AI Future Technology Science

The Future of artificial eyesight is here: EC-I

You know what is the most complex part of the human body, yes it’s the human eye. Researchers have been working on reproducing the same. Now, they have made artificial eyesight possible called EC-I, which has the ability to see better than the real thing.

Bionic eyes has the potential to restore eyesight to people who have lost it and even to those who never had it.

Currently, the most advanced versions are from companies such as Bionic Vision Australia and Second Sight, which have already been implanted into patients.

These devices are made of a pair of glasses with a camera in the center. That data is processed by a small unit worn out of the body, then sent to the user’s retina for implantation.

From there, signals are transmitted to the visual centers of the brain.

And they work. Users have reported being able to see the world for the first time in years. Unfortunately, this artificial eyesight is not enough for them to rely on to navigate the world.

Some studies have shown that this type of bionic eye can produce strange images and are too slow to capture sharp movements. 

But this new device can be improved. A team led by scientists from the Hong Kong University of Science and Technology (HKUST) has developed what they call Electrochemical Eye (EC-Eye).

artificial eyesight
A cross-section of the make-up of Electrochemical Eye (EC-Eye)

Instead of using a camera-like two-dimensional image sensor, the EC-I is designed with a concave curve following an actual retina.

This surface is studded with an array of small light sensors designed to mimic photoreceptors on the human retina.

These sensors are connected by a wire made of liquid metal, which then acts like an optic nerve.

The team tested the EC-I and showed that it could capture images much earlier than expected.

It was set up to display large individual letters in front of a computer screen, and it was able to display them clearly enough to be read.

Although this is a huge improvement over existing bionic eye designs, EC-I vision has a long way ahead to achieve the natural human eyesight. But, the team says, this may not be the case forever. 

The technique has the ability to outshine the real thing by using a dense array of sensors and attaching each sensor to a separate nanowire.

The team even goes so far as to say that using other materials in different parts of the EC-I can provide users with a higher sensitivity to infrared – essentially, night vision.

Of course, a lot of work remains to be done in the future, but the EC-I is promising.

The research was published in the journal Nature. The device can be seen testing its eyes in the video below.

Electrochemical eye (EC-eye) -Eye images sensory images

AI Future Technology How to

Latest text generator through voice and Open AI

Open AI has quietly unveiled the latest incarnation of its title-grabbing text generator: the GPT-3. The research lab initially stated that its predecessor’s ability to spread disinformation made it too dangerous to share. The decision warned journalists of imminent robot apocalypse – generating very useful publicity for the GPT-2. Now, Open AI has unveiled its elder brother. And it is very big. The language model in the GPT-2 has 175 billion parameters – 10 times over 1.6 billion, which was considered heavy upon its release last year. GPT-3  ‘s research paper has also outgrown GPT-2 from 25 to 72 pages. We haven’t got through the whole thing yet, but after flicking through have spotted some striking stuff.

Bigger and better?

GPT-3 can perform an impressive range of natural language processing tasks — without needing to be fine-tuned for each specific job. It’s now capable of translation, question-answering, reading comprehension tasks, writing poetry — and even basic math:

The model can perform three-digit addition and subtraction. Credit: OpenAI

It’s also pretty good at bettering correcting English grammar:

Credit: Open AI

GPT-2 also seems to have improved upon the vaunted writing ability of its predecessor. The research team tested its skills by asking evaluators to distinguish its works from those created by their humans. The one they found most convincing was a thorough report on a historic split of the United Methodist Chuch:

The evaluators were most impressed by this article on a church split. Credit: Open AI

However, my favorite example of its writing was the one that humans found the easiest to recognize as made by a machine:

Joaquin’s shape-shifting claims didn’t convince the critics. Credit: Open AI

That report may not have convinced the reviewers, but it certainly showed some flair and a capacity for the surreal. By comparison, here’s an example of a GPT-2-penned article that Open AI previously published:

GPT-2 did a decent job reporting on the discovery of talking unicorns. Credit: Open AI

GPT-3’s ability to generate multiple paragraphs of synthetic material that people find difficult to distinguish from human-written text … raises concerns in this regard. However, the system is unlikely to take the jobs of two-bit hacks, for now, thank God. Not because it lacks the skill — it’s just too damn expensive.

That’s because the system needs enormous computation power. The CEO of AI communications firm Hyperia, Elliot Turner explained:

“Reading the OpenAI GPT-3 paper.  Impressive performance on many few-shot language tasks. The cost to train this 175 billion parameter language model appears to be staggering: Nearly $12 million dollars in computing based on public cloud GPU/TPU cost models (200x the price of GPT-2)”

That should also reduce its powers to be used for evil, as presumably, the only people who could afford it are, er, nation-states, and multi-national corporations.

For now, we’ll have to wait and see what happens when the model’s released to the public

AI Future Technology How to Science

Using your Voice as Passwords: May become reality

How efficient is it to use voice recognition as a password? Let’s find out. Smartphones are getting smarter and faster with each passing day and so is their safety.

People love it when they include new features to unlock their brand. Apple shook the world with its FaceID when it debuted.

The USA and China are so advanced in technology that they can recognize your face even when you are wearing a mask.

It seems that Face ID will maintain its awe for at least the next few years. But using voice recognition as a password is also able to achieve the same importance.

A new feature of Google Assistant

Google Assistant

Google is slowly sliding a new feature in its AI assistant. Before we proceed to this discussion, you should know that it is not yet universally available.

But there is a chance that all your payments and purchases may soon be certified with a voice match. You can find this feature on both Android and iOS in a very short time.

Right now Google plans to limit this feature to only one function. It may be able to make Google Play purchases and also take restaurant orders.

Is it a new thing?

No matter what feature it is, it has been among us for a few years now. Google tried to implement a voice match in Google Pixel 2 and its home speakers.

It was back in 2017. But this was not so accurate so Google did not yet want to use voice recognition as a password since it can only match your voice with personal content and nothing more.

Each time it recognizes your voice it can boot up your personal calendar, email, and other similar things. Google is working really hard to achieve its voice recognition accuracy.

The future

Futurists predict that future payments may be fully automated. If you want to certify a payment, all you have to do is say “pay $10”.

You do not need to manually type in card numbers and struggle to get a one-time password. Voice authentication does not require any new hardware, it reduces cost which makes it scalable.

Thumbprint or facial recognition requires hardware improvements. While with your voice, you can achieve this with some software updates as well.

This not only reduces the cost but also reduces the time required for its implementation.

So what’s holding it back?

Any technology that makes its way into our lives has a downside. Using voice recognition as a password is hard to achieve.

The user can set up the feature in a quiet environment that may not work when he tried to access it in noise. Some users try to sound correct when installing.

But recording in a natural voice and place is the only way for you to get the best possible result. The system might detect your voice when you are sick or when you are old and your voice becomes deeper.

The conclusion

We are still a long way from using voice recognition as a password. But this is something that we can actually accomplish. As impossible as it sounds, we are really close to making it a reality.

No technology is perfect when in debuts. But with time and user feedback, voice authentication can completely change our perception of security.