Real User Monitoring (RUM) is a technique used to collect data on how real users interact with a website or application in real time. It involves tracking user interactions, such as page views, clicks, and load times, directly from the users’ web browsers. This data provides insights into the actual user experience, including performance issues, errors, and user behavior. 

Benefits of Real User Monitoring (RUM) for Businesses

Enhanced User Experience: RUM provides real-time visibility into how users experience your website or application, allowing businesses to identify and address performance issues quickly to ensure a smooth and responsive user experience. 

Data-Driven Decision Making: By analyzing RUM data, businesses can make informed decisions about website or application optimizations, prioritize improvements based on actual user behavior, and measure the impact of changes over time. 

Improved Customer Satisfaction: Monitoring and optimizing the user experience can lead to increased customer satisfaction, loyalty, and retention, ultimately driving business growth and success. 

How RUM Works in Elastic

Elastic’s Real User Monitoring (RUM) solution allows businesses to collect, analyze, and visualize real-time data on user interactions with their web applications. It involves embedding a lightweight JavaScript agent into web pages or applications to capture performance metrics and user interactions directly from users’ browsers. This data is then sent to Elasticsearch for storage and analysis, where it can be visualized and explored in Kibana dashboards. 

Real User Monitoring use cases 

1. Performance Monitoring

Identify Performance Bottlenecks: RUM allows businesses to pinpoint specific areas of their website or application that are experiencing performance issues, such as slow page loads or high server response times. By analyzing performance metrics collected from real user interactions, businesses can diagnose the root causes of these bottlenecks and take proactive measures to address them. 

Diagnose Slow Page Loads: With RUM, businesses can track page load times across different devices, browsers, and geographical locations. This granular level of insight enables them to identify pages with the slowest load times and prioritize optimization efforts to improve overall website performance. 

Detect Errors Affecting User Experience: RUM provides real-time monitoring of errors occurring within web pages or applications, such as JavaScript errors, HTTP errors, or server-side errors. By promptly detecting and analyzing these errors, businesses can minimize their impact on user experience and prevent potential revenue loss. 

2. Feature Adoption 

Track User Engagement: RUM allows businesses to track how users interact with specific features or pages within their website or application. By monitoring user interactions, such as clicks, scrolls, or form submissions, businesses can gain insights into feature adoption rates and usage patterns. This data helps businesses identify which features are most popular among users and which may require further optimization or promotion. 

Understand Feature Adoption Rates: By analyzing RUM data, businesses can measure the adoption rates of new features or updates rolled out to their website or application. This enables them to gauge the effectiveness of their feature releases and make data-driven decisions about future development efforts. 

3. Conversion Optimization 

Analyze User Journeys: RUM allows businesses to track user journeys through their website or application, from initial interaction to conversion. By analyzing user behavior, such as page views, clicks, and session duration, businesses can identify friction points in the conversion funnel and optimize the user experience accordingly. 

Identify Friction Points: RUM helps businesses identify areas of their website or application where users are dropping off or encountering obstacles in the conversion process. Whether it’s a complex checkout flow, unclear calls-to-action, or technical issues, RUM enables businesses to pinpoint these friction points and implement targeted optimizations to improve conversion rates. 

Improve Conversion Rates: By continuously monitoring and analyzing RUM data, businesses can iteratively optimize their conversion funnels to increase conversion rates. Whether through A/B testing, usability improvements, or performance optimizations, RUM provides valuable insights that enable businesses to maximize the effectiveness of their digital experiences and drive revenue growth. 

Guide Through RUM in an Example 

Let’s say you’re a retail company with an e-commerce website. With Elastic’s RUM solution, you embed the RUM JavaScript agent into your website’s pages. This agent collects data on user interactions, such as page views, clicks, and load times, directly from users’ browsers. 

In the Kibana dashboard, you can visualize this data to gain insights into how users are navigating your website, which pages are most popular, and where users are experiencing performance issues. You notice that the checkout process has a high abandonment rate due to slow page load times. 

Using this insight, you prioritize optimizing the checkout process to improve performance and reduce friction for users. After implementing optimizations, you monitor the impact on user behavior and conversion rates through RUM data. Over time, you see a significant improvement in checkout completion rates and overall user satisfaction, leading to increased revenue and customer loyalty. 

Conclusion

In conclusion, Real User Monitoring (RUM) with Elastic offers businesses invaluable insights into user interactions, performance metrics, and conversion rates, enabling them to enhance user experiences and drive business growth. Contact Elk Factory today to implement Elastic’s RUM solution and optimize your digital presence. 

Het bericht How Real User Monitoring (RUM) helps to improve your website or application with Elastic verscheen eerst op Elk Factory.

Synthetic monitoring offers many benefits for businesses that would like to deliver exceptional digital experiences: 

• Early detection: Identify and address performance bottlenecks and errors before they affect real users, minimizing downtime and revenue loss.
• Functionality validation: Ensure the functionality and uptime of critical applications, providing users with reliable and consistent experiences. 
• Business continuity: Monitor critical transactions and workflows to maintain business operations and continuity, even during peak demand periods. 
• Performance benchmarking: Compare performance metrics against industry standards and competitors to gauge competitiveness and drive continuous improvement. 

Implementing synthetic monitoring within Elastic is a streamlined process that empowers organizations to: 

• Configure synthetic checks: Define and configure synthetic checks to simulate user journeys and interactions, tailored to specific application functionalities. 
• Define alert thresholds: Establish alert thresholds and notification criteria to promptly identify and address performance anomalies, ensuring proactive resolution. 
• Integrate data sources: Seamlessly integrate synthetic monitoring data with other observability tools within Elastic, enabling comprehensive analysis and actionable insights. 

Real-world examples showcase the versatility and effectiveness of synthetic monitoring within Elastic Observability: 

• E-commerce Performance: Monitor the availability and performance of e-commerce websites during peak shopping seasons to maintain user satisfaction and drive sales. 

• Critical application validation: Validate the functionality of critical business applications, such as banking or healthcare systems, to ensure uninterrupted service delivery. 

• Cloud-native performance: Proactively identify and resolve performance issues in cloud-native environments and microservices architectures, optimizing resource utilization and enhancing scalability. 

Synthetic monitoring, as part of Elastic Observability, helps organizations to proactively monitor and optimize digital experiences. If you use synthetic monitoring, you can ensure optimal application performance and user satisfaction. 

Ready to elevate your observability capabilities with synthetic monitoring? Explore the possibilities of Elastic Observability and synthetic monitoring solutions. Contact our team for personalized guidance and support.

Het bericht Master monitoring and detect performance issues with Elastic Synthetic Monitoring verscheen eerst op Elk Factory.

Het bericht The power of Elastic: An interview with Yassin Tabla, Elastic Consultant at Elk Factory verscheen eerst op Elk Factory.

In the automotive sector, real-time data and system performance are critical. Porsche Informatik, a subsidiary that provides digital solutions to the automotive industry, recognized the need for a robust tool to monitor and analyze their vast IT infrastructure. The company chose Elastic Observability to overcome this challenge, using its powerful abilities to ensure optimal system performance and enhance operational efficiency. 

Challenges and Objectives

Porsche Informatik’s primary goal was to streamline their monitoring processes across a complex network of applications and services. They needed a solution that could not only handle large volumes of data but also provide real-time insights and diagnostics to preemptively address potential issues. 

Implementation and Results

By integrating Elastic Observability, Porsche Informatik was able to create a centralized observability framework. This setup enabled them to efficiently track and analyze data from various sources, including applications and infrastructure. The real-time analytics features of Elastic Observability allowed them to quickly identify and resolve issues, reducing downtime and improving user experience. 

The implementation of Elastic Observability has empowered Porsche Informatik to maintain a proactive stance on system management. It has improved their ability to predict potential system failures and handle anomalies efficiently, ensuring continuous operational flow and better service delivery to their clients. 

Key benefits 

• Enhanced Monitoring: Comprehensive visibility into system operations, facilitating quicker response times. 
• Proactive Issue Resolution: Ability to anticipate problems before they affect system performance. 
• Improved Operational Efficiency: Streamlined processes and reduced downtime, leading to higher productivity. 

The success of Porsche Informatik with Elastic Observability underscores the critical role that advanced monitoring tools play in today’s digital landscape. As businesses continue to handle increasing volumes of data and demand higher performance from their IT environments, solutions like Elastic Observability become indispensable. This technology not only enhances system transparency and responsiveness but also drives significant improvements in operational reliability and efficiency. 

For companies aiming to elevate their IT capabilities, the integration of Elastic Observability offers a clear pathway to achieving these goals. It provides the tools necessary to navigate the complexities of modern IT systems, ensuring businesses can thrive in an increasingly competitive market. 

As an Elastic solution provider, we at ELK Factory are committed to helping you discover and implement these transformative solutions. Let us help you unlock the full potential of your IT infrastructure. Click here to find out more about Elastic Observability. 

This article is based on the following source.

Het bericht Enhancing automotive IT performance with Elastic Observability: A Porsche Informatik case study verscheen eerst op Elk Factory.

Observability is fundamental to understanding and improving how systems and applications perform and maintain their health. This specific approach integrates the extensive gathering and examination of data via logs, metrics, and traces, presenting a complete insight into the mechanics of IT environments.  

A critical component of Observability is APM, also known as Application Performance Monitoring. APM focuses specifically on the performance and reliability of applications, providing targeted insights into user experience impacts like delays and errors. This approach to monitoring allows teams to quickly identify and address issues, making sure that there is an optimal application performance within the broader scope of system health. 

Why Application Performance Monitoring (APM) matters 

Optimal application performance is essential for businesses to function. Application Performance Monitoring (APM) plays a crucial role in achieving this goal by providing valuable insights into application behavior and performance. Let’s explore how APM benefits companies: 

Improved user experience: APM helps ensure a smooth and responsive user experience, critical for customer satisfaction and retention by monitoring application performance in real time.  

Enhanced operational efficiency: APM tools enable quicker identification and resolution of performance issues, therefore, and as a result, reduces downtime and operational costs. 

Optimized root cause analysis: APM provides detailed insights into application behavior, it becomes easier to perform root cause analysis and prevent future issues. 

Reduced MTTR: With APM, businesses can significantly lower the mean time to resolution for application issues, this ensures minimal impact on business operations. 

DevOps practices support: APM integrates into continuous integration and continuous deployment (CI/CD) pipelines. It promotes more efficient development and deployment practices. 

Optimized resource allocation: By identifying underperforming resources or bottlenecks, APM allows for better allocation of IT resources, increasing overall system performance.

How Application Performance Monitoring (APM) works in Elastic 

APM works by monitoring, collecting, and analyzing data from applications to track their performance and identify issues that could affect user experience. It uses tools to gather detailed information about how applications behave and interact with underlying infrastructure. This data is analyzed to detect anomalies, performance bottlenecks, or errors, allowing IT teams to perform root cause analysis and resolve issues quickly.  

APM provides visibility into the application’s operations, from the end-user experience through to the backend processes, ensuring that applications meet performance standards and user expectations. 

Elastic APM, part of the Elastic Stack, extends this functionality by offering seamless integration with logging and metrics, providing a unified view on system and application health. It can use several alternatives to gather performance metrics, errors, and transaction data related to your applications. This data is then visualized in Kibana, allowing for in-depth analysis and monitoring, enabling teams to quickly address issues and improve application performance.  

A more specific outline of the process includes the following:  

Instrumentation: Elastic APM supports various methods for instrumenting your applications to gather data, ensuring flexibility to suit different environments and requirements. While the integration of the Elastic APM agent directly into your codebase is a common approach, collecting detailed metrics on application performance, there are alternative methods available. Specifically, Elastic supports the use of OpenTelemetry, a set of APIs, libraries, agents, and instrumentation that allow you to collect and send telemetry data (metrics, logs, and traces) to monitoring tools like Elastic APM. This approach allows for the use of OpenTelemetry agents and SDKs, which can be particularly useful in environments where the Elastic APM agent is not the preferred option. Furthermore, the OpenTelemetry Collector, configured as an OTLP (OpenTelemetry Protocol) endpoint, can be used to aggregate and forward telemetry data from your applications to Elastic APM without needing direct instrumentation with an APM agent. 

Data Collection: Through the alternatives provided by OpenTelemetry integration, data collection can extend beyond the traditional agent-based approach. When using OpenTelemetry agents or the OpenTelemetry Collector (OTLP Collector), your applications can still capture comprehensive performance metrics, traces, and errors. This includes information on transaction durations, database queries, external service calls, and more, similar to what’s collected via the Elastic APM agent. The OTLP Collector, in particular, offers a versatile and scalable method to gather telemetry data from various sources and forward it to Elastic APM. This flexibility ensures that regardless of the specific instrumentation strategy—be it direct agent integration or utilizing OpenTelemetry components—you can efficiently collect the necessary data to monitor and analyze your application’s performance effectively. 

Data Transmission: The collected data is transmitted securely to the Elastic Stack, where it’s stored in Elasticsearch, a distributed search and analytics engine.  

Data Analysis: In Elasticsearch, the APM data is indexed and analyzed using built-in APM-specific indices and mappings. This allows for fast and efficient querying and visualization of the collected data. 

Visualization in Kibana: The analyzed APM data is visualized in Kibana, Elastic’s data visualization platform. Kibana provides various pre-built dashboards and visualizations tailored specifically for APM data. These visualizations offer insights into application performance, error rates, transaction latency, and other key metrics. 

Monitoring and Alerting: Elastic APM also supports monitoring and alerting capabilities. Teams can set up custom alerts based on predefined thresholds or anomalies detected in the APM data. This allows for proactive monitoring and quick response to performance issues. 

Integration with Logging and Metrics: One of the key strengths of Elastic APM is its integration with the broader Elastic Stack. This integration allows APM data to be correlated with logs and metrics collected from the same infrastructure. By combining APM data with logging and metrics data, teams gain deeper insights into the root causes of performance issues and can troubleshoot more effectively. 

Overall, Elastic APM provides a comprehensive solution for monitoring and optimizing application performance. By instrumenting applications, collecting detailed performance data, analyzing it in Elasticsearch, and visualizing it in Kibana, teams can gain valuable insights into their applications’ behavior and ensure they meet performance expectations. 

Discover how our expertise in implementing and optimizing APM solutions, including Elastic APM, can transform your business. Let’s ensure your applications deliver exceptional user experiences and drive operational efficiency. Reach out to us to schedule a consultation and take the first step towards maximizing the potential of your IT infrastructure. 

Het bericht Improve application reliability and performance with Application Performance Monitoring (APM) from Elastic verscheen eerst op Elk Factory.

The demand for efficient and intuitive search capabilities has never been higher. Traditional keyword-based search methods, while effective for certain tasks, often fall short when it comes to understanding the nuances of language or visual information. This is where Elastic vector search steps in, offering a shift in how businesses approach information retrieval and recommendation systems. In this article, we will take a closer look at the different use cases in which vector search can bring added value. 

A. Why Vector Search Matters for Businesses

At its core, vector search is driven by the concept of similarity. Instead of relying solely on exact keyword matches, it understands the semantic context of data, allowing for more accurate and relevant results. This capability opens up a world of possibilities for businesses across various industries. 

1. Semantic search

E-commerce: Imagine a shopping experience where you’re not limited by specific keywords. With vector search, customers can find products based on their preferences, even if they express them differently. For instance, Shopify has launched a semantic search tool for its customers to make the client’s online shopping experience more intuitive. Read more here.

(Picture source: Shopify) 

Media & entertainment: Discovering content becomes a breeze as vector search analyzes audio or textual features, enabling recommendations based on mood, genre, or style. Think about Spotify, which makes playlists based on your listening behavior.  

(Picture source: Promoly)

Recommendation Systems: 

News & articles: Vector search goes beyond surface-level recommendations, understanding the underlying themes or sentiments of content to offer personalized suggestions. 

E-commerce: By considering diverse factors like demographics and browsing behavior, businesses can enhance product recommendations, leading to higher engagement and conversions.  For example, you are browsing through a clothing website and clicked on a pair of pants, then the website will recommend other similar pants or other clothing items which you might be interested in. 

(Screenshot source: Zalando)

Customer Service Chatbots  

Providing tailored responses to user queries becomes seamless with vector search, enhancing customer satisfaction and reducing response times. This personalized approach enhances customer satisfaction and significantly reduces response times, ensuring a seamless and efficient experience. 

Natural Language Processing (NLP) Tasks: 

From document classification to sentiment analysis, vector search streamlines NLP tasks by deciphering the meaning and tone of text, facilitating more efficient data processing. 

An example focused on document classification illustrates the following: Suppose we have a collection of news articles categorized as “Technology” and “Sports.” Using vector search, we represent each article as a numerical vector based on its content. When a new article is introduced, vector search compares it with existing articles and assigns it (automatically) to the category with the most similar articles, enabling efficient document classification.  

To illustrate this example, we searched for an article related to ‘Tiktok’ on Forbes, and automatically got related Tiktok news articles on their website:  

(Screenshot source: Forbes)

2. Image search

Vector search cannot only be used for semantic search use cases. Moreover, image similarity search can bring a lot of added value for different industries.

Image Search:
From identifying objects and scenes to aiding accessibility for visually impaired users, image search powered by vector technology transforms visual information into actionable insights. Let’s explore the example of PcFruit. In the agricultural industry, image search technology revolutionizes berry farming by enabling quick and accurate identification of berry varieties. By capturing images of berries and analyzing key features like shape and color, farmers can optimize harvesting schedules and improve inventory management. Explore the full case here. 

(Picture source: Brainjar)

B. Generative AI Integration

By harnessing the capabilities of generative AI alongside semantic search, businesses can streamline processes like customer service, document summarization, and information synthesis, boosting productivity and decision-making efficiency.  

Examples of a synergy: generative AI and semantic search 

Enhanced Customer Service: GenAI-driven question-answer solutions empower service desk employees and customers alike, improving response accuracy and efficiency.

Document Synthesis: Vector search combined with generative AI can synthesize findings from disparate sources, enabling quick access to relevant information. 

Information Summarization: By summarizing key insights from extensive research, businesses can make informed decisions faster, driving innovation and growth.

Automated Legal Research: Legal firms can utilize semantic search to sift through vast amounts of legal documents and precedents. By integrating generative AI, they can automatically generate briefs, summaries, or analyses based on specific case details, saving time and resources. 

Medical Diagnosis Support: Healthcare providers can employ semantic search to analyze patient records, medical literature, and diagnostic reports. Combined with generative AI, this technology can assist in generating differential diagnoses or treatment recommendations, aiding physicians in decision-making processes.

Financial Analysis and Forecasting: In the finance industry, semantic search can be used to extract insights from financial reports, news articles, and market data. When integrated with generative AI, it can assist in generating financial models, forecasting trends, and automating report generation for investment analysis or risk assessment. 

Conclusion: Embracing the Future of Information Retrieval  

Incorporating vector search into business processes isn’t just about staying ahead of the curve, it is about reshaping the way we interact with data. By embracing the power of semantic understanding and image recognition, businesses can deliver personalized experiences, streamline operations, and unlock new avenues of growth. As technology continues to evolve, the possibilities of vector search and its synergies with generative AI are limitless, promising a future where information retrieval is not just efficient but truly transformative. 

Het bericht Get ahead of the competition with AI verscheen eerst op Elk Factory.

Efficient search capabilities are important for businesses to extract meaningful insights and stay ahead of the competition. Traditional keyword-based search methods often fall short in delivering accurate and relevant results, leading to frustration and inefficiency among users. Discover more about the topic of semantic search, a technology that leverages the power of natural language processing (NLP) to revolutionize your search experience. In this article, we explore how semantic search works together with Elastic technology to deliver superior search capabilities and drive business success. 

Semantic search goes beyond simple keyword matching by understanding the context, intent, and meaning behind search queries. By analyzing the semantics of language, including synonyms, concepts, and relationships between words, semantic search engines can deliver more precise and contextually relevant results. This capability is especially valuable in industries such as government, e-commerce, healthcare, and finance, where precision and accuracy are very important.  

Hereby an example to visualize the key difference between keyword search and semantic search: 

A closer look at the differences between traditional “keyword search” and “AI-powered search results and answers” reveal the transformative potential of the AI-powered elasticsearch. We highlight the improved results delivered by AI-powered searches, making it clear why industries are increasingly adopting this technology. Note that the enhanced results come from out of the box elasticsearch functionalities: 

To find out more about this example, click here.  

Het bericht Enhancing search efficiency with semantic search and Elastic technology verscheen eerst op Elk Factory.

In the dynamic landscape of artificial intelligence, understanding the nuances between Generative AI, Large Language Models (LLMs), and their integration with company data is crucial for businesses seeking to innovate and thrive. With Elastic serving as the essential bridge between company data and Generative AI, let’s explore the distinctions between Generative AI and LLMs, and how Elastic facilitates the seamless integration of company data for enhanced insights and security. 

Het bericht Bridging the gap: Generative AI, LLMs, and Elastic’s role in private company data integration verscheen eerst op Elk Factory.

AI has been around for some time. Use cases were sometimes not achievable due to the complexity of the implementation or limits within AI or computing power. Elastic technology makes it easier to enjoy the benefits of AI. Elastic makes it less complicated for developers to implement semantic search, image search and more. The proof of the pudding is in the eating. That’s the reason why we did an experiment and developed image similarity in Elastic. In this blog we will talk about the how and most importantly the impressive results achieved from the experiment.  

This article explores Elastic’s application in image similarity searches, focusing on icons like the recycle symbol and the European letter ‘E’. Specific experiments with these symbols are conducted, alongside discussions on technical challenges and solutions. This article particularly highlights the promising results achieved with the recycle icon and the European letter ‘E’. The potential of Image similarity technology for future projects is promising, underlining its simplicity and effectiveness in embedding images and text for searching purposes.  

Fun fact: The implementation of this experiment with the Elastic stack took less than a day, including the repairments and adjustments that were needed during the development process.   

First we will explain the results in this article, afterwards we will provide the technical details how we realized the image similarity search. To conclude will provide some insights with respect to sizing. 

We have a dataset of 47 images to be precise:

When searching for an image, typing “European E” we receive European E’s as the first results, which is good.

Instead of using text terms to search for images, we also used one of these “European E” symbols to search by clicking “Find similar images”. This also provided the expected similar images as search results. 

Green recycle symbol 

In our dataset, we have several recycle symbols of which a few are green. We did a test searching for “green recycle”. This did in fact only return our green recycle icons in the dataset, which is a great result:

2) The technical journey: how we achieved results 

The start 

We started our exploration based on this informative blog post from Elastic. 

We meticulously followed the instructions outlined in Elastic’s blog post and GitHub repository. After thorough perusal of the README.md file, we proceeded to clone the repository and integrated the required model from Hugging Face into our cloud instance using Elastic’s eland from Github. Keep in mind that this model does not need to be used. Other models can be used as well however then the backend would need to be modified.

Modifying the code 

Encountering outdated packages during the process, we swiftly modified the requirements.txt file, ensuring compatibility and smooth installation. Additionally, we addressed image pixel limitations and fine-tuned settings for optimal performance. Below illustrates the improved requirements.txt file: 

Following these adjustments, the pip install process proceeded seamlessly. Additionally, the ‘.env file’ was updated with the necessary credentials for our cloud instance. However, while attempting to generate image embeddings and ingest them, a subsequent issue was encountered. This challenge, which will be elaborated upon shortly, came from an oversaturation of pixels within our images. To address this, a simple line addition under the imports section of the create-image-embeddings.py file sufficed. Note: The code makes sure there is no maximum image pixels limit so watch out with how you use this (Decompression bomb):  

The embedding has 512 dimensions. Keep in mind that this image went through the clip-ViT-B-32 model. This is a free public model and is sufficient for our use case. The interface has a search box, which when submitted sends text to elastic which goes through the model (clip-ViT-B-32-multilingual-v1) that was imported with Elastic eland. This happens in the background (Flask Backend) and therefore, there is no need to look at that in this demo application, nor will we look at the Flask backend in this blog post. The interface has an image upload field. This can be used to upload images and search for other images that are similar to the uploaded one. 

Size of Images 

As mentioned above our dataset consists of 47 images. These Images have a combined size of around 3.6Mb. When looking at the index with the embedded values, the size is 469.9Kb.   

Size of Text  

We compared the sizing of embeddings for images with embeddings for text. We used a dataset containing about 8000 documents and having a size of 17.7Mb in total. Each document has only a few lines of text. When ingesting this data for ‘text search’ into an index, the index takes 36.3Mb. When ingesting for semantic search, using the E5 model the index takes 119.8Mb. In case both indexes are used for example to use RRF (Reciprocal Rank Fusion), the total index storage is 156.1Mb. 

Why did the size decrease for the images but increase for Text 

This is because of the dimensions and number of documents. The images index only had to keep 47 documents while the Text dataset’s index had to keep around 8000 documents. The Images dataset had the size of 9,99Kb per document While the Text index had around 14.97Kb per document. Which is relatively close to the Images index. If we take a look at the total vector dimensions, the Images index has 512 dimensions. However the Text index has around 768 dimensions, both the title and overview embedded fields have 384 dimensions. So after all it depends on how much dimensions are being used, if we used only one field to embed the size would greatly decrease, same can be said for the images index, if we used another model that generates more dimensions the size would greatly increase. 

If we go into even deeper detail, the images index has about 19.51bytes per dimension while the text index has 19.49bytes per dimension. As can be seen here they are about the same. So by this logic we could take 19.5b as average for each vector dimension. Keep in mind these are dense vectors, not sparse vectors.  

We were highly impressed with the ‘Image Similarity Search’ functionality, particularly appreciating the straightforward process of obtaining a model from Huggingface, embedding both images and text, and initiating searches.  

Elk Factory is the Elastic partner to implement your Elastic stack. We always strive for a win-win! Together, we’ll explore how this platform can make your business more efficient, so you can benefit while we gain a satisfied customer!

Get to know us, or feel free to contact us.

Het bericht Image similarity search with Elastic verscheen eerst op Elk Factory.

Het bericht LOG MANAGEMENT IN JUST 3 DAYS verscheen eerst op Elk Factory.