I met him once, he came for a conference at my university. During lunch he sat next to him. What had struck me was how kind, humble and considerate he was. Sometimes people of his caliber have the "I am too important or to busy to talk to some no name student like you" attitude but he was the opposite of that.
Lotfi Zadeh has died
21–30 of 34 posts
Re: Lotfi Zadeh has died
#22Darnit. Fuzzy sets and logic just made sense to me in so many ways and I found his writings quite interesting. RIP.
Would you mind concisely explaining one example and application of it? I've read (for example) about possibility theory, but I hardly understand what it helps us do that probability theory didn't.
Other than some of the stuff on the Wikipedia article and the double inverted pendulum, I used it with a weird heating problem and some social stuff.
So, there is this material that needs to be heated on a curve (time 0 = room temp, 20 minutes = 400C, 120 minutes = room temp - cannot remember after 20 years - it cooled sorta quickly because of fans when heater off). Software package with oven basically turned heaters full blast, the off, then full blast, etc. about half the time material wasn’t cooked properly. I wrote a bunch of rules that computed a % the heater should be at. Worked really well and was an amazingly small program.
Social wise I wrote rules to judge success based on about 20 factors. Never bothered to pursue it beyond my own satisfaction.
Re: Lotfi Zadeh has died
#23Earlier quoted context omitted.
How is this different from increasing the temperature until p(cold) = p(hot)?
Since they belong to the same probability distribution, the states are not independently defined. A change that increases the probability of hot MUST decrease the probability of cold, since the probability of all states must sum to 100%. For example, you can't have a temperature that is 80% hot AND 80% cold. This problem becomes even more apparent as you increase the number of states (lukewarm, warmish, very warm), a…
"you can't have a temperature that is 80% hot AND 80% cold. This problem becomes even more apparent as you increase the number of states"
I don't see why this is an issue. You can also define each of the states as its own binary random variable and then the probabilities of two states conditioned on the temperature can add up to more than 1.
I thought your original post was meant to explain a practical application of fuzzy theory and how it differs from probability theory. Perhaps there is another example that better illustrates how fuzzy theory simplifies a problem where using probability theory would be messy / impossible?
Re: Lotfi Zadeh has died
#24One of my favourite subjects during college. I recommend the book by Timothy Ross - Fuzzy Logic with Engineering Applications for anyone interested.
Re: Lotfi Zadeh has died
#25I'm sure that there will be a renissance of fuzzy logic, its sad to know that he will not witness it.
His character will be missed, but his contributions won't be forgotten.
Re: Lotfi Zadeh has died
#26Most unfortunate. I did have the pleasure of seeing him when he came to Villanova University to accept an award from the Franklin Institute.
Re: Lotfi Zadeh has died
#27Earlier quoted context omitted.
Since they belong to the same probability distribution, the states are not independently defined. A change that increases the probability of hot MUST decrease the probability of cold, since the probability of all states must sum to 100%. For example, you can't have a temperature that is 80% hot AND 80% cold. This problem becomes even more apparent as you increase the number of states (lukewarm, warmish, very warm), a…
"you can't have a temperature that is 80% hot AND 80% cold. This problem becomes even more apparent as you increase the number of states" I don't see why this is an issue. You can also define each of the states as its own binary random variable and then the probabilities of two states conditioned on the temperature can add up to more than 1. I thought your original post was meant to explain a practical application of…
> You can also define each of the states as its own binary random variable and then the probabilities of two states conditioned on the temperature can add up to more than 1.
Yes, but then you're defining a probability distribution over the space of fuzzy states.
Re: Lotfi Zadeh has died
#28I'm sure that there will be a renissance of fuzzy logic, its sad to know that he will not witness it. His character will be missed, but his contributions won't be forgotten.
If you strip away some cosmetic differences, most deep learning is basically fuzzy logic, so he did witness it.
Re: Lotfi Zadeh has died
#29A kind, patient mentor when I needed much kindness and patience.
Dear God, who is going to clean out his office?
Re: Lotfi Zadeh has died
#30I'm sure that there will be a renissance of fuzzy logic, its sad to know that he will not witness it. His character will be missed, but his contributions won't be forgotten.
If you strip away some cosmetic differences, most deep learning is basically fuzzy logic , so he did witness it.
Out of curiosity can you elaborate on this?