Wednesday, November 24, 2021

GPT-3 Now Available to All Developers

OpenAI has just made GPT-3 generally available to all developers (in supported countries). What is GPT-3? Wikipedia defines Generative Pre-trained Transformer 3 as an autoregressive language model that uses a deep learning neural network to produce human-like text. In other words, GPT-3 can automatically write news stories and articles, create blog posts, compose poetry, and more, in addition to generating working program code in more than a dozen programming languages. GPT-3 is also being used to develop conversational models which can respond to comments or questions with replies that fit the context of the conversation.

The potential for GPT-3 continues to grow with new applications showing up almost every day. For instance, one developer has combined Figma's user interface prototyping with GPT-3 to create websites just by describing the desired type of website in a couple of sentences. Another example of GPT-3's flexibility is its ability to convert program code from one language to another using CodeX, GPT-3's code generator. Mark Ryan has even been able to use CodeX to convert old COBOL legacy code into usable JavaScript routines (https://towardsdatascience.com/codex-translates-cobol-into-javascript-266205d606c0). You can find dozens of other examples of what's possible with GPT-3 by going to:https://gpt3demo.com/. 

There are a number of known weaknesses in GPT-3 though: 

  • The model is pre-trained - it doesn't continue to learn from each interaction.
  • If the Internet material that was used to train GPT-3 included biased content, GPT-3 will tend to exhibit those same biases.
  • GPT-3 can only accept a limited amount of input for each interaction, so the input has to adequately describe the desired result.

Plus GPT-3 also has a dark side. According to an article on VentureBeat.com, "the Middlebury Institute of International Studies’ Center on Terrorism, Extremism, and Counterterrorism found that GPT-3 can generate “influential” text that could radicalize people into far-right extremist ideologies. A group at Georgetown University has used GPT-3 to generate misinformation, including stories around a false narrative, articles altered to push a bogus perspective, and tweets riffing on particular points of disinformation."

OpenAI does have guidelines that specify what type of content GPT-3 can and can't be used to create and filters to try to detect models that violate those guidelines. However, like many other advances in technology, it's still an open question as to whether tools like GPT-3 can be supervised effectively and what their overall effect will be. 

So what's the future of GPT-3 and similar deep learning models? The best answer  is that OpenAI and other organizations specializing in neural network applications are already hard at work attempting to develop even more advanced deep learning models. There's no turning back at this point, so hopefully GPT-3 and its successors will avoid the "AI is going to take over" stigma and turn out to produce more benefits than problems.

Wednesday, November 17, 2021

Glide Pages

 Glide (www.glideapps.com) has been one of the top choices for anyone looking to create mobile apps without coding, but they have just added the "missing piece" to their platform. With the introduction of "Glide Pages", you now have the ability to create responsive web apps and websites in addition to native mobile applications. 

You build Glide "pages" by dragging and dropping "Collections" onto the design screen. There are different types of Collections for different types of data - "cards" for images and "grids", "table b" and "lists" for tabular data. You can also attach "actions" to collection objects, items within a collection and other design components. Actions include updating records, showing notifications and calling webhooks (webhooks are a mechanism for applications to communicate with each other). 

Since apps created with Glide Pages are actually web-based, it's easy to provide people with access to your application just by setting up a web "portal" and providing users with the URL for that web page. Setting up a portal also allows you to create a secure environment by requiring user logins, where each user only has access to the data specified for their particular group or role.

You can find more detailed information on Glide Pages at:

"https://docs.glideapps.com/all/courses/introduction-to-pages"  and 

"https://www.youtube.com/watch?v=qhGbbUUFzVk"


Wednesday, November 10, 2021

Data Cleaning with No-Code

What is “data cleaning” and why do it?

Almost every organization has missing, incomplete, duplicated, inaccurate or otherwise unreliable records in their spreadsheets or database. In order to work with that data to produce meaningful information those records have to be repaired or deleted - “cleaned” in other words. Doing that by hand is time-consuming, error-prone and expensive and creating a program to do the work requires coding expertise.  Now however, there are no-code platforms that can help businesses repair their data automatically, without doing any program coding. 

What is “cleaned” data used for?

In recent years more and more organizations are using their accumulated historical data to create data visualizations and to train machine learning models. Errors in the input data due to missing, incomplete or invalid fields will lead to invalid or inaccurate results from machine learning models trained on that data.

How does a no-code data cleaning site work?

Generally you import your raw data, make sure all the columns are showing with the correct type of data in each column, and then you select to start cleaning. At that point you can normally set up cleaning rules or "transformations" for each particular column. A cleaning rule might be "Remove duplicate values" or "Replace any missing values with the word 'Unknown'" or "Remove any rows with a value over $999 in this column". Once the data has been cleaned you can visually spot check it to make corrections manually or to make sure that you didn't leave out any necessary transformations. Then you should be able to export the cleaned data as a CSV file that can be used as input for a "no-code" machine learning platform.

Where can I find a no-code data cleaning platform?

There are a number of platforms that allow you to clean raw data without needing to do any coding. Probably the most complete platform at the moment is Amazon Web Services' Glue DataBrew; you can find a detailed explanation of how it works and what options you have available with it at:

https://aws.amazon.com/blogs/aws/announcing-aws-glue-databrew-a-visual-data-preparation-tool-that-helps-you-clean-and-normalize-data-faster/ 




Wednesday, November 3, 2021

Walmart Joins the No-Code Revolution


Walmart is in the process of acquiring Botmock, a no-code startup that has developed specialized software that allows anyone with a little technical knowledge to build and deploy conversational apps. Botmock's interface creates program code automatically as the user constructs conversational flows.

Walmart's customers rely more and more on being able to shop using voice and text apps, which normally take a lot of trial and error to develop. Cheryl Alnoa, a senior vice president with Walmart presented this example:

“Building seamless interactions for voice or chat is a rather difficult design problem where we have to consider all kinds of conversational flows that depend on the individual situation and customer needs,” said Ainoa. “For example, if a customer uses their voice to create their weekly shopping cart, they could say, ‘Add milk to my shopping cart.’ The correct action and reaction to the customer will depend on several factors, including if the customer has bought milk in the past, what type of milk they prefer (e.g. 2%, non-fat, etc.). Do you already have a type of milk in your shopping cart? If so, should we ask if you want to change the quantity or let you know that you already have it in your cart? “

With Botmock's technology, applications like this one can be put together in days rather than weeks or months. Considering the savings in time and effort that are possible now, Walmart hopes to have the new voice and text app features available to all the companies in the Walmart family in the very near future.