Interlinked.

Caveat: This is old, please check the new version.


1. Time Machine

From my childhood I have been fascinated with cells. I remember pricking my finger using a safety pin during 2nd grade with my friend to try out her new microscope. Sadly, we couldn't get any blood out ( cursed by parents, I guess ), so we took drops of water, oil, juice and anything we came across ( such as lasagna bits, Kinder Überraschung1 (it's so so good, isn't? ), leaves, etc ) and experimented.

We didn't know how to prepare microscope slides so many weren't clear but even the blurred picture showed something magical.

It was amazing.

The microscopic world. This hooked me up to learn about plants cells, DNA and everything biology. I would dream to be a doctor, a nephrologist (didn't know this involved urine and stuff alike for I was a kid! and thought anything that had a prefix, 'ne/nu-' sounded cool, I still think so, ;) ).

So a nephrologist, neurologist, nuclear medicine specialist, neonatologist etc. Those would define me, I had dreamt.

Neuroscience though stuck for a long time. It still burns brightly. Anyways, let's move on.

I'm a big science fiction fan and grew up watching Star Trek, E.T, Star Wars, Jurassic Park etc and being at Germany during the rise of the Personal Computer (PC), I had access to a computer back in 2000s, a Pentium II with 128MB RAM, Windows 95/Millennium and most importantly, Internet or Ethernet Cable2 which made my cravings and the flame for everything technology and science never extinguish.

Back then, the Internet was a small place, peaceful without the unnecessary cookies, far away from those redundant ads. It was just like a private island you bought off somewhere in the Pacific and are exploring slowly and trying to connect with your other island neighbours.

It was a golden time learning science, reading articles and checking mail, searching everything in Altavista3. I remember sending web greeting cards through Yahoo which was special. The card was electronic but still carried some emotion that would bring the festive joy at the other end.

Video games. I was addicted to flash browser based video games. I remember playing Lego Ninjago, Samurai Jack, Mini-golf and many others.4

I was so happy when I got Toy Story 2: Buzz Lightyear to the Rescue as a Christmas present back in 2002. That game is a classic.

My dad is a math professor5 so for bedtime stories I had math history and problems and sometimes, no story at all. My interest for mathematics was in the blood, just wasn't awake till a year where I would find it the most beautiful subject to explain the universe.

The beauty I find lies in the definitions which birth proofs to justify the rationale, the logic that fits so perfectly across the Universe.

As predicted by nosy relatives, computer access from childhood put me in a serious relationship with the our computer. I used to sit most of the time surfing or playing stuff. I learnt Logo and Basic. I loved the in built calculator of Windows Millennium. ( My parents used to hate and love this relation though and still do )

The computer tricked me to forget what my heart set it's sight in the first place. Biology and neuroscience.

Till 2010.

2. The sleepover of the century.

A night party with friends ended with all of us watching our favorites movies together.

This is when I saw, I, Robot, the Will Smith movie adaption of Isaac Asimov's6 book having the same name.

I, Robot.

Over the next few months, I buried myself into books and probably created a minimal outrage at Google servers to learn about Artificial Intelligence, can AI kill?, how can we create emotions, robotics. ( and Space Exploration, The Mughal Empire, Calligraphy. Blame the hyperlinks.)

Fast forward 10 years and I started college majoring in computer science to learn more about computers in general.

I, Computer.

I graduated this year and probably I'm ready to write down my fundamentals, my research statement that I would devote my entire life for making it possible.

As a side note and clearing your doubt beforehand, (you might ask me about this) my resume says much about augmented reality for education. How is it aligning you would ask? I argue that bringing technology (AR) to the kids and playing with them using AR helped me understand psychology of children, why visual objects are important and that interactions help to grasp ( lol ) concepts faster.

3. Behave.

Recently, I had a great time while reading Behave, the biology of humans at our best and worst, by Robert Sapolsky, the MacArthur Genius Awardee, Stanford Professor of Neurobiology and Biology and an amazing science writer.

He says about emotions, in particular ( aggression leading to ) violence, empathy, co-operation and competition in humans which can be only defined (properly) by biology.

I'm very hopeful that in my lifetime that I'll live in a world where artificial general intelligence (AGI), exists. AI to play, help, make life easier and efficient, solve crucial problems and maybe just sit and watch the pinkish orange sun going down in awe. ( Vaporware Aesthetics, anyone? )

4. Ready to Love Again.

To be human, an AI needs emotions, they need to fight, show love and empathy, help each other and compete. They need to dream. Every AI needs to be unique and have a vision for being valued.

Isn't a pretty sight to think about?

Yes, it is. I want to paint this picture, this vision.

I want to learn about the biology of emotions, the physics of consciousness,7 the nature of the human brain and behavior, the science behind thought and feeling.

5. The Vision

I want to recreate everything that biology does now, computationally and mathematically, bring emotions to AGI, bring feelings, thought and dreams to our new friends for if we want to live together we need to be alike in thought. Vision is a big part of learning, brining out emotions in us much more than auditory or speech. I want to recreate computationally how we see, store the visual information in the memory and process it. It's a mix of computer vision / vision science, computational cognitive psychology, neuroscience, neurobiology and biophysics. Interlinking to create an AGI which behaves much like us. It shows aggression (the good violence, more on this soon), logic, empathy, co operation and dreams about the future.

6. Final Thoughts.

There's really a lot going on when you think about this. The top ones in my head are:

  1. Is it possible if the brain is computationally recreated, neurological diseases such as Alzheimer's, Parkinson's are discovered in AGI? Would we scapegoated vectors of the disease?
  2. Will this allow us to tune the AGI's emotion engine to understand more about Alzheimer's, Parkinson's, etc and develop a cure?
  3. AGI safety is a major concern in near future and would this be a solution?
  4. The brain relies on a nervous system to decide, feel and work. How would we implement a nervous system for an AGI?8
  5. Dynamical systems are heavily used to model the nervous system and its functions. 9 I want to find better tools to solve ODEs/PDEs, algorithms that don't task the machine. Nature uses chemistry to solve the equations. It's inspirational when you think about it.
  6. What if AGI have the power to fire neurons at their will?
  7. Yawning is contiguous,10 so is it possible to transfer data through this mechanisms? Tricking the brain (maybe mirror neurons? ) to do something else? Can this be used to hack AGI?

7. Headfakes.

Here's the head fake. Emotions don't mean anything for an AGI. An AI won't know that violence can be bad and good. Empathy and co operation can have selfish intentions.

How will the algorithms nurture these? How will we teach AI systems that your favorite sports team destroyed your not so favorite team, in a friendly match is not violence rather a figure of speech? It's not the same as terrorism.

A benevolent AI is likely to fail, get bullied and thrashed in it's life. Humans are hard, strong and pull through pain. Tolerance. Emotional or psychological pain is a very, very difficult to recreate computationally. Will it be possible?



Only time will tell.

I really want to pursue researching, writing and exploring all these topics to make the future we dream a better place not just for humans but also for AI. A quest for a noble cause, a sustainable future for us living with synths, AIs and possibly aliens? Beautiful, peaceful life, :)

I always ask this question, why do we need wars if we can from the start remove them? A communal world can be seeded right when we make AI systems to help, protect and enjoy the good days life has to offer. It has to start with the culture, with the thought. I want to hit the right places.

There are a lot of questions and heaps of ideas stored in me. It's the time to make these dreams beautiful, push science farther and probably, if everything goes right, the popular catalogues of the future have AGI and Humans together, as the front cover, interlinked by emotions and respect not by pure bred, synthetic bred, caste, religion or colour.

We, the future.

Raising a glass or two full of electric potential for the binary hearts of the future,

Ankit Panda.
14-Oct-2018, India.


Footnotes

  1. Kinder Joy in India, Kinder Surprise in other parts I guess, which though is much worse. Kinder Überraschung is a godly milk chocolate.↩︎

  2. 1Mbps connection was fast, during 2001-2002 ↩︎

  3. a popular search engine back then, RIP now. ( Google came soon enough ) ↩︎

  4. I never thought in the future these would be played again not by humans though but AI/Deep RL. ↩︎

  5. my mom hates math which proves love is stronger after all ↩︎

  6. did you know he was a Professor of biochemistry? ↩︎

  7. Gescheider G (1997). "Psychophysics: the fundamentals". Somatosensory & Motor Research (3rd ed.). 14 (3): 181–8. ↩︎

  8. please don't say 'the wires' in the robot :( ↩︎

  9. Wikipedia says, Hopfield networks were specifically designed such that their underlying dynamics could be described by the Lyapunov function which in turn is a nonlinear technique used to analyze the stability of the zero solutions of a system of differential equations. ↩︎

  10. a number of resources conclude this, look at mirror neurons, herd support, myth busters concluded this too. ↩︎