Tech Friday

Can Artificial Intelligence understand us better than we understand ourselves?

AI is now the main power of public healthcare systems transformation. As a result of AI-related technology, self-care became easier and more accurate than ever before.
More and more of us start using a variety of gadgets and apps that allow tracking health status. 

Some companies take it to the next level. This week we chatted with a nice guy from South Korea, who is working on a contactless health analysis gadget. The app's name is CardiVu, and their main tech is called Automatic Vital Sign Extraction Algorithm, promising over 90% of the accuracy of vital signs.

It appears that regular cameras from your smartphone or laptop can reveal your stress level just by analyzing your eye. Particularly, by the way, your iris moves. You don't even have to interrupt your work for the analysis - just keep doing what you’re doing, and the app will tell you how stressed you are. Oh, and something about your health as well.

So, how does it really work? 

As usual, it starts with collecting data. Lots of data. And finding out what’s the correlation between the vital signs and human eye iris micromovements. The more data we gather, the more accurate the results are. The data monitors include heartbeat rate, fluctuations, low-frequency to high-frequency band ratio, and others.
Turns out, when the iris is too calm, you’re in trouble.

And sometime later these measurements are applied to you, and you can learn a lot about your current self. If you’ve been using the app for a while, you can also see the history of your past measurements.

Not bad for contactless technology, right?

Heart Rate Variability, or HRV, remains one of the core indicators. The Smart Diagnosis researchers also believe that “HRV has been proven to be a very powerful biomarker”. Also, it’s an accurate, non-invasive measure of the Autonomous Nervous System.

Healthy people are supposed to experience large and complex heart rate changes, but complexity is significantly reduced if they have a disease or stress.

Having got the approval of KFDA, Smart Diagnosis went further and applied the algorithm to the education industry. 
The rise of online educational courses & platforms leads to a question on how to measure the involvement and attentiveness of students.
Even if students’ cameras are turned on, who knows what’s going on in their head, do they really try to grasp the material? Or maybe there’s a game open in a separate browser window?

Of course, it’s hard to achieve involvement similar to personal lessons and discussions, but we could at least start by tracking it. 

In this case, the camera is more focused on gaze tracking, as opposed to iris contractions. The app tracks time spent with focus, as well as time spent without focus on the screen. The app can also show us a heat map of the student's attention, highlighting the main areas of the screen where the student was focused on most during a certain time.

Some other tracked factors are blinking frequency, pupil reaction, and head position. 

What do you guys think about technology? Is it useful? Dangerous? Ethical enough? Would you use it if you could? These guys are currently in the alpha stage, so we still have some time to think about the options before it hits the market.

While people are used to paying attention to the tone and meaning of the words that have been heard, developed AI-based algorithms are being taught to detect connections between words themselves and to see the irony and intentional lies.

Now intelligence officers/ agencies are ready to prove the ability of artificial intelligence to detect the human art of sarcasm. The tool, funded in part by the U.S. military, lets them analyze the social media content (including comments, posts, forum talks) while ignoring posts that shouldn’t be taken seriously. 

A couple of scientists from the University of Central Florida conclude the following: “Сertain words in specific combinations can be a predictable indicator of sarcasm in a social media post, even if there isn’t much context”. 

To simplify, there are some words - “triggers”, that get more attention from the system and help it to identify sarcasm. Thus, words as ‘just’, ‘again’, ‘totally’, or exclamation mark make it easier to recognize the suggestion of sarcasm in text. The algorithm learning base datasets include posts from social networks (Twitter, Reddit),  dialogues and headlines from The Onion and other news sources. 

So, we have here one of the most difficult forms of communication. Moreover, we`re dealing with the text format, which eliminates the voice cues. 

How can an algorithm handle this? 

Obviously, only neural networks can handle this. 

In a nutshell, AI is trained to give more weight to some words than to others, depending on what other words appear nearby. Practical significance is in helping the military to understand what’s happening in key areas where they might be operating.

The work was supported by DARPA (the Defense Advanced Research Projects Agency). They initiated and launched the program to seek a “deeper and more quantitative understanding of adversaries’ use of the global information environment”.

Scientists are now trying to take into account all the results of previous attempts to recognize human emotions. In previous releases of the podcast, we have been discussing the brightest of them.

This attempt was preceded by a more in-depth study, which improved on the previous ones. The main difference lies in the approach to the search for trigger words.

Earlier algorithms were developed to seek words suggesting specific emotions or even emojis.  As a result, they skipped expressions that did not contain them. Consequently, most sarcastic comments were skipped.

Neural networks were also used before and tend to perform better. However, it was still impossible to define for sure how the neural network reached the conclusion that it reached. 

Finally, research has gone a little further and AI has come to understand us a little better. What "challenge" will be next in this direction? Is it possible to teach algorithms how to analyze slang and spoken words soon? Can it also predetermine the mentality of the commentator depending on his remarks?

And, on a more personal level, do you think the AI can detect your sarcasm?

Most Americans are unsure of how consciously companies behave while using and protecting personal info. Nearly 81% of them report being insecure about potential risks of data collection, and 66% claim to feel the same about government data collection. 

It`s really difficult to weigh the level of potential risks and understand the anticipated harm that irresponsible behavior with personal data can cause. How can this affect our further actions, what restrictions and changes will it bring?

Let`s analyze the main points of what happens when disclosing personal data and how to properly protect yourself using the example of the recent situation in South Korea.

Korean company ScatterLab launched a “scientific and data-driven” app, which was supposed to predict the degree of attachment in relationships.

In December 2020, the company introduced an A.I. chatbot Lee-Luda. 

The bot was positioned as a well-trained AI consultant, taught on more than 10 billion conversation logs from the app. “20-year-old female” Lee-Luda is ready to set a true friendship with everybody. 
As the company`s CEO mentioned, CEO the purpose of Lee-Luda was to become “an A.I. chatbot that people prefer as a conversation partner over a person.”

Just after a couple of weeks of the bot launch users could not help but pay attention to the harsh treatment and statements from the bot towards certain social groups and minorities (LGBTQ+, people with disabilities, feminists, etc.).

The developer company, ScatterLab, explained this phenomenon by the fact that the bot took information from the basic dataset for training, not from personal user discussions.
Thus, it is clear that the company did not properly filter out the set of phrases and profanity before starting the bot training.

Lee-Luda could not have learned how to include such personal information in its responses unless they existed in the training dataset. 
But there some “good news” as well: it is possible to recover the training dataset from the AI chatbot. So, if personal information existed in the training dataset, it can be extracted by querying the chatbot.

Still going not so bad, huh? 
To make things worse, ScatterLab had uploaded a training set of 1,700 sentences, which was a part of the larger dataset is collected, on Github. 
It exposed names of more than 20 people, along with the locations they have been to, their relationship status, and some of their medical information.

Despite the fact that this situation has become a high-profile event in Korea, it has not received attention on a global scale (and we think quite unfairly).
It's not about the negligence and dishonesty of the creators, this incident reflects the general trend in the development of the AI industry. Users of software based on technology have little control over the collection and use of personal data.
Situations like this should make you think about more careful and conscientious data management.

The pace of technology development is significantly ahead of the adoption of regulatory standards for their use. It is hard to foresee where the technology will lead us in a couple of years.

So, the glocal question is “Are AI and tech companies able to independently control the ethical component of the used and developed innovations?”.
Is it worth going back to the concept of "corporate social responsibility"? And where is this golden mean (Innovation VS Humanity)?

Any software is traditionally based on the programmer`s designed algorithm and is aimed to perform a specific function. But in the case of Artificial Intelligence software, we face taking decisions that can be unpredictable.

A Head of global software standards at Philips, Pat Baird, said: “For these machine learning systems, the programmer doesn't tell the software how to solve the problem”. 
In the particular case of AI programming, the result of development is the software being able to find patterns in the data. The reality is the programmer has no idea of the reasons software made a certain decision, he just threw together an engine that calculates a ton of stuff.

The guide of potential problems with AI software avoidance should start with something like “make sure to get good data before all else”. As Pat Baird says about the traditional software, 

“It's garbage in, garbage out. But what's going to happen is you have garbage in your data, and since you don't know how the software works, it's going to be a problem.”

So, what are the common problems we face using AI apps? Pad gave a couple of real-life examples of how bad-collected data can cause untrue results delivery. 

Example one - “Bad data to start with”
It demonstrates the inability of wearable devices to recognize the mode of travel while the tracker showed over 20,000 steps after an off-road adventure in a Jeep. “This was because of all the potholes and how much I was thrown around in the Jeep,” Baird said.

Example two - “Overfitting”
Here we are talking about the excessive data sets that also lead to inaccuracies. From a huge mass of able data and values, the program selects patterns for analysis and training that are not targeted ones for us. One of the great examples is the image recognition software, the goal of which was to determine the difference between an Alaskan Husky dog and a wolf. Baird commented it in the following way: “The data performed well, but it was picking up on background cues, rather than the ones the programmer intended. What actually happened, was that most of the photos that people had of their dog were taken sometime during the summer or fall in their backyard, whereas the photos of the wolves were taken during the winter out in the wild. The software that looked great to detect the difference between the Alaskan Husky and the wolf was actually picking up whether or not there was snow on the ground.”

Example three - “ Underfitting”
Another issue is appearing as a result of the collected data shortage. Such a simple reason, right..? As a rule, it ends up with making a decision based on noise, not on something real.

Autonomy level - questions to think about
One more question appears on share responsibility we are ready to give the technology. What autonomy level should it have? Will it be enough just to give you the right driving direction or we need a self-driving car that actually does the driving for us?
The good news is that most of the companies that worked with data collection have already found better ways to do it in terms of quality control. But still, can we remove errors in developing an excellent AI system and how should we deal with it?

So, how do we avoid failures in AI development?

Here are the very basic principles:

1. Collect relevant and verified data
2. Collect enough data
3. Make sure the data is diverse
4. Carefully consider if the data justifies the power we are giving to the AI

Fortunately, the development and implementation of technologies by leading companies provide an opportunity to learn from the mistakes of others and create a "cleaner" technology now.

Well, that’s all for now. Thanks for listening! Be sure to subscribe now to stay on top of the world community news with Zfort Group!

Scientists have recently published an application that will lead to conclusions about the use of AI by completely different categories of consumers:

• ordinary users - showing that they can be deceived by facial expressions, 
• critics - to see once again the violation of human rights. 

What did the app study let us uncover, why is it appeared so interesting to scientists and cause a public outcry?

While most users are passionate about testing the application and trying to prove it invalid, scientists try to draw public attention to the ethical side of the issue and provoke public debate.  Thus the researchers hope to unmask the reality of emotion recognition systems. Critics, in turn, argue that technology violates confidentiality rules, and is even racist.

To raise awareness of the technology and promote conversations about its use, researchers developed an emojify.info website - a program to try an emotion recognition system using PC or mobile camera.

Dr Alexa Hagerty, a scientist at Cambridge University who is also a project lead of Leverhulme Centre for the Future of Intelligence, commented on it in the following way: "These developments are based on one form of facial recognition. But technology has gone beyond that, not just by identifying people, but by asserting that it is capable of accurately counting internal experiences and emotions from our faces".

Let’s go directly to emojify.info website and see what functionality the application provides. In one of the ing, users are asked to make multiple expressions of faces on the camera and see if they have been deceived by the technology.

"The developers of this technology claim to read emotions," said Hagerty. In fact, the system can only count the movements of the face and combine with the assumption of what emotions may be behind it (for example, a smile = happiness).

Facial expressions do not always reflect true emotions, nor do people’s habits of accepting a particular expression. For example, smirks, sarcasm and other ambiguous states cannot be recognized by a robot (that is, not all people and not all people are able to recognize them.).

There is already some scientifically sound evidence that the expression of internal state is not quite as simple as the creators of such developments would like. 

Which one of us didn’t fake a smile trying to look cheerful at one of the holidays, huh?

Some scientists have already expressed their view that it is now time to pause the development of emotion recognition systems in the growing market.

What do you think - should AI specialists stop playing with fire now? 

Our opinion here in Zfort Group is, there could be potential dangers in almost any technology. If we as humanity were to stop exploring the world and science out of fear of something going wrong, we would be stuck in the Middle Ages forever. We are where we are only because we tame the technologies to serve us.

Let us know if you disagree - open to being wrong, as always.

Hello, this is TechFriday again. It’s great to have you with us!

Today we’ll talk about the challenge of teaching AI relationships or at least generating some pickup lines. A recent scientific experience showed us how AI works on the GPT-3 model in terms of relationship building.

It’s been almost 10 years since we were having fun asking Alexa and Siri. Smart assistants are already "used" not only to perform their core functions but also to being a full-fledged conversation buddy. Now, with the help of AI, scientists decided to go further and teach it the art of flirting.
Let’s see what came out of it and see if the scientists were satisfied.

The main idea belongs to Janelle Shane, a research scientist from Colorado, who implemented it into reality with the help of GPT-3.

How did Shane handle that?
Firstly, the article template was fed to 4 different GPT-3 program variants, which were supposed to enhance the starter file with generated predictions. 

Some of the example pickup lines were from Tinder DM screenshots and sound like
“I'm losing my voice from all the screaming your hotness is causing me to do”.
Another one says: “I will briefly summarize the plot of Back to the Future II for you”.

Let's start with the 1st AI, DaVinci, and its variants. Shane herself characterized it as “the largest and most competent” of the 4 AIs. DaVinci’s pickup lines bordered on cute ridicule, ice-humping sentences, and a few intricate phrases, issuing such gems:

• I love you. I don't care if you're a doggo in a trenchcoat.
• You have a lovely face. Can I put it on an air freshener? I want to keep your smell close to me always.
• Wait, this beanie hat, is it fashionable?
• You look like Jesus if he were a butler in a Russian mansion.

Curie, the 2nd candidate, which is not so powerful a piece of software,  created some poetic, mysterious game of words. The “best” variants included:

• Your eyes are like two rainbows and a rainbow of eyes. I can't help but stare.
• I like the ice cream… You can keep me in the freezer for a while but then I melt!
• Hey, my name is John Smith. Will you sit on my breadbox while I cook or is there some kind of speed limit on that thing?

Babbage, our next pipeline maker, appeared more sophisticated in comparison with previous AIs. This program produced such an interesting “piece of arts” in the pickup world :

• You're looking good today. Want snacks?
• It is urgent that you become a professional athlete.
• (In your best Albert Einstein voice) "I wouldn't change a thing."

The last one, Ada, gave the smallest number of options, which looked more like email headings. Here are the examples:

• Body Softening Pads
• 2017 Rugboat 2-tone Neck Tie Shirt, and
• Future Pop-Tarts by Tracey Thorn

The most memorable line of all the pipelines spouted is still the one that was voiced by a more primitive neural network, which had a trend back in 2017: “You look like a thing and I love you.”
What can we take from the experiment results and trends in the use of AI in the industry of “digital love” and dating?

There are areas of human activity that should still be left to people. The use of artificial intelligence in these areas not only suppresses lively communication, eliminating all manifestations of social groups but also seems at least bizarre and ridiculous. 

We’re going to be much more productive and happier using these technologies in business, freeing up time for our loved ones. But not the other way around. 

Can a robot become more human than a human being?

Today we’ll go through another great battle of the e-commerce industry in recent years - the confrontation “ Magento vs. Shopify”. Both have a lot going for them, but which one to choose when building an online store? Let’s take a little look at each of them.

Talking about the Design and Themes, there are many ready-to-use themes for both Shopify and Magento, but Magento also allows you to have custom designs of your own. 

Next comparison step is Cost and Fees
Magento Open Source is available for a free download. Still, your business may need to splash out a little extra cash if you want to use premium plugins and themes or if you need to hire a developer to build out some custom functionality. 
Shopify is more of a monthly thing, in that you pay a monthly fee and they give you all the functionality you’ll need, at least for a smaller business. If you’re a more significant business, you might be better off going for Magento if you want to customize your build heavily.

The next major battleground “Ease of Use” and it’s one where there’s a clear winner. 
Shopify is undoubtedly one of the most intuitive ecommerce platforms that we’ve come across. It sets something of a benchmark and is only a little trickier to pick up than WordPress. Shopify wins the battle of Shopify vs Magento, but that win comes at a cost. 

Now we came to the “Third-Party Integrations” factor
This one’s pretty simple, and indeed there’s almost no competition in terms of Magento vs Shopify. With Shopify, what you see is what you get. Third-party integrations are few and far between, and so if it’s not included in the core installation, then you might be out of luck.
With Magento 2, however, there’s a whole community around the core product, with literally thousands of third-party Magento developers working on extra functionality, from new themes to plugins. 

What about Merchandising features?
Magento 2 comes with the Visual Merchandiser, a suite of tools to add and modify products and categories as smoothly as possible. There is a range of other ecommerce essentials, from product tags and order tracking to their dedicated visual mode, which allows you to view all of your products in a grid.
Shopify has a visual merchandising plugin of its own, and both platforms offer pretty similar features for creating a library of products. 

Okay, almost all comparison parts are done. The one we've missed is Market Scalability
Both Magento and Shopify have pretty decent scalability, and we’d be hard-pressed to pick a winner here. If we had to pick just one, we’d go for Magento – but only because again, it’s so customizable that even if some aspect or another doesn’t scale, you can probably tweak the code to change that. 

The last but not less important is SEO inbuilt tools battleground
Shopify is arguably easier to just pick up and run with, mainly because an essential part of SEO is creating and deploying high-quality content. 
Magento 2 is more robust and feature heavy, and while it can take a little longer to wrap your head around it, there’s no reason why it can’t be a powerful centerpiece for a solid SEO strategy. 

Summary

By now, you know everything you need to know about the pros and cons of Magento Open Source and Shopify, and so you should also be ready to make the call one way or the other.

Hear you in a couple of weeks, another Friday.

Today we’ll talk about the challenge of ecommerce platform selection and will look deeply into the Magento VS WooCommerce battle as a distillation of the dedicated article. 

Batman or Superman? Iron Man or Captain America? Blair or Serena? I bet you have some favs among these long-time competitors. 

But what about Magento vs. WooCommerce? If you own an ecommerce business or are about to build it, you must have heard of them. Both ecommerce platforms are popular workable solutions, but how can one choose rightly between the two? 

We have been working with both ecommerce technologies long enough and can help you out. If you are equipped with all the relevant, up-to-date info, you'll know for sure which ecommerce solution is the one for your business. Let's start with diving into Magento features!

Both ecommerce platforms work well for numerous domains. Here is a comparison of some basic aspects to bear in mind. 

We`ll start with Themes factor: 

Magento and WooCommerce offer free and paid themes for various purposes: customer support, accounting, site optimization, etc. 
With WooCommerce, there is a chance you'll be able to install the extension yourself. As to Magento, you'll need Magento development services. 
However, despite the extra time and money, Magento website additions are more cutting edge, so your ecommerce store will be more powerful. Here we definitely give credit to Magento. 

More or less clear, right? And what about their support service? 

In this match, Magento vs. WooCommerce, we'll call it a draw. Being open-source, the platforms have huge active communities ready to look into any problem you may face. 

Now, here we came to the Pricing factor

As we already discussed, Magento's and WooCommerce's code is out there available to be adjusted according to your needs. WooCommerce is easier for beginners, but you may end up paying a fair amount for any extra feature you'd like. 

Magento is more suited for ecommerce businesses, which means that its basic functionality can be sufficient for you. But there is a hosting plan you'll have to purchase. Depending on your ecommerce website specifics, you'll have a monthly hosting package to maintain.  

Summing up the Magento vs. WooCommerce comparison, it's safe to say that each has its strengths and weaknesses. WooCommerce is claimed to be more user-friendly and more comfortable to start with, while Magento is more powerful and well-known for its extensions. All of us know that selecting the appropriate tech stack is crucial. But it's not the thing we recommend as a starting point. First, you should elaborate on your business goals and requirements, realize your company needs and define your target audience. The more specific you are, the better we'll know which ecommerce solution will fit best. 

That’s all from us for now. For a detailed version, you are welcome to check the dedicated article.

Until next time! Bye!

We've been discussing how to choose a domain name for your future e-commerce store, analyze competitors` strategies during the previous episode. We've also revealed the main pros and cons of using website builders (remember, yeah?).

You can think about marketing your ecommerce site in terms of three main audiences:

1. People who have shopped with you before or at least know about your store
2. People who are looking for the kind of product that you offer
3. People who would want the product that you offer if they knew about it.

For your first audience, you should have a direct line to them — email communications, content marketing, or social media. You can communicate for “free” (except the budget spent on tools).
- You can use smart remarketing tactics from CRMs like HubSpot to send highly customized emails to shoppers that have abandoned carts. If a customer has logged in, send them customized emails with pictures of the items they were interested in and customize the emails.
- You can also get them back with remarketing. Facebook, Google, and a few others give you the opportunity to track the actions of your sites’ visitors in order to optimize the way you market to them.
- Connect via social media. Grabbing a user's attention begins with something they're interested in, not something you want them to do. By demonstrating your topical expertise, interested users then have a reason to visit your website and see what you have to offer.

Your second audience should be able to discover the products you have that meet their needs. This can involve search SEO, content creation (like blogging), and search engine marketing. 
The main idea is to bring traffic to your site providing helpful info that positions your brand as an authoritative voice in your vertical:
- Create power posts (massive articles that are full of useful information on a particular topic that relates to your industry). 
- Publish these power posts as unique pages on your website and they’ll start driving traffic as Google recognizes their value and sends searchers your way. 
- Create guides that show off how your products are used in real-world applications. Make sure to keep it visually interesting and focus on creating highly valuable content. 
- Try out Google Ads. Using Google Ads for search engine marketing gives you an opportunity to bid on keywords so that you can show up first in search engine results pages.

To reach out to the third audience (people who would want the product that you offer if they knew about it), you’ll need to think in terms of brand awareness. Are you offering a fix for a problem they don’t know about yet?
- You can use influencer marketing as one of the ways of being promoted to the new audience. Influencer campaigns help present your products in front of a potential audience by a trusted person (advisor, blogger, vlogger, etc). But make sure to work with influencers whose brand and audience are in line with your products!
- Participate in online events and discussions. As online communication and events surge, there may be opportunities to increase brand awareness with target communities.

So, we've covered the major online store promotion options. Apart from them, you can certainly sell your product through Instagram, Facebook, or Amazon, but it's a bit of a different story and is better studied separately.  

Well, you've chosen a name, went through the hosting process and analyzed your competitors` activities. The point is you are on your way. Let's move on to Design and Development tricky moments discussion. 

Here we'll review what ecommerce solutions the market has to offer today. 

You must have heard of website builders like Wix or Shopify. Their significant advantages are affordable pricing, usability, and simplicity. You can build an ecommerce shop literally in a day. What is also attractive is the perk of having a hosting plan included in the package.

The other useful factor is buying a ready-made design theme if you don't want to invest much in design. These two benefits indeed make such an ecommerce solution look like a perfect choice; however, it's not always the case. 

Your choice is limited by what the website builder foresees if you want to add any additional functionality. Moreover, if it's not included in the package you've purchased, it'll cost extra. 

Secondly, such websites can't be scaled. You'll have to migrate to another ecommerce platform.

Thirdly, SEO promotion is complicated for these websites, so if you have any long-term plans to promote your ecommerce store, they won't happen with website builders like Wix. The only option available in this regard is paid advertisement.

Finally, your online store doesn't quite belong to you. The reality is that you are just renting it. 

The second option on our list is creating an ecommerce website using a framework. Here you'll need coding knowledge or an expert in the area. It's not a cheap alternative and is a sound choice for large businesses. They normally have a dedicated development team or reach out to software development agencies to implement their project. Such ecommerce projects imply proper long-term planning with an already existing business process that serves as a basis for software. 

As you see, there can be a whole team working on ecommerce project design and development. 

The third alternative is opening an online store on open-source platforms. It doesn't require that much time and effort compared to the previous solution described. It all depends on your requests, nowadays it's increasingly popular and works for many cases. Such platform examples are Magento, WordPress, etc. Having a convenient ready-made CMS is a huge benefit for any ecommerce project. The other is the variety of plug-ins and extensions, and each meant best for a particular e-commerce need: documentation audit, SEO, payment methods, etc. 

Like with website builders, you can go with a theme or implement a custom design to stand out. The pros list also includes scalability and multi-functionality. 

Creating an online store using an open-source platform may take a few weeks or months; everything depends on customer traffic. Such ecommerce projects are well-optimized for search engines, which allows setting your marketing strategy way ahead.

That's all for now. During the following episode, we will discuss the promotion and Marketing of an Online Store in 2021.