saints1884 Posted August 4, 2009 Share Posted August 4, 2009 Although i had a stress free journey to work,last night i boarded a bus with a green advert all over it,i think it was for a power company.Anyway the long and the short of it was that i could hardly see a bloody thing out of the windows.It was like being driven around inside a large grape. Needless to say if that bus turns up again it can carry on without me. We have one of those here to. It's Scottish hydro electric. 0 Quote Link to comment Share on other sites More sharing options...
GypsyTillIDie Posted August 4, 2009 Share Posted August 4, 2009 Urgh, I think I've got the cold. 0 Quote Link to comment Share on other sites More sharing options...
renton Posted August 4, 2009 Share Posted August 4, 2009 Anyway, my own PTTGOYN.... My PHd has ground to a halt in the last couple of days. My project has required a good amount of mechanical construction work to make it go and I have sent away a block of material to get lapped down to provide micrometre accurate uniformity across it's surface. Without this workpiece I cannot carry out any experiments. It went away yesterday and given that the guy who does this work is so damn busy I don't expect to see him before friday. My theoretical work (which is building towards a different experiment) is being hamstrung becuase of the presence of variables that can only be solved empirically, so I can't build an accurate mathematical model, nor can I find any data on the subject. My boss is away for two weeks, and the other two guys who confirm signatures for buying equipment are away, so I can't even buy in the stuff I need for future work which I've been planning. It's fucking frustrating. 0 Quote Link to comment Share on other sites More sharing options...
vikingTON Posted August 4, 2009 Share Posted August 4, 2009 My gripe that puts the P in PTTGOMN: Being called 'wee man' by 5ft 6, 40 year old unemployed scumbags on the way to the shops. I'm 6ft 3 and 12 stone 5, I'm not by any definition 'wee' in comparison. -1 Quote Link to comment Share on other sites More sharing options...
xbl Posted August 4, 2009 Share Posted August 4, 2009 My boss is away for two weeks, and the other two guys who confirm signatures for buying equipment are away, so I can't even buy in the stuff I need for future work which I've been planning. It's fucking frustrating. If you're not up to anything then, know anything about 64 bit matlab, or multi core Matlab programming? I've also been hamstrung because my models have eaten all the available Matlab, and so I can only use half finished training sets rather than the full matrices I need. 0 Quote Link to comment Share on other sites More sharing options...
theentomologist Posted August 4, 2009 Share Posted August 4, 2009 it's a mild case, they have isolated the client in her home, filled her with Tamiflu and are monitoring her vitals.The clinical team want us to continue to offer support, but we are a mental health org, not a nursing one. My team leader's going 'bugger that for a game of smarties' and saying she should be admitted, but it's a hard one, as the clinical team are saying she is not at that stage yet. you could, but get dressed up as if its the andromeda strain/outbreak when you go to see her. I mean granted if she believes in aliens, is paranoid, or whatever it probably might not help. But you wont get the flu! 0 Quote Link to comment Share on other sites More sharing options...
renton Posted August 4, 2009 Share Posted August 4, 2009 If you're not up to anything then, know anything about 64 bit matlab, or multi core Matlab programming? I've also been hamstrung because my models have eaten all the available Matlab, and so I can only use half finished training sets rather than the full matrices I need. Sorry old chap, I'm afraid not, if you can't get access to a better machine, can you not reduce your training sets, maintain the functionality with less samples? 0 Quote Link to comment Share on other sites More sharing options...
xbl Posted August 4, 2009 Share Posted August 4, 2009 (edited) Sorry old chap, I'm afraid not, if you can't get access to a better machine, can you not reduce your training sets, maintain the functionality with less samples? I'm using much smaller training sets (6000 vectors rather than 60,000) but they arent' quite right. Also you can see significant improvement as you add more vectors, so to be academically thorough, I need to use a full training set. I am just currently using limited functionality, but its annoying! Edited to add, its not a hardware thing either. By switching to a better machine, I got my training time down from 4 days to an hour, but its a software thing. 32 bit Matlab is physically unable to cope. Edited August 4, 2009 by Vectron 0 Quote Link to comment Share on other sites More sharing options...
theentomologist Posted August 4, 2009 Share Posted August 4, 2009 (edited) If you're not up to anything then, know anything about 64 bit matlab, or multi core Matlab programming? I've also been hamstrung because my models have eaten all the available Matlab, and so I can only use half finished training sets rather than the full matrices I need. could you not just take Occam's razor to these unwieldy models that are using useful solitaire/minesweeper space? I guess what i'm saying is do the models need to be this size? perhaps you could go back and condense them down. If not, well then can you not borrow some matlab? edit: ah I see its a crappy software thing. excuse my uselessness. Edited August 4, 2009 by theentomologist 0 Quote Link to comment Share on other sites More sharing options...
Hank Scorpio Posted August 4, 2009 Share Posted August 4, 2009 We wait 9 years to play the Pars, and now we have them 5 times this season, twice in the space of 4 days. Fucking nonsense. 0 Quote Link to comment Share on other sites More sharing options...
xbl Posted August 4, 2009 Share Posted August 4, 2009 (edited) could you not just take Occam's razor to these unwieldy models that are using useful solitaire/minesweeper space?I guess what i'm saying is do the models need to be this size? perhaps you could go back and condense them down. If not, well then can you not borrow some matlab? Its a training set for a GMM. So I am using EM clustering to train an audiovisual model for predicting audio from lipreading. Sadly, I've already reduced the dimensionality of some stuff, and the rest, I need. My next plan is to learn how to parallel program and then get access to the departmental cluster, which gives me more power!! Power, by Vectron! How long you got left anyway Renton? Is it time to start thinking of a thesis outline yet? Edited August 4, 2009 by Vectron 0 Quote Link to comment Share on other sites More sharing options...
Toma_BullyWee Posted August 4, 2009 Share Posted August 4, 2009 Stop bloody moaning, you're still breathing, aren't you? beeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee eeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee eeeeeeeeeep. I only kid, I am still breathing. I had some very interesting dreams last night.... still bloody sore though. 0 Quote Link to comment Share on other sites More sharing options...
renton Posted August 4, 2009 Share Posted August 4, 2009 (edited) I'm using much smaller training sets (6000 vectors rather than 60,000) but they arent' quite right. Also you can see significant improvement as you add more vectors, so to be academically thorough, I need to use a full training set. I am just currently using limited functionality, but its annoying!Edited to add, its not a hardware thing either. By switching to a better machine, I got my training time down from 4 days to an hour, but its a software thing. 32 bit Matlab is physically unable to cope. Grim times. At the moment I'm struggling with ironing out reliability issues with my setup, having done that I have a nine run experiment to conduct, and then i should have enough for a good journal paper, based on real time sensing of structure growth in air/polymer interfaces. I've developed the maths that predicts both the electorstatic force and the thermal gradient which are the primary dirvers of polymer growth, but I really want to try introducing an acoustic wave into the interface, an acoustic casimir effect. The maths however is a bitch ,the materials I've found requiring the use of hamaker constants, and I can't get a practical experiment set up to try it at the moment either..... Edited August 4, 2009 by renton 0 Quote Link to comment Share on other sites More sharing options...
Guest The Phoenix Posted August 4, 2009 Share Posted August 4, 2009 beeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee e eeeeeeeeeep. I only kid, I am still breathing. I had some very interesting dreams last night.... still bloody sore though. You need to get out and listen to a decent band. If you are still with us, you up for The Box on the 14th? 0 Quote Link to comment Share on other sites More sharing options...
xbl Posted August 4, 2009 Share Posted August 4, 2009 ...and then i should have enough for a good journal paper.... Always a good thing! I've got a system in place, and all the components seem to work, but when you actually run it, it fails because of the crappyness of the models. Like you, once I've got that, and finished the integration with the audio only approach, then I should also have enough to publish again. By Vectron, do a PhD folks, its a barrel of laughs! 0 Quote Link to comment Share on other sites More sharing options...
renton Posted August 4, 2009 Share Posted August 4, 2009 (edited) Always a good thing! I've got a system in place, and all the components seem to work, but when you actually run it, it fails because of the crappyness of the models. Like you, once I've got that, and finished the integration with the audio only approach, then I should also have enough to publish again. By Vectron, do a PhD folks, its a barrel of laughs! I've been second author on a few group papers but this would be my first solo effort....... I spent the first 8 months (at my bosses insistence) trying to solve a 150 year maths problem that was tangential to my project at best. That was fun. It's taken me to this point to assemble a system that can do what I need it to do, I just need to iron out the bugs. Little reliability issues like a rough surface with bumps in the 100 micron region (which is an order of magnitude bigger than what I'm attempting to measure!) or a hairline crack in a nickel mask disrupting the polymer surface! If I sort it, I'll be able to run any number of experiments quickly and should rack up resutls bloody fast. The frustration of relying on others or on bad equipment, the inability to escape to another project, leaving you stuck in no man's land, is demoralising to say the least. Edited August 4, 2009 by renton 0 Quote Link to comment Share on other sites More sharing options...
xbl Posted August 4, 2009 Share Posted August 4, 2009 I've been second author on a few group papers but this would be my first solo effort....... I spent the first 8 months (at my bosses insistence) trying to solve a 150 year maths problem that was tangential to my project at best. That was fun. When you say solo, you mean entirely on your own, or do you mean with your supervisors name on as well (although his contribution will be minimal at best!). I've been first author on 2 papers, and second author on one so far. I am limited as I am not part of a group, and so in the name of Vectron, there is nobody else I can turn to get my name added spuriously. 0 Quote Link to comment Share on other sites More sharing options...
Guest The Phoenix Posted August 4, 2009 Share Posted August 4, 2009 The frustration of relying on others or on bad equipment, the inability to escape to another project, leaving you stuck in no man's land, is demoralising to say the least. Have a w**k. HTH 0 Quote Link to comment Share on other sites More sharing options...
Ayrgirl Posted August 4, 2009 Share Posted August 4, 2009 True, but we're not docs or general nurses. Not real nurses then 0 Quote Link to comment Share on other sites More sharing options...
Toma_BullyWee Posted August 4, 2009 Share Posted August 4, 2009 You need to get out and listen to a decent band.If you are still with us, you up for The Box on the 14th? Listen to to a decent band, pah! I'll be playing in a shite hot one tonight. Normally I would be up for box, but I'm playing in Clydebank that night, so I won't be able to go. One day though! 0 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.