Blog Post

Educator Developer Blog
3 MIN READ

How University Texas Austin Students Study Machine Learning with Microsoft Azure

Lee_Stott's avatar
Lee_Stott
Icon for Microsoft rankMicrosoft
Jun 11, 2021

Written by Patrick Chao, the Teaching Assistant for the course


 

In the spring semester of 2021, 19 students from The University of Texas at Austin took the Introduction to Machine Learning course taught Patrick Chao, Teaching Assistant for the course and Professor Danna Gurari

   

 

Before taking the course, nearly half of the students had no programming background.  By the end of the course, the students learned the foundations of Machine Learning and gained hands-on experience developing many machine learning models for different problems in Python. As part of the course, the students also paired up to develop final projects of their own design, with many groups achieving their research goals using the power of Microsoft Azure. 

 

 

In one of team’s final project, their goal was to solve a problem for small family-owned businesses with Optical Character Recognition (OCR) powered by Azure. The problem they aimed to address is that many business owners must manually enter data from printed or handwritten tickets to spreadsheets, which can be time-consuming. In the experimental results, the team explains that there is lower accuracy for tickets of certain types and provide possible solutions to improve the performance.

 

Another team aimed to develop a model for poetry classification based on time period. The problem is that it is not easy for human readers to grasp the characteristics of poetry in all periods. Moreover, the classification of time periods is an intense debate topic in the fields of humanities. The team used natural language processing techniques to analyze the title and content of poetry. With text analytics in the Azure Cognitive Services, they extracted key phrases from each poem as a helpful feature to train their model. Ultimately, their model could distinguish the period of poetry better than most English majors. The next step is to understand how the model can support the study of English literature today.

 

Lastly, another team aimed to address that vaccine hesitancy is a problem, and it slows down the speed of the population receiving the vaccine. They proposed a chatbot to answer questions and lessen worries. The chatbot detects a person’s sentiment from input questions and returns the appropriate information in a way that is tailored towards the person’s sentiment. Azure Sentiment Analysis API is used to detect the sentiment.

 

Detailed documentation and an easy-to-use interface are big advantages of Azure. After a brief introduction to Azure, most students could explore the Azure services on their own and leverage the services in their final projects. 

Interested in Learning more see the following Microsoft Learn Modules 

How to build a basic chatbot - Learn | Microsoft Docs
Create Intelligent Bots with the Azure Bot Service - Learn | Microsoft Docs

Build a chat bot with the Azure portal - Learn | Microsoft Docs

Discover sentiment in text with the Text Analytics API - Learn | Microsoft Docs

Explore Natural Language Processing in Microsoft Azure - Learn | Microsoft Docs

Evaluate text with Azure Cognitive Language Services - Learn | Microsoft Docs

Process natural language with Azure Cognitive Language Services - Learn | Microsoft Docs

Introduction to Natural Language Processing with PyTorch - Learn | Microsoft Docs

Read Text in Images and Documents with the Computer Vision Service - Learn | Microsoft Docs

Updated Jun 11, 2021
Version 2.0
No CommentsBe the first to comment