Discover Top Posts Tagged with #neuromorphic computing

Popular Recent

Artificial intelligence and high-performance computing are driving up the demand for massive sources of energy. But neuromorphic computing,

Artificial intelligence and high-performance computing are driving up the demand for massive sources of energy. But neuromorphic computing, which aims to mimic the structure and function of the human brain, could present a new paradigm for energy-efficient computing. To this end, researchers at Lawrence Livermore National Laboratory (LLNL) created a droplet-based platform that uses ions to perform simple neuromorphic computations. Using its ability to retain short-term memory, the team trained the droplet system to recognize handwritten digits and play tic-tac-toe. The work was published in Science Advances.

#Science #Physics #Soft Matter #Neuromorphic Computing #LLNL #Lawrence Livermore National Laboratory

Beyond silicon: These shape-shifting molecules could be the future of AI hardware

For more than 50 years, scientists have searched for alternatives to silicon as the foundation of electronic devices built from molecules. While the concept was appealing, practical progress proved far more difficult. Inside real devices, molecules do not behave like simple, isolated components. Instead, they interact intensely with one another as electrons move, ions shift, interfaces change, and even tiny differences in structure can trigger highly nonlinear responses. Although the potential of molecular electronics was clear, reliably predicting and controlling their behavior remained out of reach. At the same time, neuromorphic computing, hardware inspired by the brain, has pursued a similar goal. The aim is to find a material that can store information, perform computation, and adapt within the same physical structure and do so in real time. However, today's leading neuromorphic systems, often based on oxide materials and filamentary switching, still function like carefully engineered machines that imitate learning rather than materials that naturally contain it.

Understanding the Anatomy of Future Hypothetical Robots

By Arjuwan Lakkdawala

Ink in the Internet

I’m taking a radical look at future hypothetical robots. Sort of like a Frankenstein of modern science. A new era of tech is coming and have no doubt it’s going to turn the world upside down.

As a tech enthusiast and someone who is actually rather frightened of how far scientists are pushing the threshold of “progress.” I found myself thinking about that in the future which is happening before me, and that which one might say is impossible to happen.

I’m going to write about both prospects because I think each is a variable important to evaluate in order to deduce the mental and material outcome of the present state of things in this regard.

What my eyes see: Neuromorphic computing it appears is almost all set to eclipse deep learning, machine learning, take Artificial Intelligence to a whole new level.

Basics of it: a memristor sits in the heart of each chip, which has synapses that simulate the passing of information like in the human brain. Please note: the “simulation” is crude and in its infancy, because the human brain is not a solid electrical device.

In our brains and down to our spinal cord there are billions of neurons with chemical and electrical signal processing synapses. That send commends to our whole body.

The electrical synapses which our brain uses less often than the chemical ones, are responsible for our reflexes among perhaps other things. Like when you’re suddenly about to trip and fall but on reflex gain your balance. You didn’t think of balancing yourself, but that super instant decision made with lightning speed that saved you is the work of the electrical synapse.

Leading tech companies are trying to make the chips smaller.

The idea is that the processing and memory components on the memristor, work with the least energy consumption, and highest speed possible.

Presently in computers the processor and memory are located separately and the flow of electricity which sends signals to transistors in the binary language of 1s and 0s doesn’t permit for programming which could let Artificial Intelligence simulate any form of self-derived creativity or interpretation.

Deep learning which so far is considered the ultimate edge of AI is still based on logical computation but with layers of data from which the algorithm outputs the highest number of matches.

This logic based thinking which is an artificial simulation of how the left side of our brain operates, is what scientists want to surpass into the territory ruled by the right side of our brains – creativity.

It is believed the Neuromorphic chip is the first step in this direction.

However, don’t get confused by the difference in the Neuromorphic chip’s ability to process information and the programming needed for it to get “creative.” Artificial Intelligence is still hungry for human produced data and programming.

Neuromorphic computing is simply the beginning of a clever way of processing information. But it does seem soon enough it’s going to become a needed option.

Because according to Moore’s Law (the doubling of transistors every couple of years) means we are going to reach a point where the transistors produced won’t be enough to process rapidly growing new data in the world.

Suppose Neuromorphic computing improves to somewhat science fiction levels, in this case it would be a good summary of the brain of our hypothetical robot.

I’m certain it won’t actually go in the robot’s head space. Because it’s not necessary.

Cloud computing is so much more convenient.

Next for its eyes just imagine Google Glass. (You get the picture)

Hearing: Ultra-sensitive, can pick up frequencies humans can’t.

Muscles: depending for what the robot is used for, it could be soft plastic, or titanium metal.

What about the hands and touch? Here is where my fear lies. Over ambitious scientists can’t stop making robots that are “human-like” or animal-like” the purpose of these isn’t to replace humans or pets, but for people who prefer to be served by robots in restaurants, in sales, etc.

As for pets it’s thought “robotic pets are so much easier to care for.”

This is the biggest fault in the idea. There is no caring, there is no sense of guilt if it breaks, no pity if you haven’t fed it for days.

I recently saw a video on Youtube in which a man beats a humanoid looking robot, and kicks a dog-like robot. The footage meant to express that treat them as you want. This thing looks like a human but it’s not, and this looks like a dog but it’s not.

As someone who grew up taking care of actual pets, and cried rivers when they died. This footage doesn’t change a thing about how I feel about real humans and animals.

But imagine if you’re a kid born in this tech era. And you grow up with a mechanical pet dog. Won’t it distort your balance of touching with care, in comparison to rough touching?

Won’t it distort your sense of humanity of what hurts and what doesn’t?

The obsession to make robots human-like is so bad. I wouldn’t be surprised if future robots had human hearts grown in human tissue sacks, connected to its brain via prosthetics.

This too is serious bad news. Because if the robot is indeed given some form of artificial intelligence independence.

I wouldn’t expect it to have good feelings. Humans are the best example of how hard it is to fight evil temptations.

Arjuwan Lakkdawala is an Author, Science Communicator, and Journalist. @Spellrainia Twitter Instagram

References:

Brain-Like (Neuromorphic) Computing - Computerphile

Neuroscience 2nd. Edition (book)

Exam Sam (website) Neurotransmitters, Reflexes, Synapses.

What is Cognitive Computing - and Futurlogy (YouTube)

Recent research has found that the construction of logic gates from light is, in fact, possible. This has been done by ‘bouncing’ the light between one wavelength and another.

Optical computing can conceivably move at the speed of light. A group of UCLA researchers has demonstrated that it can also be used to construct trainable neural networks.

#computing #neuromorphic #neuromorphic computing #light #optical computing #neurons #deep neural network #artificial synapses #refraction #training #3D printing #3D-printed #speed of light

Air Force Tests IBM’s Brain-Inspired Chip as an Aerial Tank Spotter

Chips with silicon “neurons” could make satellites, aircraft, and drones smarter.

Satellites, aircraft, and growing numbers of drones—the U.S. Air Force has a lot of electronic eyes in the sky. Now it’s exploring whether brain-inspired computer chips could give those systems the smarts to do things like automatically identify vehicles such as tanks or anti-aircraft systems.

#neuromorphic computing #ibm

Light-sensitive materials mimic synapses in the brain

An interdisciplinary research team has engineered a new class of organic photoelectrochemical transistors (OPECTs). These tiny devices can convert light into electrical signals and mimic the behavior of synapses in the brain. The research results have now been published in the research journal Advanced Science. The team included Professor Francesca Santoro and Dr. Valeria Criscuolo from the Institute of Biological Information Processing—Bioelectronics at Forschungszentrum Jülich, in cooperation with colleagues from RWTH Aachen University—Professor Daniele Leonori and Junior Professor Giovanni Maria Piccini (now University of Modena and Reggio Emilia). Our brains work by passing signals between nerve cells, adapting over time to learn and remember. Scientists are trying to recreate this kind of behavior in electronic devices, a field known as neuromorphic electronics. One way to do this is by developing materials that can "learn" in similar ways to how the brain does.

Bio-inspired nanochannels provide experimental evidence for uncovering brain memory mechanisms

A research team from the Institute of Modern Physics of the Chinese Academy of Sciences and Lanzhou University has obtained important experimental evidence for revealing brain memory mechanisms and developing a new type of neuromorphic computing. Their findings are published in Advanced Functional Materials. Human learning and memory originate from highly dynamic connections between neural synapses. These structures act as intelligent switches that transmit signals and undergo dynamic remodeling, thereby enabling the encoding, storage, and retrieval of information and serving as the biological basis for cognition and behavioral adaptation.

Tuning magnetism with voltage opens a new path to spintronic neuromorphic circuits

A team of researchers has discovered a new way to control the magnetic behavior of quantum materials using applied voltages. Specifically, the material lanthanum strontium manganite (LSMO), which is magnetic and metallic at low temperatures but non-magnetic and insulating when relatively warm, can be influenced by voltage. The work is published in the journal Nano Letters. Quantum materials like LSMO are materials that possess special properties because of the rules of quantum mechanics. Researchers discovered that applying voltage to LSMO in its magnetic phase causes the material to split into regions with distinct magnetic properties. The magnetic properties of these regions depend on the applied voltage. This is important because normally, magnetic properties don't respond to voltage.