Tag AI

Posts: 20

How to Recognize AI Snakeoil

My colleague posted an interesting Princeton presentation to our group Slack channel. It's about how AI is being used in social interactions such as hiring decisions. I completely get why HR departments and companies want to do this. They get 100's of resumes from people and want to weed through the lot of them to get to the right candidates. This is how AI can help, optimizing processes to assist people to make the right decisions.

For example, AlphaGo is a remarkable intellectual accomplishment that deserves to be celebrated. Ten years ago, most experts would not have thought it possible. But it has nothing in common with a tool that claims to predict job performance. via Princeton

AI can scan an applicant's resume for keywords and rank job hunters based on the universities they attend, but it can't discern the human factor. You can be socially awkward but excel in your field. You can be great on a paper and have all the right credentials but be a complete asshole. AI can't predict the subtleness of being a human.

Princeton AI Snakeoil Presentation

As AI is being used in these social interactions more and more, people are starting to game the system. Applicants are scrutinizing job applications and hiding in white text words like "Cambridge" or "Yale." Humans can't read it and it's not there when you print it out, but a computer can read it in digital form. This skews the results, in your favor.

I already showed you tools that claim to predict job suitability. Similarly, bail decisions are being made based on an algorithmic prediction of recidivism. People are being turned away at the border based on an algorithm that analyzed their social media posts and predicted a terrorist risk

Another big problem I see is how AI is being used to assist in judgment. These have big consequences if you happen to share similar 'features' of a terrorist or criminal. This goes to the notion that fairness is hard and that human lives are messy. Our Judicial system uses these systems to 'predict' whether to deny someone bail. There have been cases where these systems aren't fair at all. Gender bias is another big thing. The recent high profile Apple Card credit limits highlights potential gender bias.

What can we do in this brave new world to fight these problems? The first step is awareness that there is a problem. The next step then is to use the same technology to solve the problem. Then track the system to make sure it's not gamed and fair.


Go Master Quits, AI too powerful

I just read that Lee Sedol is retiring from the competitive Go world because of AlphaGo.

The South Korean said he had decided to retire after realizing: "I'm not at the top even if I become the number one."

This is a bit sad but understandable. I wrote about Lee's historic match with AlphaGo a while ago. He ended up losing four matches out of five, which was an ass-kicking for the #1 Go player in the world. Just think about it, AlphaGo from Deepmind is just getting better and better.

What is Go?

If this is the first time you're reading about Go, then you probably don't understand the importance of this particular AI beating a world champion. Go (aka Weiqi) is an ancient Chinese game that is about getting territory on a board made of 361 intersections.

You place these black and white stones at the intersections and then build 'territory.' The person who has the most territory after all moves are complete (or resigns), is the winner.

Wikipedia defines it:

Go is an abstract strategy board game for two players, in which the aim is to surround more territory than the opponent. The game was invented in China more than 2,500 years ago and is believed to be the oldest board game continuously played to the present day. A 2016 survey by the International Go Federation's 75 member nations found that there are over 46 million people worldwide who know how to play Go and over 20 million current players, the majority of whom live in East Asia.

It's been said that the combinations of Go moves is greater than the amount of atoms in the universe. I'm not sure if that's correct but the lower estimate of moves is somewhere around 2 × 10^170. So you can understand why AlphaGo beating a human is so powerful.

I've played it a bit of Go over the years and got to about 14 Kyu level. I don't play competitively but I enjoyed a few friendly games and have played against a few Dan level players. They expertly wiped the board with me with such grace and humility. It was a thing of beauty to see how they lead me into traps 20 and 30 moves ahead.

Why you should play Go

I recommend that every parent expose their kids to both Chess and Go, these type of games are awesome for strategic thinking and teaching patience. I believe it helps wire their brains in ways that become apparent years later.

I don't get much time to play Go now, but perhaps over upcoming holidays I might play a game online or with my children. If you wanted to try it out, go check out the KGS Server or buy a simple board from Amazon. Let me know how you like it.


AI Trading & Lehman Brothers?

WOW! I posted this over 12 years ago. Talk about being spot on the money! I hate to do this but I'm going to quote myself here:

I firmly believe that data mining, AI, and machine learning trading will accelerate over the years. Who knows, maybe my little model will move markets one day! :)

As far as the trend in AI goes, it's really getting started now. AI is being adopted by the next tier of companies. We're out of the "innovator" and now "early adopter" phases. We're starting to move into the "early majority" phase, and this is where it gets good.

Adoption Phases Adoption Phase Cycle

I'll write more on this late but I want to go on a tangent here. I'm cracking up. LEH-FUCKING-MAN BROTHERS. WOW. It's amazing what I can find in my archives back then. Remember Lehman Brothers? They blew the fuck up because of bad bets in the MBS markets.

They used to be so 'bad ass' but then they kept too much risk on their balance sheet and the Fed refused to bail them out, they set off the financial crisis. I posted a link to an article a few days later where Richard (Dick) Fuld got punched in the face at the gym. A 150 year company that was destroyed in a matter of days because everyone was greedy.

LEH filed for bankruptcy on September 15, 2008, a little over a year later from my AI post. It's mind boogling, it had $639 Billion in assets with $619 Billion in debt. The charts are damn right frightening but the good news is that AI is still here.

LEH, Daily Chart to collapse Lehman Bros. (LEH) Daily Chart prior to collapse

LEH, Daily Chart prior to collapse Lehman Bros. (LEH) Weekly Chart prior to collapse


Investing in the S&P500 beats AI Trading

I forgot to link to this image back in early May 2019. It's from Bloomberg and it makes a lot of sense, maybe eye opening for some.

Investing in the S&P500 beats AI Trading

Essentially it shows that the market is the be and end all. Of course the market is an average and there are funds or AI traders that beat the average, but others will underperform the average. Over time, your little 'edge' will eventually underperform and you have to keep writing new code and new strategies. You can write all the trading algorithms you want but in the end it'll be futile.

If you started using the AI back in 2016, you would've outperformed the S&P500 for a short while but then...

Save your time and just do passive investing. Buy high quality funds and ETF's with low expense ratios over a diverse asset class. That's it. Then rebalance once or twice a year.

Yeah, but Tom you post Forex and Stock stuff from time to time. Yes, I do and I have some 'mad money' that I trade to satisfy that itch. Do I trend follow with that mad money, yes. Do I go long or short currencies, yes. Have I started writing Python scripts to crunch market data and build charts? Yes, I do that so I can learn Python better and it's what I do to keep myself in the market Zeitgeist.


Neural Market Trends is the online home of Thomas Ott.