December 16, 2015 at 9:48 pm
Comments posted to this topic are about the item The Machines are Learning
December 17, 2015 at 12:21 am
The potential has been there so long that when it started to appear viable I just could not believe it. It is a bit like the cold fusion of the computing world.
Gaz
-- Stop your grinnin' and drop your linen...they're everywhere!!!
December 17, 2015 at 7:12 am
I'm not so sure about the article's assertion that the credit card companies are suffering losses. In my experience, it is the business that accepted the fraudulent card that suffers the loss because the credit card company does not pass the money to them.
December 17, 2015 at 7:30 am
Back in the late 1980's, I worked for a retailer that would require the customer to show their drivers license whenever they presented a credit card or check. If the customer refused, then we would call the manager who would either approve the transaction or politely deny it. We even had photos of known grifters from the local area behind the counter. In the following decades ecommerce has removed that face to face aspect of the transactional process, making identity theft easier and more common, and we're still playing catch-up in terms of security.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
December 17, 2015 at 8:25 am
kmpa (12/17/2015)
I'm not so sure about the article's assertion that the credit card companies are suffering losses. In my experience, it is the business that accepted the fraudulent card that suffers the loss because the credit card company does not pass the money to them.
It depends. Sometimes its the vendor, sometimes the CC company.
December 17, 2015 at 11:06 am
kmpa (12/17/2015)
I'm not so sure about the article's assertion that the credit card companies are suffering losses. In my experience, it is the business that accepted the fraudulent card that suffers the loss because the credit card company does not pass the money to them.
I thought the same thing but I guess sometimes its the credit card company or bank. I don't know what conditions are involved.
December 17, 2015 at 12:38 pm
I was shopping at Wal Mart a few months back. The machine was messing up - 3 customers in a row. While I was in line, my I got a call from a number I didn't know. Because I was dealing with someone face-to-face, I ignored it. When I was done, I listened to the voice mail and it was fraud department informing me that there were several hits on my card in very rapid succession, so they locked it. I was impressed.
The overall capacity of machines to learn is cool. It's also a bit scary because we don't know that the conclusions they reach will always be logical or correct. Humans certainly aren't always logical or correct either, but there is still some common sense left in the world, no matter how badly sensitized or diluted.
December 17, 2015 at 3:13 pm
The trick is in training the system in the first place. I remember playing with a package called 4Thought back in the 1990s. I think you have to be careful with expectations. Yes these systems do increase the hit rate but there is a lot of human skill in getting the high hit rates
December 17, 2015 at 6:17 pm
How applicable is machine learning for most of us?
I would say that machine learning has it's usefulness. However, it's not easy to make work. If so, we could predict the stock market. We could product a sale. We could even predict how likely any one of us is going to commit a crime in the next 15 minutes.
Unfortunately, like the whole big data craze, machine learning is up there. It can be your friend or your enemy. It can help you land a lot of money or sink your business with a poor decision. Clearly it can work. We use it for fraud and to predict the weather. But, it takes time, research and a lot of smart people.
Statistics on the other hand, which is about drawing valid conclusions, is a lot more useful. Database professionals can do a lot more with a total of sales per region than a prediction of total sales in the next 2 years.
That's just me though.
December 18, 2015 at 5:27 am
xsevensinzx (12/17/2015)
How applicable is machine learning for most of us?
I would say that machine learning has it's usefulness. However, it's not easy to make work. If so, we could predict the stock market. We could product a sale. We could even predict how likely any one of us is going to commit a crime in the next 15 minutes.
Unfortunately, like the whole big data craze, machine learning is up there. It can be your friend or your enemy. It can help you land a lot of money or sink your business with a poor decision. Clearly it can work. We use it for fraud and to predict the weather. But, it takes time, research and a lot of smart people.
Statistics on the other hand, which is about drawing valid conclusions, is a lot more useful. Database professionals can do a lot more with a total of sales per region than a prediction of total sales in the next 2 years.
That's just me though.
You raise a good point. We can predict events based on data, but not on natural or random things.
The stock market is driven by people, specifically their brains and emotions. The weather is driven by what occurs in nature.
Scientists are still trying to understand these things. Granted, we know a lot, but certainly not everything.
Therefore, any conclusions or predictions we reach are based on a partial understanding. It's no wonder they're wrong.
December 18, 2015 at 9:37 am
Ed Wagner (12/17/2015)
I was shopping at Wal Mart a few months back. The machine was messing up - 3 customers in a row. While I was in line, my I got a call from a number I didn't know. Because I was dealing with someone face-to-face, I ignored it. When I was done, I listened to the voice mail and it was fraud department informing me that there were several hits on my card in very rapid succession, so they locked it. I was impressed.The overall capacity of machines to learn is cool. It's also a bit scary because we don't know that the conclusions they reach will always be logical or correct. Humans certainly aren't always logical or correct either, but there is still some common sense left in the world, no matter how badly sensitized or diluted.
It is kind of amazing. After my last trip, where I was in Orlando, UK, Houston, I got a call the day I arrived home. They'd flagged fraudulent purchases in Ohio that day. Despite the fact I'd used my card in the UK and Houston that day, those weren't seen as fraud.
That's impressive to me.
Viewing 11 posts - 1 through 10 (of 10 total)
You must be logged in to reply to this topic. Login to reply