Menu
What's new
By:
Filters

What ideas interest you?

In general terms:

* People's interaction with technology

* Information security

* Recommendation systems

* Data aggregation & inference (Business intelligence I guess)
 
I'm into Cosmology and its physics. I also have a 10" Schmidt Cassegrain but wish my nuts didn't freeze off every time I use it in the winter...:confused:...
I also try to get a handle on all things quantum and string theory related, but without the math, I'm ultimately stymied. Even so, wow the wonder of it all.
 
Last edited:
in addition to improving my piano and compositional skills, I'm very interested in mathematics, relativity, and quantum mechanics. I'm also working a bit on various ML techniques for a few AI ideas.
 
AllanH said:
in addition to improving my piano and compositional skills, I'm very interested in mathematics, relativity, and quantum mechanics. I'm also working a bit on various ML techniques for a few AI ideas.
Click to expand...
What are your thoughts on the technological singularity?

Can you talk about your ML projects or would you rather not?
 
whiskers said:
What are your thoughts on the technological singularity?

Can you talk about your ML projects or would you rather not?
Click to expand...

Most teenagers already live the singularity - they cannot think and function without a smart phone!

Regarding Machine Learning: I can describe one project that was rewarding. Back in 1996 I started developing statistical models for trading stocks and futures. I quickly figured out that I needed to develop some custom statistics as stock/futures have time-varying distributions. As such, many of the common tools fail as the time series are non-stationary. I ended up developing a mechanism for time-evolving distributions and the mathematics to make decision with confidence (in a statistical sense). I have a PhD in mathematics, so I was pretty comfortable with that. In about 2003, I switched to real-time trading and the systems went fully autonomous in about 2005. I finally lost the speed battle some years ago and turned it off. I did not want to make the required investment to get down into the low micro seconds. In view of today ML algorithms, most of what I created would be considered ML but without any of now well-known frameworks.

EDIT: I coded the entire thing myself, including the ML algos and trading engine with broker interface. Pretty wild in retrospect.
 
AllanH said:
Most teenagers already live the singularity - they cannot think and function without a smart phone!

Regarding Machine Learning: I can describe one project that was rewarding. Back in 1996 I started developing statistical models for trading stocks and futures. I quickly figured out that I needed to develop some custom statistics as stock/futures have time-varying distributions. As such, many of the common tools fail as the time series are non-stationary. I ended up developing a mechanism for time-evolving distributions and the mathematics to make decision with confidence (in a statistical sense). I have a PhD in mathematics, so I was pretty comfortable with that. In about 2003, I switched to real-time trading and the systems went fully autonomous in about 2005. I finally lost the speed battle some years ago and turned it off. I did not want to make the required investment to get down into the low micro seconds. In view of today ML algorithms, most of what I created would be considered ML but without any of now well-known frameworks.

EDIT: I coded the entire thing myself, including the ML algos and trading engine with broker interface. Pretty wild in retrospect.
Click to expand...
wow, that's seriously impressive.

Re: singularity, I mean more of the 'runaway point' of AI being more or less self sufficient, and better at learning things (still is only good at very specific directions to my understanding.) But I suppose from a 'previously unfathomable technological changes' general aspect, yeah, we're already there :)
 
whiskers said:
wow, that's seriously impressive.

Re: singularity, I mean more of the 'runaway point' of AI being more or less self sufficient, and better at learning things (still is only good at very specific directions to my understanding.) But I suppose from a 'previously unfathomable technological changes' general aspect, yeah, we're already there :)
Click to expand...

That singularity! - By the time a true AI (assuming access to network resources) becomes self-aware, we'll be gone in 24 hours. Probably less. The numbers are simply stacked against us.

EDIT: let me substantiate my statement. The thinking processes in the brain are essentially chemical and operate in the kilo hertz range. Computers operate in gigaherz, i.e. 1,000,000 times faster. Over time we will likely have enough compute power to substantially equal a brain. So for each 24 hours the AI will have had 3,200 human years of experience. That's a few einsteins, wittens, etc. Unless the AI really likes us, we'll be rodents in no time. The moment the AI becomes self-aware, we'll immediately lose control of it. My two cents.
 
Craig Duke said:
Sam Harris also has a say about that.
Click to expand...
I read the Sam Harris book in which he opined in the affirmative about all of this nonsense. Predetermination, schmedetermination, sez I. That we have predetermined influences that affect us before we act? OMG-what an amazing theory...say it ain't so!
 
AllanH said:
Most teenagers already live the singularity - they cannot think and function without a smart phone!

Regarding Machine Learning: I can describe one project that was rewarding. Back in 1996 I started developing statistical models for trading stocks and futures. I quickly figured out that I needed to develop some custom statistics as stock/futures have time-varying distributions. As such, many of the common tools fail as the time series are non-stationary. I ended up developing a mechanism for time-evolving distributions and the mathematics to make decision with confidence (in a statistical sense). I have a PhD in mathematics, so I was pretty comfortable with that. In about 2003, I switched to real-time trading and the systems went fully autonomous in about 2005. I finally lost the speed battle some years ago and turned it off. I did not want to make the required investment to get down into the low micro seconds. In view of today ML algorithms, most of what I created would be considered ML but without any of now well-known frameworks.

EDIT: I coded the entire thing myself, including the ML algos and trading engine with broker interface. Pretty wild in retrospect.
Click to expand...

Did you happen to read Michael Lewis's book "Flash Boys"?
 
AllanH said:
That singularity! - By the time a true AI (assuming access to network resources) becomes self-aware, we'll be gone in 24 hours. Probably less. The numbers are simply stacked against us.

EDIT: let me substantiate my statement. The thinking processes in the brain are essentially chemical and operate in the kilo hertz range. Computers operate in gigaherz, i.e. 1,000,000 times faster. Over time we will likely have enough compute power to substantially equal a brain. So for each 24 hours the AI will have had 3,200 human years of experience. That's a few einsteins, wittens, etc. Unless the AI really likes us, we'll be rodents in no time. The moment the AI becomes self-aware, we'll immediately lose control of it. My two cents.
Click to expand...
On the optimistic side, we'll probably blow up the planet before that occurs. I like to remain positive.
 
NYC Composer said:
Did you happen to read Michael Lewis's book "Flash Boys"?
Click to expand...

I have not yet read Flash Boys. There are also many exchange order flow issues that are unfair to the little guy. As that became known, it explained several things as well. It's definitely a rigged game.
 
Last edited:
You must log in or register to reply here.
Top Bottom