Khuyen Tran’s Post

To clean, gain insights, and create embeddings from massive unstructured text datasets, use Galatic

🙌🏼 Our latest issue (6.3) is out! 📖 For an in-depth preview of every article, start the issue by first reading the excellent editorial, "From COVID-19 to GPT-4o: The Groundbreaking Quinquennium for Harvard Data Science Review (and Humanity)" by Editor-in-Chief Xiao-Li MENG. 🧐 Check it out now! https://lnkd.in/eN3iPPT5

Just finished the course “Artificial Intelligence Foundations: Machine Learning” by Kesha Williams! Check it out: https://lnkd.in/gq5ujSuQ #machinelearning.

Just finished the course “Artificial Intelligence Foundations: Machine Learning” by Kesha Williams! Check it out: https://lnkd.in/gs-rBmBM #machinelearning.

Thinking about what to watch next? 📺 Superlinked got your back 😅 Check out the recording of Mór's hands-on workshop on building RAG solutions with complex data, in collaboration with Data Science Festival 🤓 Learn how to - 📖 Embed timestamp, numeric “helpfulness” rating and unstructured text data 📖 Apply weights at query time to increase the quality and consistency of the results 📖 Set up a RAG system from your data to power a chatbot 📖 Connect to a vector DB and a simple deployment to run your application Leave your questions in the comments 🙌 #datascience #RAG #LLM #GenAI #webinar

Exploring t-SNE vs. SNE in Simple Terms When it comes to visualizing complex data, t-SNE and SNE are two popular methods. Let's break down the differences: 🔹 SNE (Stochastic Neighbor Embedding): ▪ Keeps Neighbors Close: SNE focuses on keeping similar data points close together. ▪ Slower for Big Data: Takes more time and power to process large datasets. ▪ Crowding Issue: Can struggle to show data clearly when reduced to 2D or 3D. 🔹 t-SNE (t-Distributed Stochastic Neighbor Embedding): ▪ Better Visuals: Uses a special method to prevent the crowding problem, making clusters more distinct. ▪ Faster with Large Data: More efficient at handling bigger datasets. ▪ Widely Used: Preferred for creating clear and insightful visualizations of data. Why Choose t-SNE? 🔹 Speed: t-SNE is quicker for large datasets. 🔹 Clarity: Produces clearer and more meaningful visuals. 🔹 Ease of Use: Simple to implement and understand. #DataScience #MachineLearning #DataVisualization #tSNE #SNE #DataAnalysis

🚀 Exciting news! Check out our latest blog post on "Relational Graph Convolutional Networks for Sentiment Analysis" on arXiv:2404.13079v1. As textual data continues to grow, sentiment analysis becomes key for extracting insights from user-generated content. Our post delves into leveraging Relational Graph Convolutional Networks (RGCNs) for sentiment analysis, capturing complex relationships between entities. Don't miss the opportunity to explore how RGCNs offer interpretability, flexibility, and effectiveness in capturing relational information for sentiment analysis tasks. Dive into the full post here: https://bit.ly/4b7hjal.

The response to our launch of Alchemer Pulse has been amazing. Researchers and other survey builders understand the power of automatically tagging open-ended feedback with sentiment and theme. Check out how AI-powered text analysis can help you understand unstructured feedback at scale: https://lnkd.in/ggYy6qex

Our Topological Data Analysis for Machine Learning tutorial with Baris Coskunuzer is now available. We plan to expand this into a book eventually. Whether you are looking for a reference for your topology course or considering integrating a section into your machine learning class, our tutorial and codebase will be invaluable resources.

"This is achieved by organizing the graph into multiple layers, with each layer representing a different granularity of the dataset. The top layers contain fewer nodes, which represent clusters of data points, while the bottom layers contain individual data points." #knowledgegraphs #artificialintelligence #searchsystems #HNSWgraphs #HNSW #searchalgorithms #vectordatabases