 ——Improving Information Retrieval and RAG with HyDE.png

thumbnail_July.31 ——Improving Information Retrieval and RAG with HyDE.png

July.31 ——Improving Information Retrieval and RAG with HyDE.png

Better RAG with HyDE - Hypothetical Document Embeddings

Improving Information Retrieval and RAG with Hypothetical Document Embeddings (HyDE)

Knowledge Graphs (KGs) store and link data based on their relationships. KG-enhanced RAG can significantly improve retrieval capabilities and answer quality. 

large_Enhancing Your RAG with Knowledge Graphs (1).png

medium_Enhancing Your RAG with Knowledge Graphs (1).png

small_Enhancing Your RAG with Knowledge Graphs (1).png

thumbnail_Enhancing Your RAG with Knowledge Graphs (1).png

Enhancing Your RAG with Knowledge Graphs (1).png

Enhancing RAG with Knowledge Graphs

Enhancing Your RAG with Knowledge Graphs Using KnowHow

Using LLMware and the Milvus vector database, we can combine the power of vector similarity search and LLMs to ask questions on our private documents.

large_July.10 ——Safeguarding Data Integrity_ On-Prem RAG Deployment with LLMware and Milvus.png

medium_July.10 ——Safeguarding Data Integrity_ On-Prem RAG Deployment with LLMware and Milvus.png

small_July.10 ——Safeguarding Data Integrity_ On-Prem RAG Deployment with LLMware and Milvus.png

thumbnail_July.10 ——Safeguarding Data Integrity_ On-Prem RAG Deployment with LLMware and Milvus.png

July.10 ——Safeguarding Data Integrity_ On-Prem RAG Deployment with LLMware and Milvus.png

Safeguarding Data Integrity: On-Prem RAG Deployment with LLMware and Milvus 

This article discussed the popular voyage AI embedding models and rerankers and their integration with Zilliz Cloud. 

large_Voyage AI Embeddings and Rerankers for Search and RAG.png

medium_Voyage AI Embeddings and Rerankers for Search and RAG.png

small_Voyage AI Embeddings and Rerankers for Search and RAG.png

thumbnail_Voyage AI Embeddings and Rerankers for Search and RAG.png

Voyage AI Embeddings and Rerankers for Search and RAG.png

Voyage AI Embeddings and Rerankers for Search and RAG 

Image Embeddings are the core of modern computer vision algorithms. Understand their implementation and use cases and explore different image embedding models.

large_June 06—— Image Embeddings for Enhanced Image Search (2).png

medium_June 06—— Image Embeddings for Enhanced Image Search (2).png

small_June 06—— Image Embeddings for Enhanced Image Search (2).png

thumbnail_June 06—— Image Embeddings for Enhanced Image Search (2).png

June 06—— Image Embeddings for Enhanced Image Search (2).png

Image Embeddings for Enhanced Image Search

Image Embeddings for Enhanced Image Search: An In-depth Explainer

Integrating the Milvus vector database and MemGPT has taken one step further in streamlining the development of AI Agents with connections to external data sources. In this post, we share an example demonstrating how to use this integration to build a chatbot with external memories. 


large_Introduction to MemGPT and Its Integration with Milvus.png

medium_Introduction to MemGPT and Its Integration with Milvus.png

small_Introduction to MemGPT and Its Integration with Milvus.png

thumbnail_Introduction to MemGPT and Its Integration with Milvus.png

Introduction to MemGPT and Its Integration with Milvus.png

Introduction to MemGPT and Its Integration with Milvus

In this blog post, we will understand the concept of vector embeddings and explore its applications, best practices, and tools for working with embeddings.


large_Everything You Should Know about Vector Embeddings (1).png

medium_Everything You Should Know about Vector Embeddings (1).png

small_Everything You Should Know about Vector Embeddings (1).png

thumbnail_Everything You Should Know about Vector Embeddings (1).png

Everything You Should Know about Vector Embeddings (1).png

An Introduction to Vector Embeddings: What They Are and How to Use Them

Understand how vector databases process data to enhance customer experience and drive business growth. 

Exploring VectorDB Use Cases

large_May 21——Enhancing Customer Experience with  Vector Databases.png

medium_May 21——Enhancing Customer Experience with  Vector Databases.png

small_May 21——Enhancing Customer Experience with  Vector Databases.png

thumbnail_May 21——Enhancing Customer Experience with  Vector Databases.png

May 21——Enhancing Customer Experience with  Vector Databases.png

Enhancing Customer Experience with Vector Databases: A Strategic Approach

Learn the mechanics, applications, and benefits of lexical and semantic search and how to perform it in Zilliz.

large_Navigating the Nuances of Lexical and Semantic Search with Zilliz.png

medium_Navigating the Nuances of Lexical and Semantic Search with Zilliz.png

small_Navigating the Nuances of Lexical and Semantic Search with Zilliz.png

thumbnail_Navigating the Nuances of Lexical and Semantic Search with Zilliz.png

Navigating the Nuances of Lexical and Semantic Search with Zilliz.png

Navigating the Nuances of Lexical and Semantic Search with Zilliz

Explore different ranking algorithms, their usage, and best practices. Uncover the real-world success stories of ranking algorithms and their evolution over time.


large_Ranking Models_ What are they and when to use them_.png

medium_Ranking Models_ What are they and when to use them_.png

small_Ranking Models_ What are they and when to use them_.png

thumbnail_Ranking Models_ What are they and when to use them_.png

Ranking Models_ What are they and when to use them_.png

Ranking Models: What Are They and When to Use Them?  

Mastering color histograms is indispensable for anyone involved in image processing and analysis. By understanding the nuances of color distributions and leveraging advanced techniques, practitioners can unlock the full potential of color histograms in various imaging projects and research endeavors.

Vector Database 101

large_Demystifying Color Histograms_ A Guide to Image Processing and Analysis (1).png

medium_Demystifying Color Histograms_ A Guide to Image Processing and Analysis (1).png

small_Demystifying Color Histograms_ A Guide to Image Processing and Analysis (1).png

thumbnail_Demystifying Color Histograms_ A Guide to Image Processing and Analysis (1).png

Demystifying Color Histograms_ A Guide to Image Processing and Analysis (1).png

 Demystifying Color Histograms: A Guide to Image Processing and Analysis

This article will explore CLIP's inner workings and potential in multimodal learning, with a particular focus on the clip-vit-base-patch32 variant.


LLM 101

large_Exploring OpenAI CLIP_ The Future of Multi-Modal AI Learning.png

medium_Exploring OpenAI CLIP_ The Future of Multi-Modal AI Learning.png

small_Exploring OpenAI CLIP_ The Future of Multi-Modal AI Learning.png

thumbnail_Exploring OpenAI CLIP_ The Future of Multi-Modal AI Learning.png

Exploring OpenAI CLIP_ The Future of Multi-Modal AI Learning.png

Multimodal AI learning can get input and understand information from various modalities like text, images, and audio together, leading to a deeper understanding of the world. Learn more about OpenAI's CLIP (Contrastive Language-Image Pre-training), a popular multimodal model for text and image data.

Exploring OpenAI CLIP: The Future of Multi-Modal AI Learning

This article will discuss the Softmax Activation Function, its applications, challenges, and tips for better performance.


large_Mar 26 ——Vector Database 105.png

medium_Mar 26 ——Vector Database 105.png

small_Mar 26 ——Vector Database 105.png

thumbnail_Mar 26 ——Vector Database 105.png

Mar 26 ——Vector Database 105.png

The softmax function is a mathematical function used in machine learning, particularly in the context of classification tasks. 

Understanding Softmax Activation Function

Decoding Softmax Activation Function