PhD s from Convolution Integral Thread
in [DSP]
|
Prev: Inverting bits
Next: ifft after modifying a signal
From: HardySpicer on 4 Jan 2010 05:02 On Jan 4, 9:39 pm, Rune Allnor <all...(a)tele.ntnu.no> wrote: > If a problem has not been solved yet, there are several > possible reasons: > > 1) Irrelevance. No one cares if the problem is solved or not. > 2) Feasability. The problem is relevant, but no solution can > be found. > 3) No one have yet discovered that off-beat approach > exploiting those half-mad properties of that arcane > theorem yet. > No no no! This reminds me if the famous quote (actually a missquote) that everything of use has already been invented. We don't know what is yet to be discovered and if the sum toal of all scientific knowledge is 100%, I am willing to bet that we know about 10% tops. What about quantum computers, biological computers with almost human AI, robotics to a level that is nearly human,Fusion power,human- computer interfacing, space exploration, alien life-forms, grand unified theory of forces, and so on... Even in humble dsp there are so many things to be discovered and being discovered that would never have been dreamed of 30 years ago. For example, real-time independent component analysis - can anybody make it work? Can you separate out multiple convolutive mixtures in real-time? Accurate speech recognition in a noisy environment - is it possible? Easy to give up and say these things are of no use, practicing engineers frequently do since they cannot make money out of them at present. Hardy
From: steveu on 4 Jan 2010 05:26 >On Jan 3, 9:43=A0am, Fred Marshall <fmarshallx(a)remove_the_xacm.org> >wrote: >> It got too long and I just couldn't spend the time to read all the >> posts. =A0So, if this is redundant then .. OK: >> >> 1) Folks with common sense and experience are invaluable. >> 2) Folks with an ability to focus on the real issue are invaluable. >> 3) Folks who can figure something out quickly and put what they know to >> use are invaluable. >> >> So, if in the process of earning a degree a person becomes a "quick >> study" (thus the search / research part) and are able to convey the crux >> of what they know or have quickly learned to everyone else (thus the >> thesis and defense) then they might be really good to have around - as >> long as they meet (1) and (2) at the same time. >> >> I have known folks who weren't really great at (1) and (2) or even (3) >> who had advanced degrees. =A0They were still helpful when we needed a >> strong technical basis for pondering a thorny issue. =A0But, in general, = >I >> didn't find these latter guys in industry - more in academia or >> government labs. =A0There's nothing wrong with this, it's partly an >> environmental emphasis. >> >> What's the difference between a plant manager and a consultant? >> >> If something goes wrong in the plant, the manager may know what to do >> about it but maybe not why it happened or why the solution works. =A0The >> consultant may not know what to do about it but may know why the >> solution works. =A0Working together they may come up with a lasting >> solution.... >> >> Fred > >What about problems that have not been solved to date? Can your >average Joe engineer solve them or does it require some advanced >knowledge/Maths that Joe may not follow. You seem to be equating "average Joe" with "someone who doesn't have a PhD". A PhD is purely a career choice. It is unrelated to ability or knowledge. The average Joe engineer typically doesn't do anything really deep, whether or not he has a PhD. Smart people, with and without PhDs, do. Between completing an MSc and a PhD a person has typically only worked within a very narrow area. Unless some specific maths was needed for that topic, they probably won't be any more mathematically capable at the end of the PhD. > eg pre-Widrow and co there >was no LMS,adaptive filters etc. Would an engineer in industry have >had time to sit down and work all that out on his own or in a group? >Doubt it. Engineers in industry had worked out quite a lot of adaptive techniques before Widrow. He did interesting work, and then others did interesting work after him. Being an academic or not is coincidental. >Maybe in a research lab environment for IBM or Bell labs but >most companies are interested in one thing only - making money! That's >why they are in business. Scholarly activity for its own sake is the >domain of the University. You have a very high minded view of University life, that most people don't share. Scholarly activity is very much tailored to sustaining funding. Nothing more. On the other hand Bell labs produced much of the best academic work of the 20th Century, often as a spinoff. People from Bell labs won 7 Nobel prizes, and those were mostly not for focused industrial work. I remember an interview with the head of Bell labs in the 70s, when people from Bell labs has won a prize for an outcome of their radio astronomy work. He said funding motivated people to do work like that has proven highly profitable in the long run. He quoted some interesting examples, too. Even when the likes of Bell and IBM produced commercially interesting work, it was often been outside their sphere - e.g. a lot of the core techniques used in digital comms came from IBM labs, although IBM has never been a major supplier of digital comms equipment (e.g. the work of Godard and Ungerboeck) - so, commercial labs are quite able to be quite broad in their coverage. We are living in a rather sad time where both commercial and academic research are narrowly focused on short term goals, and quick financial exploitation, rather than deeper research with potentially greater but slower returns. Its not an academia vs industry issue at all. >Then again occasionally a company does >discover something new - happens a lot in Japan. The knowledge will >disseminate eventually but it's a secret for a long time - or it's >patented so nowbody else can use it. I spent a long time working in a >company with a very experienced engineer who told me every day how >rubbish academics were and that they knew virtually nothing of any >real use that wasn't in a text book. I eventually managed to squeeze >out of him some great stuff indeed but it was a struggle. He was tight >fisted when it came to knowledge. he considered it all bellonged to >him since he had discovered it! It was his property and not to leave >the company. Contrast this with a Uni. >Yes companies are great but they exist but for one reason - to make >money. Now remember that next time you discover something new. Haven't you noticed that industry and academia are now essentially the same in how they work? Take a patent minefield, like speech compression. CELP was a pretty smart idea. It was developed in a commercial environment, yet the basic idea was not patented. Later, Sherbrooke U patented every little idea surrounding the core idea of CELP, that anyone in the department could come up with. You seem to extrapolate the whole world from experience with one guy. What happened to your precious scientific method? There are plenty of people in both industry and academia who think they are so key that only they can take the next step. Many of them think patents should essentially last forever. They kinda ignore the reality that anything smart they've done is just the top layer on a huge pile of knowledge. Without free access to all the lower layers, their patentable ideas would be in gridlock. This is a personality issue, and has nothing to do with industry or academia. Steve
From: Rune Allnor on 4 Jan 2010 06:06 On 4 Jan, 11:02, HardySpicer <gyansor...(a)gmail.com> wrote: > On Jan 4, 9:39 pm, Rune Allnor <all...(a)tele.ntnu.no> wrote: > > > If a problem has not been solved yet, there are several > > possible reasons: > > > 1) Irrelevance. No one cares if the problem is solved or not. > > 2) Feasability. The problem is relevant, but no solution can > > be found. > > 3) No one have yet discovered that off-beat approach > > exploiting those half-mad properties of that arcane > > theorem yet. > > No no no! This reminds me if the famous quote (actually a missquote) > that everything of use has already been invented. No, I don't. I am saying that no one ever have said "I think I will invent something today" and then actually done it. Any non-trivial discovery or invention is the result of a long-term activity paired up with a not at all insignificant portion of coincidence and luck. There is the anecdote about Edison searching for a suitable filament for his light bulb. The assistant came in and said, "this was the 2000th (or whatever it was) test that failed" upon which Edison supposedly replied "the we know 2000 materails that don't work." In those days the test were all in a day's work. These days that would have been 2000 PhD theses. Rune
From: invalid on 4 Jan 2010 06:37 "Rune Allnor" <allnor(a)tele.ntnu.no> wrote in message news:32ea28e3-d644-4035-9f4c-e7488f994843(a)k23g2000yqa.googlegroups.com... > There is the anecdote about Edison searching for a suitable > filament for his light bulb. The assistant came in and said, > "this was the 2000th (or whatever it was) test that failed" > upon which Edison supposedly replied "the we know 2000 materails > that don't work." > > In those days the test were all in a day's work. These days > that would have been 2000 PhD theses. Which brings us to the correct and reverential way to address a recent PhD, which is .... "A Big Mac with Fries, please"
From: steveu on 4 Jan 2010 07:48
>On 4 Jan, 11:02, HardySpicer <gyansor...(a)gmail.com> wrote: >> On Jan 4, 9:39=A0pm, Rune Allnor <all...(a)tele.ntnu.no> wrote: >> >> > If a problem has not been solved yet, there are several >> > possible reasons: >> >> > 1) Irrelevance. No one cares if the problem is solved or not. >> > 2) Feasability. The problem is relevant, but no solution can >> > =A0 =A0be found. >> > 3) No one have yet discovered that off-beat approach >> > =A0 =A0exploiting those half-mad properties of that arcane >> > =A0 =A0theorem yet. >> >> No no no! This reminds me if the famous quote (actually a missquote) >> that everything of use has already been invented. > >No, I don't. I am saying that no one ever have said "I think >I will invent something today" and then actually done it. > >Any non-trivial discovery or invention is the result of a >long-term activity paired up with a not at all insignificant >portion of coincidence and luck. > >There is the anecdote about Edison searching for a suitable >filament for his light bulb. The assistant came in and said, >"this was the 2000th (or whatever it was) test that failed" >upon which Edison supposedly replied "the we know 2000 materails >that don't work." > >In those days the test were all in a day's work. These days >that would have been 2000 PhD theses. > >Rune > |