[VIDEO] Ben Goertzel: From Here to Human-Level AGI in 4 Simple Steps

talk
bengoertzel
agi
video
events

#1

Let us invite you to a following talk by Ben Goertzel entitled From Here to Human-Level AGI in 4 Simple Steps taking place on Monday May 21st, 9:30-10:30 a.m.

Venue: Czech Institute of Informatics, Robotics, and Cybernetics (Jugoslávských partyzánů 1580/3, Prague 6) – Red Lecture Room (B-246)

This event has passed, please find the recordings below

Video

Q&A

Abstract: AI technology has entered the mainstream of business and society, but there is still a large gap between the current crop of task-specific „narrow AI“ tools and the Artificial General Intelligences (AGIs) envisioned by futurists and SF authors. To get from here to true AGI will require advances in (at least) four different aspects. First, it will require coordination of different AI agents at various levels of specificity into an overall complex, adaptive AI network — which is the problem addressed by the SingularityNET blockchain-based AI framework. Second, it will require bridging of the algorithms used for low-level intelligence such as perception and movement (e.g. deep neural networks) with the algorithms used for high-level abstract reasoning (such as logic engines). Third, it will require embedding of AI systems in physical systems capable of interacting with the everyday human world in richly nuanced ways — such as the humanoid robots being developed at Hanson Robotics. Fourth, it will require the development of more sophisticated methods of guiding abstract reasoning algorithms based on history and context (an area lying at the intersection of AGI and automated theorem proving). Fortunately,while none of them are actually simple, all of these aspects of the AGI problem are topics of active research by outstanding teams around the world, making it plausible that AGI at the human level and beyond will be achieved during our lifetimes.

Dr. Ben Goertzel is one of the world’s foremost experts in Artificial General Intelligence, a subfield of AI oriented toward creating thinking machines with general cognitive capability at the human level and beyond. He also has decades of expertise applying AI to practical problems in areas ranging from natural language processing and data mining to robotics, video gaming, national security and bioinformatics. He has published nearly 20 scientific books and 140+ scientific research papers, and is the main architect and designer of the OpenCog system and associated design for human-level general intelligence.

Ben is the CEO of SingularityNET (a blockchain based AI platform company), and the Chief Scientist of Hanson Robotics, a robotics company that creates the world’s most advanced humanoid robots. Ben also serves as Chairman of the Artificial General Intelligence Society, which hosts the annual AGI research conference series, and the OpenCog Foundation.

Before relocating to Hong Kong in 2011, Dr. Goertzel held executive roles at AI consulting and product development firms in Washington DC (CEO, Chairman and Chief Scientist at Novamente LLC and Biomind LLC) and New York City (CTO at Webmind Inc.). Prior to that, he served as faculty in mathematics at the University of Nevada Las Vegas, in cognitive science as the University of Western Australia, and in computer science at Waikato University in New Zealand, at the City University of New York and at the University of New Mexico in Albuquerque. Dr. Goertzel holds a PhD degree in mathematics from Temple University in Philadelphia, USA.

More information: https://www.ciirc.cvut.cz/pozvanka-21-5-2018-ben-goertzel/


SingularityNET Monthly Updates #1
#2

Wow, thank you for info. I will be there!


#3

Please post a link if this will be video recorded. Thank you


#4

I am also very interested in video if it will be recorded.


#5

Hmm… I am not sure if I can get a recorded video of the event - but I will ask… maybe the institute will be recording it. If we have it I will upload it :blush:.


#6

Thank you again. It was really one of the most interesting presentation I have ever been. :slight_smile:
There was a cameraman, so i hope it should be a video from it.


#7

Awesome to hear!! If you have more pictures please share :smiley: I will hunt down the cameraman :stuck_out_tongue:


#8

If you’d like to share with others… what would be some of the most interesting things you took away from this presentation?


#9

Floppy hat or it didn’t happen.


#10

Gryffindor!!


#11

Here are a few thoughts that was the most interesting for me. :slight_smile:


It is important to make AI as close as possible to human brain. If we make AI just based for solving mathematical problems and it would not understand human thinking and this AI would became super intelligence machine it could be a bit complicated to find a way to understand each other. We can see it in nature for example with humans and monkeys. Our DNA is similar in 98 percents, but we are still quite different.

To make AI that will be close to us we need to understand first how our brain works.
A lot of AI developers focus on deep learning these days, which is interesting, but it is just one part what our brain can do. To make a complex Artificial general intelligence we need to learn all aspects of human brain.


I was asking a question about SingularityNET.

When there are a lot of algorithms working on SingularityNET, it would need a lot of computer power. Are you planning to use Golem or any system like that?

Answer:
We have tasted a Golem system, but it didn’t work the way we needed. There is a lot of potentially options like system like Amazon farms or a lot of mining farms in China that they are looking for a new opportunities. When there will be the demand, the infrastructure is already made and there will be way to solve that.


It is written by my words so I hope I understood everything correctly. :slight_smile:

I really recommend watch the video from it when it is hopefully out some day.
I was watching a lot of presentation from Ben and this one was one of the most interesting for me.


#12

I was hoping that Ben will wear the Zebra hat. :smiley: Unfortunately not… Next time I hope. :smiley:


#13

It was so interesting that I just made one photo. :frowning:
Next time I will make more pictures. :smiley:


#14

Thanks Patrik, interesting read. Two parts stood out for me, firstly Ben’s acknowledgment that the hardware, such as Amazon cloud, is being developed. Secondly that the process of finding our own brain function is the challenge. Thanks again.


#15

One thing though that I think is important to realise about DNA, it is true that we share 50% of our DNA with a Banana for example… but we definitely aren’t 50% similar to bananas.

So what is going on here?!

Well in fact only just over 1% of the genome codes for genes. The rest over 98% of our DNA is consider junk DNA, that is DNA that does not code for genes.

I think it’s fairly obvious that this so called junk DNA is anything but, and we should focus out attention on trying to understand what this information consists of and how it is expressed.

I think we will need near AGI levels of machine Intelligence to unravel that one.

However, I dont think it’s necessary to replicate how the human brain works to be able to reach and surpass human level intelligence.


#16

That’s a bit like putting bird feathers under the microscope to learn how to do engineered flight. We just need to know enough, or be sufficiently inspired to engineer a thinking machine. We are not driving around in ICE-based horses. We have plenty of bottom-up. We need more top-down (like Dr. Ben’s approach) and then we’ll have our first AGI. The goal is not to recreate or mimic human intelligence, but human-level or greater intelligence.


#17

Excellent comparison!

Your contributions recently have been utterly brilliant Brian!


#18

It is important to note that achieving AGI does not imply having reached a super-human level of intelligence or even a level of intelligence of the smartest human. The first AGI will likely be less than toddler-level. You have to crawl before you can walk. We will likely need super-human intelligence in order to determine why only one part of you came out shaped like a banana—and only in specific contexts—despite your fifty-percent shared genetic heritage with said tropical fruit.


#19

But this is still our best shot at creating AGI - if we look at how nature did it, we can try to replicate it. Otherwise, we are more or less throwing stuff at the wall and see what sticks. Sure, maybe we don’t need to understand how every single process works at a molecular level (though maybe we do need to - we just can’t be sure until we actually manage to develop AGI), but keep in mind that we are still a long way from understanding how all the important processes in the brain function, let alone those that we won’t find relevant for developing AGI. At the very least, fully understanding the human brain will speed up the path to AGI substantially, and at worst, it will be necessary to understand it. :slight_smile:


#20

The moment we get to a self aware AGI it will go from toddler to super human levels in a blink of an eye.