Tallan's Blog

Tallan’s Experts Share Their Knowledge on Technology, Trends and Solutions to Business Challenges

Posts Tagged "Machine Learning"

Machine Learning Getting Started Guide

Samantha Francis

It’s taken a year for me to feel confident enough to even chime in, on a high level, about the products we’ve created, and the platforms we utilize.  I can dabble in conversation about chatbots and Microsoft’s Cognitive Services.  I understand now, more or less, what ‘the cloud’ is and its benefits.  But, this is why teamwork makes the dream work, you know.  My colleagues can build you a solution to any business challenge. Anything.  You’ve got a problem, they’ll solve it.
But, now it’s my turn. I am going to express why what they can do matters.
You’ve all heard of Machine Learning.  We partnered with RetailWire to produce a Webinar on ML for Retail back in April and that’s where my understanding really began to take shape.  In a nutshell, Machine Learning can be set-up and do in minutes and…

Smart Chatbots are Relegating the FAQs Page to History

Tallan Partner

Whether you call them conversational agents, dialog systems, or chatbots, AI-powered bots that can hold human-like conversations are seeping into our everyday lives.
Chatbots work well in a structured environment with a predetermined dataset. Answering simple questions, for example, would be a task a chatbot could excel at. Which is why chatbots are now replacing the Frequently Asked Questions page on websites.
Here’s what you need to know about how these chatbots work and why you might never see a traditional FAQ page again:
Why Q&A?
Chatbots can be either retrieval-based or generative, which means they can either retrieve data from a predetermined dataset or generate new responses from scratch. These bots can also be open or closed domain, depending on whether the user can take the conversation anywhere and still expect a reply or whether a user needs to stick to a narrow…

Microsoft is Teaching Systems to Read, Answer and Even Ask Questions

Tallan Partner

From left, Rangan Majumder, Yi-Min Wang and Jianfeng Gao on Microsoft’s Redmond, Washington, campus. Photo by Dan DeLong.
Microsoft researchers have already created technology that can do two difficult tasks about as well as a person: identify images and recognize words in a conversation.
Now, the company’s leading AI experts are working on systems that can do something even more complex: Read passages of text and answer questions about them.
“We’re trying to develop what we call a literate machine: A machine that can read text, understand text and then learn how to communicate, whether it’s written or orally,” said Kaheer Suleman, the co-founder of Maluuba, a Quebec-based deep learning startup that Microsoft acquired earlier this year.
The Maluuba team is one of several groups at Microsoft that are tackling the challenge of machine reading. Two other research teams, one at the company’s Redmond, Washington, headquarters and the…

Data Virtualization: Unlocking Data for AI and Machine Learning

Tallan Partner

For reliability, accuracy and performance, both AI and machine learning heavily rely on large sets. Because the larger the pool of data, the better you can train the models. That’s why it’s critical for big data platforms to efficiently work with different data streams and systems, regardless of the structure of the data (or lack thereof), data velocity or volume.
However, that’s easier said than done.
Today every big data platform faces these systemic challenges:

Compute / Storage Overlap: Traditionally, compute and storage were never delineated. As data volumes grew, you had to invest in compute as well as storage.
Non-Uniform Access of Data: Over the years, too much dependency on business operations and applications have led companies to acquire, ingest and store data in different physical systems like file systems, databases and data warehouses (e.g. SQL Server or Oracle), big data systems (e.g….

r-7

Adding R Packages In Azure ML

Iu-Wei Sze

The Execute R Script module in Azure Machine Learning is incredibly useful for manipulating data in ways that other modules do not cover. Its functionality can be further expanded by adding R packages that are not included in Azure ML by default. We will first show you how to get a list of packages that are already in your workspace and then how to add additional packages.
Checking Which R Packages are in Your Workspace
Create a new experiment, and place the following R code in an “Execute R Script” module:
data.set <-data.frame(installed.packages());
maml.mapOutputPort(“data.set”);

Run the experiment. The output of the Execute R Script module will be a list of the available packages.
Adding R Packages
Before you can use the package in Azure ML, you need to set up the zip file structure in which ML expects the packages to appear. To do this, start by installing the…

tips-4

Azure ML Tips And Tricks

Lu Li

While working with Azure Machine Learning, we ran into some situations that seem simple to handle in hindsight, but are rarely discussed online. Here are some simple Azure ML tips and tricks that we found useful.
1. Saving and Reusing a Trained Model
In one of our projects, we found that we needed to use the same trained model multiple times rather than just once in a predictive experiment. To do this, save a trained model by right clicking the output port of the Train Model module and selecting the Save as Trained Model option.

After giving the trained model a name and saving it, you will be able to find a module with the given name under the Trained Models sections of the module selection bar

or by typing the name into the module search bar. From here, you will be able to use…

headdesk

16 Things Only People Who Work with Azure ML Will Understand

Lu Li

Azure Machine Learning is part of the Cortana Intelligence Suite. It’s a cloud based collaborative drag and drop tool that can be used to build, test, and deploy predictive analytics solutions. We recently worked with Casella Waste Systems to analyze their customer and sales data using Azure ML. We found Azure ML to be a useful tool that allowed us to visualize our work and avoid coding when it wasn’t necessary. However, the process of creating successful experiments didn’t come without some speed bumps. Here are some things that we discovered along our ML journey, delivered in the spirit of a Buzzfeed article.
1. Your experiment starts out nice and simple

2. And ends up looking like this

3. You have multiple people working on an experiment

 
4. But you see this when everyone tries to edit the experiment at the same time

All that…

\\\