Category Archives: Apps

Fine-tuning GPT3

Reading Time: 3 minutes

to improve performance in specific tasks and domains

GPT3 is a powerful language generation model, making it ideal for building chatbots, conversational interfaces, as well as other driven applications such as automated content creation. We can also use GPT3 to generate code, art and even to compose music. The main advantage in most cases is the ability to ask the right questions. While in more advanced projects, knowing how to use the GPT3 API to generate automated tools can greatly speed up our business or academic tasks. 

This is one of the reasons why tuning GPT3 is so important. Doing this for specific tasks or domains involves training a model on a similar data set that is specific to the task or domain of interest. This process is also known as transfer learning and allows the model you are using for chat to adapt to new tasks by adjusting the weights of the free-training model to better fit the new data and emphasize a specific topic. 

Take the example of tuning GPT3 for sentiment analysis. If we wanted to use chat for this we would train the model on a data set or text labeled with sentiment. Remembering that sentiment can be positive, negative or neutral. This is useful for, among other things, the work of a data analyst when he or she wants to find the sentiment of different tweets against, for example, a statement or the general character of a given politician. 

GPT3 tuning can be done using various steps, and learning libraries such as TensorFlow or PyTorch. This can be done by adjusting the parameters of a pre-trained model using new data. This process can take anywhere from a few hours to a few days, depending on the size of the dataset and computational resources.

During a usual chat with ChatPT3, you may notice that the chat remembers what you asked it a few messages ago and is able to make changes based on its pre-generated statements. The chat also learns from our conversations. Tuning works in almost the same way, but on a much larger scale. Moving into the programming area, we can tune ChatGPT3 using the OpenAI API model, which includes GPT3 chat features. 

The tuning process requires access to a dataset and a development environment to train the model, which is not directly provided in the GPT chat. So, to tune GPT3 you need to create an OpenAI API key and use it to evaluate the GPT3 model. 

You can then use the API to tune the model on your specific task or domain by providing a dataset or using the API to train and update the model. You can always get pre-trained models that are tuned for specific domains or tasks, which are available from a number of different providers such as https://huggingface.co/models and more. You can use these models without training your own data set and simply add them as an add-on to GPT3 chat. 

Another advanced tuning technique is data augmentation, which is a technique that is used to improve the performance of GPT3 models. This technique is used to artificially increase the size and variety of the training data. This can be done by using various techniques such as adding noise to the data, rotating and inverting the images and creating new data by combining existing and old data. This can help to make the model more robust and reduce overfitting. 

For example, using data augmentation techniques artificially increases the size and diversity of the medical dataset and can help GPT3 learn medical-specific language and terminology. Transfer learning, on the other hand, allows the model to adapt more efficiently to a new task or domain. I strongly encourage you to experiment with ChatGPT as it can save us many hours of tedious work and improve the end result. 

Source: OpenAI ChatGPT Master for Business and Software Applications

Another novelty in the field of artificial intelligence

Reading Time: 2 minutes

Microsoft has introduced the artificial intelligence VALLE, which is able to simulate any human voice based on an example lasting only three seconds. At the same time, the voice is imitated very reliably, preserving both the timbre and the emotional coloring of the original.

Unlike other text conversion methods, which often synthesize speech by manipulating waveforms, Microsoft’s development mainly analyzes exactly how a person sounds, breaks this information into separate “tokens” and uses training data to compare their “knowledge” about how this voice will sound if the AI pronounces other phrases. VALL-E is trained on recordings of 60 thousand hours of conversations of more than 7 thousand speakers, which is about 100 times more than in existing systems.

The most interesting thing is that it takes VALL-E only a few seconds to clone a voice, and besides, the similarity of emotional timbre and background noise. Moreover, the fact that the model is still quite new, but already has a huge advantage over others. And, of course, further improvements are expected to lead to even more human-like speech.

Google showed its AI Duplex, which can also speak almost indistinguishably from a human, back in 2018, but the essence of Microsoft’s development is not in the AI itself, namely in its ability to imitate different voices.

As in the case with all the other different AI models there is concern about the misuse of Vall-E. For example, to imitate the voices of public figures, politicians, or stars. Criminals will also be able to obtain confidential data if they make a person believe that he is talking to family, friends, or officials. Some security systems also use voice identification.

However, VALL-E developers and researchers claim that such risks can be reduced using a different model. It is possible to develop a model that will be able to determine whether the audio was generated by the VALL-E or not. However, how it will look at the moment of a telephone conversation, for example, is not entirely clear.

But so far this development is not open for public use and we ourselves cannot try out VALL-E. We cannot make sure how well the model imitates intonation and changes in the emotionality of speech, and how quickly everything works. But at the same time we have the opportunity to listen to the already generated text from VALL-E.

Sources:

https://www.business-standard.com/article/technology/microsoft-s-new-ai-tool-vall-e-can-replicate-any-voice-in-just-3-seconds-123011000696_1.html

https://valle-demo.github.io/

Internet, law, and the freedom of speech

Reading Time: 2 minutes
What the Left Keeps Getting Wrong About Free Speech

Through all the times big corporations like Twitter, Google were immune to the lawsuits against them regarding the content posted on each of their platforms. However, current legal cases are about to change this.

There are always battles between Republicans and Democrats, between right and left about the censorship on the Internet. Nowadays, the regulations of blocking content are not really understandable for a lot of people. Some unharmful posts can be deleted, while disturbing pictures still remain in place. So, the debates about how to make the Internet field not a paradise, but not hell at the same time, are on the stage.

This summer is considered to be decisive for the companies in the US, thus, in the European Union. The aim of the justices is to review all the existing cases and come with the legislations and systems that would prevent people from posting/seeing disturbing content, while still having a freedom of speech.

One of the most recent lawsuits was made against Twitter. Several antisemitic posts were noticed on the platform, that were indented to attack Jewish people and deny the Holocaust. Those posts breaking the law of Germany (from where they were posted) as well as breaking terms and conditions of Twitter. The users of the platform have expressed their feeling of being betrayed by the company after it was bought by Elon Mask.

Why was this post not deleted if it was violating terms and conditions of the platform?

The answer is rather easy. After Elon Mask’s purchase of Twitter, it was announced that the company is going to loosen rules and will stop banning so much of the content. The Campaign Against Antisemitism that was a close partner to the company said that the ‘laws’ of the firm started to fall apart after the shift of the ownership to the billionaire.

So, now we are approaching one of the biggest and most interesting changes when it comes to the censorship not only on platforms mentioned above, but also all over the Internet. A lot of people think that we should still leave space for the freedom of speech, as we want to build strong democracy, however this can threaten people life’s and feeling of safeness on the Internet. 

While on the other side, we have the fact that banning ‘inappropriate’ content, can actually be difficult to implement, as there are no strict limitations of what is good or bad. Additionally, that would threaten one of the fundamental rights – the freedom of speech.

To read more about the thought on the topic: https://www.ft.com/content/131c1c6f-7adc-4217-b251-6e7e18570581

Tagged , , ,

Żabka Nano- the innovative future of convenience stores

Reading Time: 4 minutes

Many of you are definitely familiar with a Polish chain of convenience stores called “Żabka”. The owner, has opened up their first 7 stores in 1998, and has quite literally spread around the whole map of Poland (and even slightly outside of it! You can find a few “Žabka” stores in Chech Republic! as well) . It is easily one of the most popular stores in Poland today, and recently, in 2021 a brand-new concept of Żabka has been introduced- Żabka Nano.

Żabka Nano - Pierwsze zakupy gratis!

Żabka Nano is not only the most innovative store in Poland but also the largest autonomous chain store in Europe. The overall idea of autonomous stores already exists in Australia, Japan or South Korea (and probably somewhere in the US), but Poland’s Żabka still overtook all other potential European competitors.

Nano - Żabka Group

The main difference between the two stores (the original and the autonomous one) are the employees. More precisely, the lack of them. The idea of Żabka Nano is that it can be open 24/7 and accessed using a credit card or żappka- the stores app that allows you to collect points for your purchases. However, if you wish to enter with the app, you need to make sure you have funds in your account. Moving on, once you’ve put your credit card near the reader at the entrance of the store, the doors to Żabka Nano open. The easy accessibility to enter by card has been implemented by the Żabka’s collaboration with AiFi, in all of the Nano stores.

Żappka” – nowa aplikacja rabatowa sieci sklepów Żabka - mobiRANK.pl

It is a great way of saving your time since you don’t have to stand in a queue to buy your necessities. The store itself is usually a lot smaller than the normal Żabka stores (to no one’s surprise), mostly because it offers less goods than the basic stores. For instance, you will find yourself not being able to purchase alcohol or cigarettes, but you can buy 0% beer (what’s the point in that though). You will also find basic products, such as; water, snacks, juices and refreshments, ice cream and frozen food.

Pierwsza Żabka Nano już działa w akademiku

For the Żabka customers, you all probably know about the famous Żabka hot-dogs. In case that’s you and the nearest Żabka near you happens to be Nano, fear not. You can still get your hot-dogs. The store offers an interesting service for the hot-dog lovers, where you have a robot that grills and assembles the whole hot-dog right in front of you. Keep in mind though that it may not always succeed as it’s assembled by a robot and not a person, so it has its failures at times. The robot is called Robbie and its software has over 15000 lines of code. The order of your snack takes less than a minute (unless you’re indecisive) and the display next to the hot-dog stand shows you the ordering options where you choose the size, type of the sausage and the sauces you wish to have in your hot-dog. After that, you just wait and watch Robbie prepare your meal.

Żabka inwestuje w roboty w sklepach autonomicznych, w jednym z nich działa  robot przygotowujący hot dogi | Donald.pl

Onto the security stuff, how does the store know what you take, and how does it not charge you extra? The store doesn’t actually use any shelf-sensors. It does have cameras inside the store, just like all the other stores, but this time, they’re used for camera-vision technology provided by AiFi. To respect the customer’s privacy, customers are identified by the system without the use of biometrics or facial recognition. The cameras only track the movement of the client’s, while the product recognition is performed by AI computer vision. Using key point tracking technology, the system creates a stickman from each customer who enters the store, to which the system supports three sessions running simultaneously, each with a small number of participants.

All in all, Żabka Nano is a great innovation and it does help those who crave a snack anytime after 11PM- after all the normal Żabka stores close. It’s interesting to see how the technology improves everyday and how the world easily adapts to it. Do you think Żabka Nano is a good innovation? Or maybe a beginning to something bigger in the future?

Pierwsza Żabka Nano otwarta w Decathlonie

Netflix officially branches out into the fitness world.

Reading Time: 3 minutes

Gather round all Netflix users. It’s your time to finally make better use of your Netflix subscription and instead of binge-watching various series, start to move your body (This isn’t about Netflix & Chill activities).

Nike Training Club Launches On Netflix – Deadline

Upon the surprising news, Netflix has announced their collaboration with Nike Training Club (NTC), where you will be officially offered an overall of 30-hour exercise sessions divided into two batches. Nike Training Club has great home-workouts on various difficulty levels, starting from “mild burns” for beginners to intense workouts for all. Needless to say, the majority of NTC users proudly express their satisfaction with the app, which may lead the collaboration to a great success (or at least slightly greater success than their previous collaborations).

The said batches have a set date, and upon a pleasant surprise for some, the first batch has already been released on December 30th 2022 and is available in every region that Netflix operates in! The said batch includes 45 workouts, ranging from cardio, yoga, pilates, core-strength, weight training and pretty much anything else you can think of (related to fitness of course). If you’re a fitness-freak and want to take on a challenge of doing the whole program ASAP, do not worry, the second batch may release quite soon- perhaps even in the first half of 2023, either way, regardless of whether it does release soon or not, it’s unlikely that you’ll get bored of the content you’re already given as it may genuinely be good and you might even see some change in your strength or see positive progress regarding your weight goals or body visuals, hence you may wish to do the program again, at least until the second batch releases.

Netflix – Aplikacje w Google Play

The collaboration has been inspired by the pandemic and the restrictions that have been established during that time, amongst which, one was the prohibition of going to the gym. This has caused a major problem in the society, as people were locked in their houses with no ability to train or workout in public, which resulted in a major increase in obesity amongst people. That’s where many have gone through a home-workouts phase, to the point where it became a trend to show off your progress of before and after the lockdown, along with recommending or filming special home workouts that have helped people lose weight. As such event was so popular all around the world, Netflix felt the need to take their platform to a next level and bring back the trend of a healthy at-home lifestyle, with workouts performed by professionals (whether it be yoga instructors or bodybuilders).

Tips For Exercising at Home During COVID-19 Pandemic - Scripps Health

Netflix is slowly expanding the focus of their platform to care for the well-being and self-improvement of their users. It has previously partnered with Headspace, a meditation-related platform, which allows the users to access guides into meditation and various other mindfulness-related content. If you wish to access it and check it out yourself, all you have to do is look up “Headspace” in the Netflix search bar.

Poradnik Headspace: Medytacja - #TeamRodzina

Netflix has a perfect timing for the release of the new content. Since the new year has just begun, a lot of individuals have written their New Years Resolutions, amongst which, many probably wish to begin their fitness journey, and they may not know where to start, or they may struggle with social anxiety. The collaboration of Netflix and NTC allows those people to achieve their dream goals in the comfort of their own home.

The idea seems promising and has a bright future ahead of itself. Also, whilst being a Netflix-user you get a whole wellness package included in the price. Why not try it for yourself? If your new years resolution is to change your lifestyle and start working out, look up “Nike” in the Netflix search bar, and enjoy the various workouts Netflix has prepared for you.

Sources:

https://techcrunch.com/2022/12/21/netflix-branches-out-into-fitness-content-nike-training-club-classes/

https://www.netflix.com/pl/

Romantic relationship between AI-asisstant and designer

Reading Time: 3 minutes
AI maven AiDA unveils the future of fashion at Hong Kong show | Life
https://www.reuters.com/technology/hong-kong-designers-try-out-new-assistant-ai-fashion-maven-aida-2022-12-27/

The fashion industry is one of the largest in the world, with revenues anticipated to reach $2.25 trillion by 2025.

According to reports, the global market for AI in the faction industry was $270M in 2018 and is expected to reach $4.4B by 2027. Which encourages to get in or act in it.

AI in design

AI-enabled systems can produce clothing designs utilising information such as photos from the brand’s past products or those of other designers, information about customer preferences (colour and style choices), and current fashion trends.

New innovative tool AiDA (AI-based Interactive Design Assistant), developed by students and researchers, collaborated on Hong Kong’s Fashion X AI fashion show. Featured 80 outfits from 14 designers.

“Our relationship is comparable to a romantic relationship in that I seemed to gradually get to know her (AiDA), and she gradually got to know my own designs,” Mountain Yam (one of designers) said. “In accordance with my lines, styles, and databases, the system will propose something for me that I may not have ever considered, but she (AiDA) thinks is suitable for me. Therefore I believe we are developing a long-term relationship.”

According to AiDLab CEO Calvin Wong the AIDA system is supported by AI technologies such as image recognition, detection, and image generation. Designers may upload drought sketches, materials, and color palettes to a virtual mood bard, and the software’s algorithm generates blueprints that they can modify and personalize. The technology can generate a dozen fashion patterns in 10 seconds is sort of damn fast idea machine, saving designers valuable time and being a second brain.

The industry is evolving, and new AI-related programs are giving the faction industry more sophisticated feel.

As a result, how many new stylists will be required? AI will be able to generate thousands of ideas and collections based on a certain designer’s style, which would take a lot of time for a regular person. Only the best of the best will be remaining, because many companies would want to replace them with a more efficient tool.

Measuring the human body – By taking a few photos you can create a 3D model of your body, who will be able to try on clothes and see how they fit you. And here the question arises whether the companies will not manipulate the algorithm so that the clothes look better than in reality ?And since it already came and you paid for it, you will be less inclined to give it back and the company will make money. Or wants them to be as fit as possible to increase customer satisfaction and reduce returns with app like Virtusize?

AI understand what human want humans need to wear what looks good on certain people and what people might like. It’s a new professional fashion stylist. Now also ordinary people can have access to it . With the help of artificial intelligence we will are able to create clothes based on our style, the music we listen to, and other preferences. An app like StitchFix or  Finery will help you match outfits from a variety of brands.

Fashion is one of the world’s most polluting sectors, and various companies are experimenting with technologies to make it more eco-friendly and sustainable. Nowadays, 3D clothing (The Fabricant) available only in the virtual world in which we spend more and more time every year is becoming a growing trend. In the DressX application you can buy various clothes and you pay to put them on a photo that you can share on social media and note they cost the same as normal and you can “wear” them only once. Would you guys opt for such clothes?

Already in the near future you will be able to buy elegant clothes for your character to microstft teams or zoom meeting while sitting in pajamas, sounds tempting.

https://www.reuters.com/technology/hong-kong-designers-try-out-new-assistant-ai-fashion-maven-aida-2022-12-27/

Three AI And Tech Trends That Will Transform The Fashion Industry (forbes.com)

https://timesofindia.indiatimes.com/blogs/voices/the-era-of-the-artificial-intelligence-fashion-designer-is-getting-closer/

Misogynistic tendencies in AI

Reading Time: 3 minutes

In recent years, artificial intelligence has come under fire for its role in perpetuating and amplifying misogyny. This is primarily due to the fact that AI is often created and trained by male developers, who inadvertently imbue their own biases into the algorithms. As a result, AI systems have been found to display sexist behaviour, such as calling women ‘cooks’ and ‘nurses’ while referring to men as ‘doctors’ and ‘engineers’.

Adele and Betty Friedan as imagined by the Lensa AI.

Sexist language
There are a number of ways in which AI can be misogynistic. One of the most visible ways in which AI perpetuates misogyny is through the use of sexist language. This was most famously demonstrated by Microsoft’s chatbot Tay, which was designed to learn from interactions with users on Twitter. However, within 24 hours of being launched, Tay began tweeting out sexist and racist remarks, which it had learned from other users on the platform.

While this was an extreme example, it highlights the fact that AI systems can easily pick up and amplify the biases that exist in the real world. If left unchecked, this can lead to a reinforcement of sexist attitudes and behaviours.


Algorithmic bias
Another way in which AI perpetuates misogyny is through the use of algorithms. These are the sets of rules that determine how a computer system behaves. Often, these algorithms are designed by humans, who may inadvertently introduce their own biases.
For example, a study by researchers at MIT found that facial recognition systems are more likely to misidentify women as men than vice versa. This is because the system had been trained on a dataset that was predominantly male. As a result, it learned to associate male faces with the concept of ‘person’ more than female faces.
This kind of algorithmic bias can have a severe impact on the real world, as it can lead to women being denied access to certain services or being treated differently by law enforcement.

Data bias
Another issue with AI is that it often relies on data that is biased. This can be due to the fact that the data is collected in a biased way or because it reflects the biases that exist in the real world.
For example, a study conducted by Ellen Broad, an expert in data sharing, infrastructure and ethics, found that Google Photos image recognition system is more likely to label pictures of black people as ‘gorillas’ than pictures of white people. This is because the system had been trained on a dataset that was predominantly white. As a result, it learned to associate black faces with the concept of ‘gorilla’ more than white faces. This kind of data bias can lead to AI systems making inaccurate and potentially harmful decisions. For example, if a facial recognition system is more likely to misidentify black people as criminals, then it could lead to innocent people being wrongly arrested.

Brandee Barker’s Twitter post

Moreover there’s something deeply troubling about the way AI is being used to create ‘portraits’ of people, particularly women. In the case of Brandee Barker, an AI created deeply sexualized versions of the woman.
This isn’t just a case of bad taste or something that can be chalked up to the ‘uncanny valley’ effect. There’s a more sinister element at play here: the objectification and sexualization of women by AI.
It’s not just Barker who has been rendered in a sexualized manner by AI. In an essay for Wired, the writer Olivia Snow wrote that she submitted “a mix of childhood photos and [current] selfies” to Lensa AI and received back “fully nude photos of an adolescent and sometimes childlike face but a distinctly adult body”.
The AI-generated images of women are eerily realistic, and that’s what makes them so troubling. They look like real women, but they’ve been created by machines with the sole purpose of objectifying and sexualizing them. This is a scary prospect, because it means that AI is perpetuating and amplifying the misogyny that already exists in our society.

Addressing the issue
Given the potential impacts of AI-perpetuated misogyny, it is important that the issue is addressed. The solution to this problem is not to try to create AI that is gender-neutral. Instead, we need to ensure that AI systems are designed and built with the needs of all users in mind. This includes ensuring that a diverse range of voices are involved in the development process and that training data is representative of the real world. Only by taking these steps can we create AI systems that are truly inclusive and beneficial for everyone.

Sources:

Lensa AI app

https://www.theguardian.com/us-news/2022/dec/09/lensa-ai-portraits-misogyny

https://www.adweek.com/performance-marketing/microsofts-chatbot-tay-just-went-racist-misogynistic-anti-semitic-tirade-170400/

https://news.mit.edu/2018/study-finds-gender-skin-type-bias-artificial-intelligence-systems-0212

Tagged , , , ,

How can a VPN hide you?

Reading Time: 2 minutes
https://i.iplsc.com/sieci-vpn-umozliwiaja-zdalnym-klientom-laczenie-sie-z-prywat/000AGCCWMXS9TFMV-C122-F4.jpg

A virtual private network (VPN) is a type of technology that allows you to create a secure connection over a less-secure network between your computer and the internet. When you use a VPN, all of your internet traffic is routed through an encrypted tunnel to a server controlled by the VPN provider. This makes it much more difficult for anyone on the same network as you, or for any external observers, to see what you are doing online or to track your online activity.

One of the main benefits of using a VPN is that it can help to protect your online privacy and anonymity. When you use a VPN, your internet traffic is encrypted, which means that it is scrambled and unreadable to anyone who might intercept it. This makes it much more difficult for anyone to track your online activity or to determine your location.

Additionally, when you use a VPN, your internet traffic is routed through the VPN server, which means that your IP address is hidden. Your IP address is a unique numerical label that is assigned to your device when you connect to the internet, and it can be used to identify your device and location. By hiding your IP address, a VPN helps to further protect your online privacy and anonymity but there have some country which is not allowed to use VPN because of national security such as China., North Korea, Belarus, Oman, Iraq, and Turkmenistan It can indicate a VPN ability to remain anonymous very well.

Overall, a VPN can help to hide you by encrypting your internet traffic, hiding your IP address, and routing your traffic through a secure server, making it much more difficult for anyone to track your online activity or determine your location.

Resource:

https://www.le-vpn.com/are-vpns-legal/

https://www.webhostingsecretrevealed.net/blog/security/are-vpns-legal/

https://www.kaspersky.com/resource-center/definitions/how-does-vpn-keep-me-safe-online

instagram shadowbanning and hidding posts.

Reading Time: 2 minutes

Shadowban is a term at the sound of which content creators get uncomfortable feelings. It’s a very frustrating thing , but what does it have to do with business and application development?

With one of the newest instagram updates, anyone using proffesional account will have ability to chceck wheater their posts are being shadowbanned or unable to reach the “explore” area. The whole procedure is described by CEO himself (i will link the video 🙂 ) . Finally the problem has been partly solved.

People are getting shadowbanned by violating terms concerning amounts of posts , followers likes etc. it was developed to get rid of bots , spammers and accounts that are buying followers to gain more reach. Although the problem has spread all across the regular instagramers.

For many years, this seemed to be a serious problem due to misconceptions about posting. The number of cases where ordinary people as well as full-time developers got shadowban without their knowledge is insane. The fact alone that anybody could recieve restrictions without their knowledge is ridicolous itself, and i understand that everyone needs to agree to instagram terms before using the app, but wouldn’t it be a more reasonable option to inform and remind users about them violating the terms , before applying the “hidden punishment” ?

This update is crucial for activists , reporters , humanitarian crisis posts or even artists. In each category there might be a lot of misunderstanding or agorithm fails that could cause in account block or ban .

Instagram Will Start Blurring 'Sensitive' Content On Its Platform |  HuffPost UK Tech

For example things like war-photos could be perceived as violence or body exposure in art-photography may be interpret as pornography. The border is very subjective and in many cases it should be reviewed by a special consultant.

It’s hard to call shadowbanning or sensitive content patches as censorship , because this devices were invented to save content and make a safe platform , not to stand against it. Unfortunatelly in cases that i described above the situation is difficult and i hope that IG will find a method to solve the problem so that everybody could post without worrying about their account being unfairly dispended. Tell me what do you think about it …

sources :

https://hypebeast.com/2022/12/instagram-shadowbanned-posts-instagram-update

https://img.huffingtonpost.com/asset/58d53ffa14000020000715de.png?cache=xQvFtGSIly&ops=crop_0_143_455_574%2Cscalefit_720_noupscale

Glasses that let deaf people read conversations

Reading Time: 3 minutes

“Artificial Intelligence will make people’s life easier”. I have seen this phrase lots of times. And once again I am convinced that this is true. 

Have you ever thought about how hard it is to be deaf? Having fewer job opportunities, social withdrawal, and emotional problems caused by a drop in self-esteem and confidence. And now imagine the possibility to READ speech in real-time by just wearing glasses. This is what augmented reality (AR)-powered smart glasses can allow. The XRAI Glass software converts speech into subtitles which appear on the user’s glasses screen. Moreover, it is tethered to the live-captioning app that translates over 9 languages, with more coming in the next few months. The translation takes place near-real time, allowing users to keep up with the conversation. All processing is done on the mobile phone so the glasses themselves are lightweight and comfortable to wear all day long. The glasses can be plugged even into a TV to display subtitles. In fact, one of the XRAI’s trial group, a lawyer, is using the glasses plugged directly into the mic system of the court. 

HOW HAS THE IDEA EMERGED?

The XRAI Glass technology was inspired by one of the co-founders’ 97-year-old deaf grandfather: “If he’s enjoying subtitles while watching television, why can’t we subtitle his life?” This is how the history of XRAI Glass started.

PRICE

The live-captioning app is now free to download, but users will be asked to pay for premium features if desired (up to $49.99/month). The glasses separately cost $484 and can be purchased on EE or Amazon. 

400 XRAI smart glasses turn audio into captions to let deaf people 'SEE'  conversations | Daily Mail Online

DRAWBACKS

The company is planning to reach 70,000 people by the end of 2023. However, some improvements should be made before these glasses are incorporated into daily usage. Some people state that glasses may struggle to understand group conversations where people are speaking over each other. Aside from this, a quiet environment is a prerequisite for accurate speech interpretation. That is why in July 2022 XRAI Glass launched a public trial period where people can try and share their experiences using smart glasses. The company aims to expand AI capabilities even more to provide the best service it can. 

PERSONAL OPINION

What I like the most is that people will never know what are you looking at while talking with them. Those who are suffering from hearing loss will be able to feel equal to other people and will not now see themselves as “exceptional” or “unusual”. Moreover, these glasses are not a good solution just for deaf people, but also for those who are facing the problems of consuming information in an audial way. That means that they will be able to hear and read speech simultaneously wearing smart glasses. 

I watched a couple of videos where deaf people are trying these glasses for the first time and was touched to see them broken down in tears by their hope for a better, easier life. I am happy that modern technologies are able to facilitate the daily life of people with hearing disabilities and even better solutions will come in the future as AI is progressing at unbelievable speed.

References:

https://metro.co.uk/2022/08/01/smart-glasses-allow-deaf-people-to-see-conversations-with-subtitles-17106685/

https://metro.co.uk/2022/11/17/we-tried-the-smart-glasses-that-let-you-see-conversations-17764768/

https://www.hackread.com/ai-powered-smart-glasses-deaf-speech/

https://www.businesswire.com/news/home/20220728005555/en/XRAI-Glass-Revolutionary-New-Glasses-Allow-Deaf-People-and-People-Who-Have-Hearing-Loss-to-%E2%80%98See%E2%80%99-Conversations

https://www.cbsnews.com/miami/news/captioned-smart-glasses-let-deaf-people-see-rewind-conversations/

Tagged