Author Archives: Brygida Voicu

Alexa Custom Assistant

Reading Time: 3 minutes


Today, Amazon has released a product that allows you to construct your own solutions based on their technology. This intriguing option, which will speed up and facilitate the process of making your own products more attractive is Alexa Custom Assistant.

Alexa herself probably doesn’t need an introduction, but for the record, it’s a cloud-based voice service that lets you use your voice to command her to perform some tasks for you. The possibilities of this technology are very wide, ranging from shopping, making your home smarter, providing you with information, to product control and much more. Of course, this is not the only such technology available on the market, but by introducing a unique option that I would like to present in this article, it has a great chance of becoming the strongest.

By Alexa’s  Custom Assistant, Amazon enables companies to create their own smart voice assistants by using their AI technology and then connect it to Alexa.  The solution is therefore also to enable the cooperation of two assistants, which significantly affects the attractiveness of not only the product of the person who generates his assistant based on Amazon technology, but also of  the very Alexa. Once connected, it would continue to perform its regular tasks and the new assistant would play a role of a so-called “product specific expert”. This expert would be responsible for operating with more concrete features, that are directly linked to specifics of  the product. So we can say, the tasks would be somewhat of delegated. How does it work? Amazon says that “if a customer asks Alexa to roll down a car window, or how to troubleshoot a device, the request will be routed to the brand’s assistant. If a customer asks the brand’s assistant to play an audio book, the request will be routed to Alexa,” The fact that both assistants will in a way coexists is important regarding the capabilities area. There will be no need for the Companies, who decide to take advantage of the option proposed on Amazon, to develop the basic skills of the assistant, instead they can immediately focus on improving some specific and unique, in view of their product, capabilities. So this cooperation will prove to be both more effective and cheaper. It is fair to say that this option is mutually beneficial to companies and what is more, it  fulfills the main goal of meeting customer needs.

An advantage that should be especially noted is the significant reduction of costs and time, which according to Amazon is possible due to “allowing companies to leverage Alexa’s world-class technology stack to create their own intelligent assistant without the investment, long development cycles, and resources to build it from scratch and maintain over time.”  Creating such assistants proves to be very costly due to the constant necessity of their improvement- we have to bear in mind that competitiveness on this  market is very high. Amazon by investing in Alexa Custom Assistant and making its improvement up to date in a constant manner, will let the companies simply benefit. Therefore, there will be  no need for them to monitor market changes and invest in improving the voice assistant themselves. Amazon will do it for them. So, companies will now have a chance to form a tailored to the brand, customer needs and “aesthetic” reasons (=creating your own unique voice) voice assistant, for not only a much lower price but also relatively much faster.

Amazon, within this option, provides access to all their voice technologies, experts, ongoing improvements in AI, privacy and security settings. Companies will be able to generate their own wake words, which according to Amazon ”are created using the same state-of-the-art process used for developing the Alexa wake word and deployed through the high-accuracy Alexa wake word engine”. Amazon assures that through the whole process, companies will be guided by science experts.

As we are still experiencing a technological revolution on a daily basis and the lifestyle is closely related to digitization, looking for a solution that will further facilitate this reality and make all digital experiences even more smooth is a large part of today’s business. Voice assistants are a great example of that, as the ways they make life easier are practically endless. To name one, they often save our time by automating repetitive tasks. But they also prove to be useful in enhancing e-commerce marketing, customer support, smart offices and many more. Therefore, it is important to note that  Alexa Custom Assistant can be build into any type of product and as such immediately increase its “charm”. The first customer however, is the global automaker Fiat Chrysler Automobiles (FCA). Which is quite interesting, in relation to the next paragraph.

This newly launched Alexa Custom Assistant has a potential of becoming a huge game changer in the voice assistant technology area and especially concerning the car software space. Regarding The Verge article, making Alexa an “in-bulit solution in environments that may not need a phone at all” is an interesting way to “bypass the need to directly compete with Apple and Google”. This approach to implement voice technology directly in conjunction with a specific product is a quite unusual move and will definitely trigger an interesting response from other tech giants. It definitely sets a challenge for Cortana, Google Assistant and Siri.

This one Polish startup conquering Asia you need to know

Reading Time: 4 minutes

Although, according to Statista, the value of online retail sales is growing and its global share of retail sales is expected to reach 22% in 2023, we still cannot forget that the remaining 78% is for offline trade.

When conducting online retail sales, there are many effective tools to promote your store and thereby attract customers. You can use sponsored links, SEO or take advantage of the benefits of social media and for example create a competition for customers, thanks to which they will engage more and probably also  promote the store on their own. You can also offer small discounts for first-time purchases and use a very powerful tool which are subscriptions. According to AdWeek’s analysis subscription model is perceived among clients as budget-friendly and convenient and at the same time, it makes it easier for businesses to scale and predict revenue.

It follows from the above that the possibilities to attract customers online are practically endless. The reality is quite different in offline world. Therefore, all solutions that attract customers are in high demand even more concerning offline retail.  This is where a Cosmose AI decided to step in and make a revolution.

Cosmose AI is a Polish behavioural targeting technology startup which was founded in 2014 by Miron Mironiuk and it holds that its current valuation is more than $ 100 million. It has offices in Shanghai, Hongkong, New York and Warsaw. Huge and innovative companies such as Walmart, LVMH, L’Oréal, IWC, Este Lauder,FENDI, Samsung and many more use their solutions.

What this company basically does it tracks foot traffic in traditional offline stores in order to help to predict how customers will behave. The company is able to indicate the customer’s location with 2-meter accuracy, i.e. indicate in which aisle the customer spent a certain amount of time. This 2 meter accuracy is possible due to proprietary solution based on neural networks. Thanks to this, then, a few hours or a day after the visit, an advertisement for this store with, for example, the product or subject  from/of the aisle/shop will display itself on this persons social media.  Cosmose works with online media platforms like Facebook, Google, Line, Weibo,WeChat. Regarding the prediction, a very high 73% accuracy was achieved for when where and who will go shopping (concerning beauty category). The startup also claims that it can determine whether a customer will or will not like a certain product. Of course, the possibilities for this solution do not end with the displayed ads .Through the application dedicated to interested stores, also the impact of the advertisement on the recipients can be checked, i.e. how many people who clicked or saw the add visited the physical store again. This application also allows you to track changes in the traffic of competing stores, which is very interesting feature because it provides a valuable insight and companies can mutually inspire themselves and improve. And what is even more useful- the app enables us to connect also with customers from competitor stores.

To conclude, the company offers the mentioned above location tracker- Cosmose Analytics, a prediction and analytics  platform which is aimed at helping sellers with marketing and stimulating sales- Cosmose AI and also Cosmose Media which is used for targeting online advertisments. Regarding the very targeting, the company analyzes shopping patterns, what allows them later to chose customers who will with the highest probability be interested in your product. According to Acxiom, this results in Cosmose delivering 135% higher offline conversion in comparison with leading advertising platforms.

So what is the technology behind all these wonders?  These unique marketing services are possible thanks to the company’s proprietary technology, which is OMNIcookie. OMNIicookies allow you to locate customers using GPS and a network of WIFI transmitters, but what’s important-without the need to connect with it. So no visitor action is needed- they don’t have to open any apps nor connect with anything. The company uses agentless component (a tool which collects performance metrics without the need to install any agent software on the devices that are monitored) which works somewhat like this; the company access the data they need but they don’t know who they belong to because they use anonymous mobile user IDs. They access the data from weather apps, social media platforms and also fitness, radio, gaming and news apps.  So the above explains how the company’s activities are in compliance with data protection.

As the numbers best describe the scale of operations, the company’s ecosystem consists of; 1 + billion tracked devices (this equals to 1 billion understood and influenced shopping habits), 360,000+ stores scanned (they aim to grow that to 10 million by 2022) and data from 400,000 applications.

800 million, so the vast majority, of those tracked devices are from China. Now, why China and not Europe or U.S? According to the founder and CEO, Miron Mironiuk, he first was set on chasing American dream, but in the U.S smartphone penetration was not very high. The opposite was true in China, where this penetration was 90% high. To present you with some numbers; China has over  1 billion operating smartphones. Also, Mironiuk mentioned that due to a very high competitiveness in the extremely fast growing China market, you always have to improve your product. This is both very challenging and rewarding.

Nowadays, it is impossible not to mention the COVID-19 pandemic. It would seem that since the company provides services to physical stores, its activity is currently much more difficult. Well, this is not entirely true. During these difficult times Cosmose aims to help their clients by providing useful insights on which basis clients can make decisions about  when to reopen their stores, how to enlarge revenue and what kind of inventory will attract customers. They also enable their customers to connect with those who used to shop in their physical shops in order to inform about the opportunity to buy online.

Couple months ago the startup announced that it has raised $15 million in Series A. This, according to the company is used to improve the product and expand globally. Considering the opportunities offered by the company, its valuation and the amount of investments that are made in it are not shocking. Cosmose very effectively combines the analysis of customers in-store behaviors and habits, so offline activities with the power of online marketing.It is a very interesting and innovative startup that definitely changes the situation of many entrepreneurs and creates a huge change in marketing.

Latest iOS update, privacy and Facebook.

Reading Time: 4 minutes 

The issue of our privacy on the Internet is an issue that raises many emotions. An equally high-profile issue is the debate about its violation by various applications. The topic is definitely not transparent, or at least it has not always been. The number of incidents of data breaches continues to increase, and with it, individual users’ concerns about their own data and privacy regarding their activity on the internet. Despite the awareness that such processing should not take place without our consent, there are many examples that our consent does not always play a significant role. Our vigilance was aroused specifically after the not so distant scandal at Cambridge Analytica.

Following our activity on the Internet raises a lot of emotions, but it is also certainly something we have been living with for a long time. The use of some data, of course with the consent of the user, is a very important business driver. It allows businesses to target specific ads and content at different users based on their interests. And as such, they make a lot of profit.  As long as it is  all clad in a specific legal framework, as long as we know how and what for is  our data  used and also how important role our consent plays in each matter it should be just fine.

In the view of the aforementioned, the latest iOS update seems very interesting. On Monday, Apple launched iOS 14.3 and with it an interesting feature appeared in App Store. From now on, any tracking will be or at least should be much more transparent for users. Thanks to this App Store update, before installing the application, we will be able to see a detailed list of how and what for  our data can be used by the application, how ‘developers may handle our data’.

Until now, probably everyone was aware that by installing the application, we allow access to some data, but the ability to see the entire detailed list will definitely open the eyes of many users.  As an example, below I present the list regarding Facebook.


From now on, everyone whose application is or will be in App Store will have to disclose information about collecting targeting data and seek for user consent.  This App Tracking Transparency did raise some tension, especially in the earlier mentioned Facebook circles. We cannot deny, that transparency regarding this issue is extremely important and every user has the right to know when his data is collected, for what purposes and what happens to it next. According to Apple, this option is simply about giving the user a possibility to make a more informed choice not about blocking apps from creating and targeting personalized ads. Simply saying, Apple decided to take some more steps to protect internet users, whose data is clearly becoming the currency of the internet nowadays.

However, Facebook looks at this option, requirement completely different. Facebook, regarding this matter, presented itself as a defendant of small businesses , claiming that this transparency feature will cause a lot of damage to them. They did not like it that this information about tracking and expressing consent is so transparent and easy to obtain in comparison with what was before.  They even created an ad, which was posted in the Washington Post, New York Times and Wall Street Journal and carried the  headline; “We’re standing up to Apple for small businesses everywhere”. They claimed there that this feature is ‘limiting businesses ability to run personalized ads and reach customers effectively’. They also shared their concerns that without personalized apps ‘ average small business advertiser stands to see a cut of over 60 % in their sales for every dollar they spend’. The thing is, that this statistics comes from their own data…

The uproar that this update has caused in Facebook circles is actually not that shocking but it certainly does send a clear message. Through their statements it may seem that the passing intention is not so much to protect small businesses but to save themselves from having to resign from collecting as much data as possible and therefore making the biggest possible profit out of it and all of this without much transparency. It seems that the statement about Facebook having lack of respect for the privacy of users is not so inadequate in the context of their sharp statements and protests.

Of course, we cannot deny that user data is often critical to business, but obtaining it cannot be done according to the principle that the purpose justifies the means. First things first, users consent is absolutely required. Therefore, this update, App Tracking Transparency is a huge reminder about who the data belongs to, how important they are and also it is a huge step for obtaining a healthy balance between advertising methods and privacy.


How to outsmart facial recognition systems?

Reading Time: 4 minutes

The statement that facial recognition technology is controversial will probably come as no surprise. We are concerned about the possibilities of this technology and its applications. No wonder, since in the wrong hands it can even become a tool of mass surveillance and human rights violations (here I mean its particularly alarming use in China on Uyghur minorities).

The Clearview AI company created by Hoan Ton-That deserves a mention- their application also seems controversial in the context of possible applications. It used to be available for private customers. How their software works? Basically, you can (at this moment not personally you, because it became only available for customers associated with law enforcement or some other federal, state, or local government department, office, or agency) upload a photo of a person and it will provide you with more public photos of that person and links to where this photos appeared. Clearview AI has a database of more than 3 billion pictures that were scraped from millions of websites (like Facebook, YouTube, Venmo). When the photo is uploaded, it is compared to the database and voilà- the result appears. Simple, elegant and with high potential of being dangerous. Just to provide you with a comparison of its search power to different “recognizing” tools:

It is used mainly by the law enforcement departments and governments ( in USA especially). Those experiments prove to be very effective. For example, the Indiana State Police solved a case within 20 minutes of using the app. A fight of two men which ended with firing a gun has been recorded by a bystander. The police uploaded this image into the app and got a match very quickly. The gunman’s face appeared in a video posted on social media and his name was included in the caption. As he was not in government databases it would be quite difficult to identify without the help of this app. I would like to emphasize once again, that data is downloaded from platforms, including social media (which, yes, violates their terms of service). So any photo posted by anyone can go to the base…

Clearview does not make their service publicly available but another company might. If this happens, according to NY Times, someone walking on the street could be immediately identified by just anyone (and also his address etc).

So how to protect yourself from this now that we all exist in virtual reality? Is there any way to outsmart these kinds of applications? Generated Media seems to have come up with a answer to this problem. They created a tool called Anonymizer. How it works? We upload a real photo of our face and the tool provides us with variety of fake faces that are similar to ours. Similar, yet if any facial recognition company will add the face to its database, finding the real you by using the fake image will be , according to the company, quite impossible. According to Generated media, to achieve higher security of your anonymity, you should change the fakes as often as possible.

What is the technology behind this solution?  The company uses generative adversarial networks (GANs). They were first designed by Ian Goodfellow in 2014. It works by putting two neural networks against each other in a form of zero-sum game. So, we have a generative network which goal is to  create a fake image good enough to deceive the discriminative network. Of course, the discriminative networks goal is to avoid getting tricked. In order to do that, it has to correctly guess which images are artificial images and which are original data.  So put simply, the generative network creates new classes of data and the discriminative network checks their authenticity. This game has many thousands of rounds. Networks check their moves and use backpropagation to keep improving. This technology already works on a large scale. In less than a year, Generated Media created more than 2 mln fake images. The company uses them as training data. Anonymizer works by analyzing the features of persons face on uploaded photo and finding about 20 similar fake ones within Generated Media database.

Of course, we cannot view this technology solely in the context of protecting ourselves against facial recognition systems. Like most technological solutions, it can also be a threat if it falls into the wrong hands. I mean here for example Deep Fakes. Also, what is interesting, the technology on which it is based, meaning GANs,  aroused much controversy among SAG-AFTRA (Screen Actors Guild – American Federation of Television and Radio Artists)- they even took legal action. In their opinion, GANs are a threat for actors because production companies might decide to create GANs-enabled holograms instead of hiring real actors. Back to Anonymizer; as of course the intentions of the person using this technology cannot be prematurely identified, in case they are negative this  is actually not such a convenient tool. Why? Because such users could still be tracked back to this company and thus identified.

To conclude, Anonymizer can be both fun and a potentially powerful tool. Powerful in meaning of protecting ourselves from not only mass surveillance applications but also from being targeted for stalking. I would also like to stress the positive context of GANs use. They can be used anywhere where we deal with a visual pattern, even in modelling dark matter in astronomy. We have to keep in mind that they were originally developed to train large sets of data. To my mind, Anonymizer   is an extremely interesting tool. It can definitely be described as a useful way of showing the utility of synthetic media.



No headphones, loud audio, a busy place, yet still you remain the only one hearing the sound?

Reading Time: 4 minutes

As unbelievable and drawn straight out of a science fiction reality it may seem, an Israeli company called Noveto Systems just proved it’s possible. The company is directed by dr. Christopher Ramstein and holds over 44 patents. Their new, long awaited device debuted last Friday and it enables a completely new and exciting way of listening to audio. For now, they presented a prototype, but there is a promise, that a consumer-ready device will be ready by December 2021.

So, what possibly new can there  be about listening to audio, you may ask? Let me introduce you to SoundBeamer 1.0. Simply put, it’s a device that thanks to patented by this company specific soundbeaming  technology, targets your ears and directs the sound directly at them. Only you can hear the sounds that are transmitted, but it does not cut you off from the real world- you can still experience voices from the surrounding. The feeling when enjoying this way of listening is, according to Associated Press journalists who tried it, as if it was playing in your head and happening all around you. According to company’s employee, Ayana Wallwater: ‘when you are inside the bubble, you get the sound, the volume, the immersiveness of it’. In Noveto’s opinion, SoundBeamer 1.0 represents a completely new way of sensory listening.

But let’s move to the technical aspects. Although, the sound beaming principle is nothing new, Noveto Systems managed to transform it into an extremely interesting technology that tracks your ears. SoundBeamer 1.0  is a stand-alone branded hardware which works by using ultrasonic waves and beamforming in order to create sound pockets just next to your ears. This is possible due to Noveto’s Digital Signal Processing algorithms, which are mathematical formulas that influence and manipulate digital signals. The ground of the product’s operating can be explained on the basis od directional sound. This term refers to the concept of creating fields of sound which spread less than usual loudspeakers (they create sound waves that diverge quickly-this is called diffraction- and aim to spread as much sound as possible and cover as much place as possible), by using many different devices. As those directional speakers need a minimum of diffraction, they use ultrasonic speakers. Owing to ultrasonic waves, which diffract less than acoustic waves, the sound can be aimed at more accurate locations. Here, in Noveto system, a custom-designed transducer array pushes the algorithm-designed waves into air and when those ultrasonic waves come across an object (you, to be exact) they convert into acoustic waves, which you can hear. But how is it possible that the waves don’t hit somewhere else, but the entire process of converting them happens just outside your ear? SoundBeamer 1.0 has a built-in 3D sensing module which locates and follows the trail of movements of your ears in real time. Then the ultrasonic waves target them and converts itself into small pockets of acoustic sound. What is amazing, is that the system does all of this in real time, so you can move your head freely, and the sound pockets will always remain next to your ears.  What’s worth mentioning, this product can be connected via Bluetooth to any device.


What are the benefits? To quote the Noveto company, this will be (or actually is) a world ‘where sound can be both private and spatially three-dimensional’.  Confidently, our comfort will increase, owing to the lack of headphones. We will have more freedom and more control over the sound. Also, we will always be “present” and focused, due to the possibility to experience  sounds from variety of sources.

I am more than sure, that at the moment we are not able to count even half of the device’s applications that may appear in the future. Just for now, it can prove to be extremely useful regarding conference calls- you will have a private experience, but won’t be isolated from your team. Another area this device could be used for is public speaking- it could help some people who are not that good at it, or do not have much time to prepare, to always sound and look (due to the fact that it is not visible whether you use it), professional. Maybe it is a mundane matter, but how many times when driving with someone in a car, you felt irritated because of the type of music they listen to, or navigation instructions that occurred from time to time, and as you wanted to talk, isolating yourself through headphones was not an option? Using this solution every passenger can have their own choice of audio and still enjoy the conversation. Another entertainment matter SoundBeamer 1.0  might contribute to are parties- not everywhere and not every time you can host them without interrupting your neighbours. Listening to music loudly might occur as an issue, but due to this device the problem could vanish. SoundBeamer 1.0 will revolutionize the audio market for sure, the question is whether it will happen in the short or long-run. Although most surveys show that we, as people, are resistant to change, this is usually due to uncertainty which comes along with it. I am positive, that we will let this solution play a significant role.  Yet still, I can imagine situations in which this technology could cause trouble. For example, it might prove to be a difficult nut to crack in examination situations. Checking if someone isn’t transmitting information via ultrasonic waves could be hard… However, as necessity is the mother of invention, a solution for that would probably also emerge quickly.