Ed Sandoval

Knowledge Graphs and Large Language Models Dive into our latest blog, “Knowledge Graphs with Large Language Models,” where we explore how combining KGs with LLMs can revolutionize semantic understanding and contextual reasoning. Highlights include: -How KGs enrich LLMs’ contextual understanding -Techniques for integrating KGs with LLMs -Real-world applications and future potential Read the full blog to see how this cutting-edge fusion is shaping the future of AI! https://lnkd.in/gijszfnU #AI #KnowledgeGraphs #MachineLearning #GenerativeAI #DataScience #Innovation

2

👀 Constructing #KnowledgeGraphs from unstructured text is possible using LLMs. In this blog, Noah explores a simple yet powerful approach for 👆 in 3 steps: 👉Extracting nodes and edges from the text using LLMs. 👉Performing entity disambiguation to merge duplicate entities. 👉Importing the data into Neo4j to store and analyze the knowledge graph. https://bit.ly/4aPPaFd #Neo4j #LLMS #dataanalytics #graphdatabase Noah Mayerhofer

79

3 Comments

#Knowledge is at the heart of your business—knowing better means doing better. That's the power of integrating #ReasoningAI (and not just #GenAI) & #KnowledgeGraph into your data and content strategy. And it's not just us saying it; Gartner's experts agree! 😉 #DataStrategy #KnowledgeGraph #KnowledgeOperatingSystem #CognitiveAssetManagement

6

📖 Good Reads for today The article by Ben Lorica 罗瑞卡 and Prashanth Rao offers a decision framework for evaluating vector search solutions based on deployment models, scalability, performance, data governance, and other factors. LanceDB stands out as the only vector database meeting all criteria for handling large-scale data in a high-performance manner, suitable for modern AI/ML workloads. LanceDB is built on our team's decades of experience with high-performance Big Data systems in the cloud. We created the Lance columnar format, a radical-thinking groundbreaking advancement for the modern AI era, marking the first widely adopted new columnar format since the decade-old Parquet and ORC formats. The Lance format incorporates lessons from running AI data lakes, offering features such as fast random access and multi-modal support. This compute-storage separation, inspired by data lakes, enables efficient horizontal scaling of stateless LanceDB services. By embracing cloud-native object-store-based vector search, LanceDB achieves a superior performance/cost ratio compared to memory-based solutions. Our customers love it. We chose Rust to build LanceDB because we value our developer community. From indie developers to large teams, you can easily embed LanceDB in your applications, store data on Amazon S3 or Google GCS, and package your application in Kubernetes. Start any number of pods, and you'll see throughput increase linearly with consistent latencies. We've worked hard to meet some of the most challenging deployments, returning millions of vectors per second in milliseconds over the internet. Leading AI companies use LanceDB not only for vector search but also for serving images, videos, and text. We serve multi-modal data within the same simple query interface, not just embeddings. For our developer friends who want to try out Lance and LanceDB with open source, Lance v2 is available in beta, and we encourage you to try it out. https://lnkd.in/gp_bWbTV

17

1 Comment

Which Sentiment Analysis Approach is Right for Your Contact Centre: NLP-Based or Generative AI? 📊 John Matthew Ortiz at MiaRec compares the latest two approaches to achieving Sentiment Analysis: Natural Language Processing or Simple Language Model-based versus Generative AI or Large Language Model-based Sentiment Analysis. Read his comparison here 👉 https://lnkd.in/dnNiCnUh #AI #CX #ContactCentres

7

🎙️ How do the statistics and ML communities differ? What does current research in trend filtering and conformal prediction look like? Daniel Bashir speaks with Professor Ryan Tibshirani.

6

1 Comment

EK's Lulit Tesfaye and Jeff Jonas of Senzing teamed up on EK's latest blog, in which they discuss typical problems that would necessitate Entity Resolution, how to unify enterprise data with a Semantic Layer, and the role of entity-resolved knowledge graphs. #semanticlayer #knowledgegraph #entityresolution https://lnkd.in/ePujrbve

9

🔥 In our latest research, we've used cutting-edge Deep Learning (DL) techniques, specifically using Transformer models - the same powerful architecture behind today's most potent Large Language Models, like GPT, to forecast heat demand over the next 24 hours with high accuracy. Key Insights: 🔍 Transformer models for sequence forecasting. 🌤️ Custom feature engineering, including weather data. ⚙️ Data from 38,710 time points over five heating seasons (2019-2024). ➡ MAE without weather data: 28.15 KWh (outperforms LSTM benchmark). ➡ MAE with weather data: 21.09 KWh. ➡ Consistent performance with a standard deviation of just 0.23 KWh. This research demonstrates how modern DL models can revolutionize DHS operations, making them more efficient and sustainable. 🌍 Read the full story on Medium and discover how we’re shaping the future of urban heating solutions. 👉 https://lnkd.in/diVe6jT6 Follow XAI4HEAT on LinkedIn for more cutting-edge research and insights. 🌟 #DeepLearning #Transformers #DistrictHeating #UrbanInfrastructure #AI #MachineLearning #Sustainability #Innovation #fondzanauku #prisma Milan Zdravkovic Marko Ignjatovic, PhD

10