Тёмный

NODES 2023 - Graph Machine Learning for 2024 

Neo4j
Подписаться 51 тыс.
Просмотров 2,2 тыс.
50% 1

This session will teach you how to leverage modern, relevant, and practical graph machine learning (GML) approaches to solve common data science/ML business problems. Zach will provide a technical overview of GML and walk you through modern techniques for generating embeddings and building supervised learning workflows, including link prediction and node classification tasks. He'll also touch on integrations with Generative AI to enrich/improve AI capabilities with GML. The use cases and applications of GML are fairly broad, including recommendation systems, fraud & anomaly detection, disambiguation, and search.
Neo4j GenAI: bit.ly/4eRHS6g

Опубликовано:

 

2 окт 2024

Поделиться:

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист
Посмотреть позже
Комментарии : 10   
@podunkman2709
@podunkman2709 7 месяцев назад
You do not need any "open AI" to get that. You have SBERT package and open embedding models - including this one presented here. I was expecting a lot from "vector search" but after several tests I'm very disapointed - results are really poor. U can have better similarity using FullText search (for example with solr library). We can say that vector search does not work at all with asymetric search; short text as parameter, seach in longer sentence. Even if u apply special model for such search.
@SynonAnon-vi1ql
@SynonAnon-vi1ql 7 месяцев назад
Great presentation! Can you please share the git for the demo?
@neo4j
@neo4j 7 месяцев назад
See here: github.com/neo4j-product-examples/ml-genai/tree/main/retail-hm
@SynonAnon-vi1ql
@SynonAnon-vi1ql 7 месяцев назад
You see, intuitively I think I am convinced that graphs are useful. And I am leaning towards them. But algorithmically and from examples point of view I am yet to come across an example in ML and GenAI that is made possible / more accurate by graphs but not with the help of traditional SQL database (or tediously impossible / far less accurate). Can you provide such example(s)? Is there any comparative study of (im)possibilities / (in)accuracies between the two approaches?
@neo4j
@neo4j 7 месяцев назад
We have collected a couple of resources in our GenAI Landing Page: neo4j.com/generativeai/
@kevon217
@kevon217 8 месяцев назад
Stellar overview!
@neo4j
@neo4j 7 месяцев назад
Thank you
@mikehatch4478
@mikehatch4478 8 месяцев назад
Are the original presentation decks from Nodes 2023 available somewhere? I searched around and didn’t find them.
@neo4j
@neo4j 7 месяцев назад
Not all did share the decks with us. Lets see if we can get this one
@neo4j
@neo4j 7 месяцев назад
Not all did share the decks with us. Lets see if we can get this one
Далее
🦊🔥
00:16
Просмотров 393 тыс.
Учёные из Тринидад и Тобаго
00:23
Лучше одной, чем с такими
00:54
Просмотров 754 тыс.
一分钟学懂 什么是知识图谱
1:17
Graphs for AI and ML
45:11
Просмотров 23 тыс.
Graph databases: The best kept secret for effective AI
23:07
Ontologies for Enabling Automated Graph   LLM Workflows
28:16
Yann LeCun - Gen AI Winter School, Objective Driven AI.
1:14:29
🦊🔥
00:16
Просмотров 393 тыс.