Viswanatha Swamy
Applied Information Sciences
DESCOBRIR, CONECTAR, CRESCER
Pronto para começar a usar a IA? O Microsoft Reactor fornece eventos, treinamento e recursos da comunidade para ajudar startups, empreendedores e desenvolvedores a criar seus próximos negócios em tecnologia de IA. Junte-se a nós!
DESCOBRIR, CONECTAR, CRESCER
Pronto para começar a usar a IA? O Microsoft Reactor fornece eventos, treinamento e recursos da comunidade para ajudar startups, empreendedores e desenvolvedores a criar seus próximos negócios em tecnologia de IA. Junte-se a nós!
25 fevereiro, 2025 | 10:00 AM - 11:00 AM (UTC) Tempo Universal Coordenado UTC
Tópico: IA de núcleo
Linguagem: Inglês
This session will cover the process of building custom named entity recognition (NER) models using Azure AI Language. Attendees will learn how to define and tag custom entities, train models to extract these entities from unstructured text, and evaluate model performance. Practical exercises will provide hands-on experience with the Azure AI Language platform.
How do you think a Start-up will be benefited from this session?
Start-ups can leverage custom NER for a variety of use cases:
Automating data extraction from documents: Extract key information like product names, dates, locations, and contract details from invoices, contracts, and other documents.
Improving customer support: Identify specific product mentions or issue types in customer feedback to route inquiries to the appropriate team.
Enhancing search functionality: Index documents based on extracted entities to provide more relevant search results.
Analyzing social media: Extract brand mentions, product names, and other relevant information from social media posts.
By automating these processes, start-ups can save time and resources, gain valuable insights from their data, and improve operational efficiency.
Read More - https://aka.ms/Entity-Recognition
The session will focus on:
Understanding custom named entity recognition: What custom NER is, and when to use it.
Tagging entities in extraction projects: Best practices for defining and annotating custom entities in text data.
Understanding how to build entity recognition projects: Project setup, data preparation, and model training in Azure AI Language.
Training and evaluating your model: The training process, model evaluation metrics (precision, recall, F1-score), and strategies for improving model performance.
Practical Examples and Hands-on Exercise: Attendees will work on a practical scenario, such as extracting information from product descriptions or legal documents.
Speaker BIO- Viswanatha Swamy
He is an aspirant Software Architect and currently, he works at Applied Information Sciences. He is passionate about C#, ASP.NET, Azure, Performance testing/tuning,.NET Core, and Docker. He loves to learn about new technologies.
Social Handle- https://twitter.com/vishipayyallore
AI-900
Some experience in AI
Willing to learn AI
Palestrantes
Já tem registro e precisa cancelar? Cancelar registro
O evento faz parte de #AI102 Series.
Clique aqui para visitar a página da série no qual você pode ver todos os eventos futuros e sob demanda.
Para perguntas, entre em contato conosco em reactor@microsoft.com