In theory yes you could. RL is based on trial and error. The agent senses the environment in the form of states (ligh=state1, dark=state2) and it performs actions according to a policy to receive positive or negative rewards. For example your goal is to have a room always lighten, the agent will sense the environment (it is dark) it will turn the lights on and it will receive a positive reward.
Eventually the agent will learn that in order to receive positive rewards it needs to turn on the light when it is dark and turn it off when it is light.
Actually RL it is more complex than that but that is the basic idea.
If you are interested go to scholar.google.com and search for papers about reinforcement learning. Important authors are Barto, Sutton, Watkins (Q-Learning), Littman, Kudenko, Stone, Clauss, etc.
For me it is much simpler, Apple became much greedier that it was. They saw the opportunity to monopolize the market and they took it without any remorse.
Google was important but I agree that the first search engine should have been there.
IPv6 was also missed there, still not fully operational but it has been there for a while.
Internet worms (i.e. Code Red) were also important, they raised some alarms related to security practices in OSs and applications.
I agree with some of the statements of the blog. However it misses one important point in how people use lists. For example I follow around 300 people, some of them are very noisy some times but eventually they post something interesting, so I feel that the noisy is an acceptable price to pay.
Nevertheless, because of the overflow of information, many times I miss posts of quiet people that generally have something important to say. Then, lists come to the rescue.
I just put those very interesting people in a list that I can check for time to time. If people follow my list(s), they give the authority that I created it(them)or whatever I do not really care. For me that is just marketing trying to convince of the value of twitter. For me, the value is finding good content with less effort; and I think the lists are helping on it.
Note: I use also tweetdeck that has a similar concept, which I found very useful. Perhaps that is why I liked the lists concept.
But as a entrepreneur, Do you really know that data? What are your sources? What are your tools to make the forecast?
Yes, generally you do not have any of them and they know that. Your forecast is just a very empirical calculation with a large margin of error.
I think there are different monsters. One thing is to project what are you going to do in the next five years and another completely different is to forecast what are you going to "sell". I think the post implies that the later is crap. For the former I imagine that all of us know that things are going to change (as I think you imply too about a PhD program).
There are many applications that do not require unique algorithms and can be solved using simple regressions. Also not everybody has access to a ML expert to code a solution (even if this only requires a simple ANN using an opensource library).
Machine learning as a service, very interesting. I found a similar site some months ago. I looked over my bookmarks but I could not find it. The idea was similar, you upload a data-set and they execute different learning algorithms over it. I do not recall if you also provide an evaluation or they automatically split the given data-set.
Not necessary. If you assume that people is going to spread the news in twitter (or blogs, facebook, etc.) when they find a balloon, the locations is not a restriction. You just need a smart enough engine to search and find those tweets. Now, the "smart engine technology" is the difficult task (and for sure what DARPA is looking for).
I think ties are for big consolidated enterprises that are slow and difficult to move (or at least that is my prejudge). Yahoo! does not need ties, they need to move fast and get in top of Google, MS, Facebook and Twitter. They have to re-invent themselves if they want to survive.
Look at his eyes and the expression in his face. He is not comfortable and he does not believe what he is saying. He WANTS to believe it though. Or perhaps I've just watched "Lie to Me" too much.
We have been working remotely for a while. It is difficult but it can be done. I would like to add a couple of applications that we have used to make our life easier:
- Etherpad.com for real-time document colaboration
- Google documents, if privacy is not a concern
Also, because it is very difficult for us to get together, every time that we can do it we try to maximize the time (no dumb and useless meetings, going straight to the important points, etc.). As result, our face-to-face meetings are very productive.
Adding some randomness could help to avoid over-fitted populations. Then you could prevent to converge to a sub-optimal state and force the system to look beyond of what it seems the best.
In the old times of WAP, mobile operators had gateways, so even though you could get the IP address of the phone (from a server-side for example) the information was meaningless to locate the user.
With AT&T 3G I honestly do not know. May be someone could do some basic hacking and post some results.