TonnyM
Badges: 7
Rep:
?
#1
Report Thread starter 9 months ago
#1
Hi,

I have a long standing interest in AI, Cognitive Science, Neuroscience and Machine Learning. I am 28, CNC Operator/Engineer without a degree.

My options are part-time degree in Mathematics & Statistics or Engineering (Electrical or Design/Modeling).

I am not scared of math. I have no problem to follow MIT classes on YouTube. I can most likely teach myself all the necessary mathematics but I still need a degree in something relevant to get a decent job. Would not be a degree in Maths a waste of time and money?
On the other side degree in EE lead bit off the track but at least it is a proper qualification on it's own.
0
reply
Acsel
Badges: 20
Rep:
?
#2
Report 9 months ago
#2
While Maths is useful, to work with AI you're most likely going to need programming knowledge and the only degree that's really going to teach that properly is Computer Science. Obviously learning programming is possible outside a degree but one way or another it's not something you can totally ignore. Typically speaking, people don't just go straight into AI. It's the sort of thing people specialise into after they have good foundations in Programming, Maths and anything else that may be prevalent to where they want to end up. The typical student would probably do a CompSci undergrad degree, then do a dedicated Masters (or in some cases specialise in their final year of undergrad), or get experience working in a more traditional software development job. It's can't really just get a degree in CompSci (or Maths, or whatever) and then immediately jump into the deep end with AI.

Is there a reason why these are your only part time degree options? You mention that you have a long standing interest, does that come with any knowledge and/or experience?
1
reply
TonnyM
Badges: 7
Rep:
?
#3
Report Thread starter 9 months ago
#3
(Original post by Acsel)
While Maths is useful, to work with AI you're most likely going to need programming knowledge and the only degree that's really going to teach that properly is Computer Science. Obviously learning programming is possible outside a degree but one way or another it's not something you can totally ignore. Typically speaking, people don't just go straight into AI. It's the sort of thing people specialise into after they have good foundations in Programming, Maths and anything else that may be prevalent to where they want to end up. The typical student would probably do a CompSci undergrad degree, then do a dedicated Masters (or in some cases specialise in their final year of undergrad), or get experience working in a more traditional software development job. It's can't really just get a degree in CompSci (or Maths, or whatever) and then immediately jump into the deep end with AI.

Is there a reason why these are your only part time degree options? You mention that you have a long-standing interest, does that come with any knowledge and/or experience?
Well, I have to combine work and study. I have plenty of experience when it comes to programming and quite a bit in ML/Deep Learning, Q, I mostly play around with pySC2, trying to build a system able to mimic a human player. Computer Science degrees are a very "basic" and broad intro to a very large set of sub-disciplines, useless to me. AI research in its core is a discipline of Applied Mathematics mixed with Neuroscience, especially when it comes to the research relevant to AGI models.
0
reply
Acsel
Badges: 20
Rep:
?
#4
Report 9 months ago
#4
(Original post by TonnyM)
Well, I have to combine work and study. I have plenty of experience when it comes to programming and quite a bit in ML/Deep Learning, Q, I mostly play around with pySC2, trying to build a system able to mimic a human player. Computer Science degrees are a very "basic" and broad intro to a very large set of sub-disciplines, useless to me. AI research in its core is a discipline of Applied Mathematics mixed with Neuroscience, especially when it comes to the research relevant to AGI models.
I meant more of why are those the only degree options rather than why do you have to study part time.

Not all CompSci degrees are created equally. Not all are basic and it's not like knowledge of how the hardware works is unrelated to AI. Ultimately, if you're looking for relevant degree material, you're quite limited at the undergraduate level. For the most part a Maths undergrad will be just as basic and broad as a CompSci undergrad; most dedicated AI courses are at the Masters level. That's not to say there aren't undergrad degrees in AI available though and a quick Google search throws up Manchester and Edinburgh among others.

Realistically, you have various options. You could do a relevant undergrad in Maths or CompSci, then a Masters in AI. You could look for one of the more niche undergrad AI degrees. You could look for an undergrad degree that has elements of AI; plenty of Maths, CompSci and Engineering degrees offer it in some format. And of course you can develop your knowledge with self study. There's no right or wrong answer here. To answer the original question of "would a Maths degree be a waste of time and money" that's really up to you. A Maths degree won't inherently teach you AI. Only a dedicated AI degree will scratch the surface on that. If you choose to do anything that isn't AI focused, it's entirely up to you to bridge the gap; this is true whether you pick Maths, CompSci, Engineering, etc.
0
reply
TonnyM
Badges: 7
Rep:
?
#5
Report Thread starter 9 months ago
#5
(Original post by Acsel)
I meant more of why are those the only degree options rather than why do you have to study part time.

Not all CompSci degrees are created equally. Not all are basic and it's not like knowledge of how the hardware works is unrelated to AI. Ultimately, if you're looking for relevant degree material, you're quite limited at the undergraduate level. For the most part a Maths undergrad will be just as basic and broad as a CompSci undergrad; most dedicated AI courses are at the Masters level. That's not to say there aren't undergrad degrees in AI available though and a quick Google search throws up Manchester and Edinburgh among others.

Realistically, you have various options. You could do a relevant undergrad in Maths or CompSci, then a Masters in AI. You could look for one of the more niche undergrad AI degrees. You could look for an undergrad degree that has elements of AI; plenty of Maths, CompSci and Engineering degrees offer it in some format. And of course you can develop your knowledge with self study. There's no right or wrong answer here. To answer the original question of "would a Maths degree be a waste of time and money" that's really up to you. A Maths degree won't inherently teach you AI. Only a dedicated AI degree will scratch the surface on that. If you choose to do anything that isn't AI focused, it's entirely up to you to bridge the gap; this is true whether you pick Maths, CompSci, Engineering, etc.
I can do MSc in Mechanical or Electrical Engineering without doing BEng first, but not Machine Learning or CogSci and I would not qualify for Chartered Status. Oddly enough I was told that I can do PhD (again Engineering) without having a degree (because of my work experience and references), but I would not qualify for any funding......
Degree in CompSci won't give me anything I cannot Google or figure out on my own.
I am doing part-time because I have a mortgage to pay. I need adult level income to make it work.

I feel likeI can extract more value from math-heavy path to MEng Engineering (Modelling/Desing)... I don't know...
0
reply
Acsel
Badges: 20
Rep:
?
#6
Report 9 months ago
#6
(Original post by TonnyM)
Degree in CompSci won't give me anything I cannot Google or figure out on my own.

I feel likeI can extract more value from math-heavy path to MEng Engineering (Modelling/Desing)... I don't know...
That's sadly not how the world works though. If a programme or job wanted you to have a degree in CompSci, the ability to Google it yourself won't cut it. Because let's face it, you can Google pretty much every degree, that's not the same as having the degree. I'm not saying you absolutely need to do CompSci, I'm simply saying you are going to struggle to jump in at the top without evidence from the bottom. That might mean a CompSci degree, or a Maths degree, or evidence of your own knowledge. Or it might mean doing a dedicated AI undergrad degree.

If that's where you feel you'll get the most value, then that's your answer.
0
reply
TonnyM
Badges: 7
Rep:
?
#7
Report Thread starter 9 months ago
#7
(Original post by Acsel)
That's sadly not how the world works though. If a programme or job wanted you to have a degree in CompSci, the ability to Google it yourself won't cut it. Because let's face it, you can Google pretty much every degree, that's not the same as having the degree. I'm not saying you absolutely need to do CompSci, I'm simply saying you are going to struggle to jump in at the top without evidence from the bottom. That might mean a CompSci degree, or a Maths degree, or evidence of your own knowledge. Or it might mean doing a dedicated AI undergrad degree.

If that's where you feel you'll get the most value, then that's your answer.
I think that generic undergraduate degrees in CompSci are obsolete and unnecessary. They made sense in the past, before the internet and MOOCs. Back then Universities were the only and sole source of relevant knowledge but those times are gone, although when it comes the rest of the STEM, students usually need guidance, tutoring and access to specialised and very expensive equipment. This is why I feel like I would be wasting lots of time and money by doing CS just for the magical BSc next to my surname...
0
reply
Acsel
Badges: 20
Rep:
?
#8
Report 9 months ago
#8
(Original post by TonnyM)
I think that generic undergraduate degrees in CompSci are obsolete and unnecessary. They made sense in the past, before the internet and MOOCs. Back then Universities were the only and sole source of relevant knowledge but those times are gone, although when it comes the rest of the STEM, students usually need guidance, tutoring and access to specialised and very expensive equipment. This is why I feel like I would be wasting lots of time and money by doing CS just for the magical BSc next to my surname...
You're placing an awful lot of faith in students; many CompSci students wouldn't get through without guidance and support. If learning everything online were a viable alternative, people wouldn't be paying £9250 a year for a CompSci degree. Not to mention university isn't just about learning content. It's the total opposite, CompSci is more necessary now than ever but at this point it's a total tangent to the thread.
0
reply
DFranklin
Badges: 18
Rep:
?
#9
Report 9 months ago
#9
(Original post by TonnyM)
I think that generic undergraduate degrees in CompSci are obsolete and unnecessary. They made sense in the past, before the internet and MOOCs. Back then Universities were the only and sole source of relevant knowledge but those times are gone, although when it comes the rest of the STEM, students usually need guidance, tutoring and access to specialised and very expensive equipment. This is why I feel like I would be wasting lots of time and money by doing CS just for the magical BSc next to my surname...
FWIW, here's my (20+ year old) experience:

I was always interested in both maths and computing. I eventually decided to do maths because I felt "I'm going to do computing anyhow, but I need to be taught maths". My first degree was a first in maths from Cambridge. I was a pretty good coder and had done some paid programming work, but I wanted to get a job in computer graphics (which was pretty niche back then), and I wasn't successful in finding anything.

I then did a MSc in computer science at a "mid-level" university. It honestly didn't teach me very much - I was somewhat shocked at the lack of intellectual rigour compared with maths (and I think this would have been true even if I'd done maths at a more mid-level place), There were two of us with maths/physics degrees (everyone else had a BSc in comp-sci) and we 2 came top of the year by a fairly comfortable margin.

Subsequent to getting the MSc, I got my first job in "the industry", which I've now worked in for over 20 years. Although I don't use that much of what I learned mathematically, there's certainly a fair chunk of stuff at work which falls in the "we need DFranklin to work out how to do that". It's as much the training in analytical thought as the actual material knowledge.

Overall, my feeling is that although in many ways what I got out of the MSc was "a piece of paper", it was an important piece of paper in terms of getting that first job. So it may well be best to "suck it up" and do the degree that most visibly "ticks the right boxes", rather than the one you think would work best.

Possibly an important caveat: my MSc was a one year course, and I still frequently found myself frustrated by lecturers with no "real world" experience etc. I could see a 3-4 year course being very difficult to bear.
2
reply
JudaicImposter
Badges: 15
Rep:
?
#10
Report 9 months ago
#10
(Original post by ltsmith)
come to edinburgh and study our AI degree
What percentage of the Minf Informatics course can be specialised in the study of AI?
Reference: https://www.ed.ac.uk/studying/underg...view&code=G500
0
reply
winterscoming
Badges: 19
Rep:
?
#11
Report 9 months ago
#11
(Original post by TonnyM)
I think that generic undergraduate degrees in CompSci are obsolete and unnecessary. They made sense in the past, before the internet and MOOCs. Back then Universities were the only and sole source of relevant knowledge but those times are gone, although when it comes the rest of the STEM, students usually need guidance, tutoring and access to specialised and very expensive equipment. This is why I feel like I would be wasting lots of time and money by doing CS just for the magical BSc next to my surname...
What are your actual goals? If you're not too bothered with having the letters BSc after your name, and you're already comfortable with self-teaching, perhaps you'd be better off sticking just to distance learning and focusing on bringing your skills up to the standard that will get you into the kinds of jobs you're after.

There's tonnes of good content on https://www.edx.org/ and https://www.coursera.org/ which is all available for free as long as you don't want a certificate

Also https://eu.udacity.com/ has loads of free beginner-intermediate content. It also has more advanced content and nanodegrees which aren't free, and are a little expensive, but still much cheaper than a brick university - a lot of the content has been created by various global tech giants like Google and Amazon.

https://www.pluralsight.com/ is a monthly subscription service with a lot of excellent video-based tech content for specific technologies. Again, quite a lot of content here is written by senior figures at all kinds of different tech companies, although there structure of pluralsight is a lot less formal since it's pretty much all video-based.

https://teamtreehouse.com/tracks is another decent pay-for service which has a lot of content around programming and web development - Treehouse develop most of their own content, but they seem to have quite a bit of good content.
Last edited by winterscoming; 9 months ago
0
reply
TonnyM
Badges: 7
Rep:
?
#12
Report Thread starter 9 months ago
#12
(Original post by JudaicImposter)
What percentage of the Minf Informatics course can be specialised in the study of AI?
Reference: https://www.ed.ac.uk/studying/underg...view&code=G500
As far as I can tell it's usually about 10-12 lectures in Machine Learning, rest is a generic CompSci.
0
reply
TonnyM
Badges: 7
Rep:
?
#13
Report Thread starter 9 months ago
#13
(Original post by winterscoming)
What are your actual goals? If you're not too bothered with having the letters BSc after your name, and you're already comfortable with self-teaching, perhaps you'd be better off sticking just to distance learning and focusing on bringing your skills up to the standard that will get you into the kinds of jobs you're after.

There's tonnes of good content on https://www.edx.org/ and https://www.coursera.org/ which is all available for free as long as you don't want a certificate

Also https://eu.udacity.com/ has loads of free beginner-intermediate content. It also has more advanced content and nanodegrees which aren't free, and are a little expensive, but still much cheaper than a brick university - a lot of the content has been created by various global tech giants like Google and Amazon.

https://www.pluralsight.com/ is a monthly subscription service with a lot of excellent video-based tech content for specific technologies. Again, quite a lot of content here is written by senior figures at all kinds of different tech companies, although there structure of pluralsight is a lot less formal since it's pretty much all video-based.

https://teamtreehouse.com/tracks is another decent pay-for service which has a lot of content around programming and web development - Treehouse develop most of their own content, but they seem to have quite a bit of good content.
I am doing various MOOCs for nearly a decade. I've started with MIT OCW in 2009 and watching stuff via Edx somehow became my hobby Udemy is good too. Watching Youtube lectures and tutorials got me my current job. But to get into AI/DeepLearning seems to be much harder, all ads require a degree in CompSci, EE, Maths or Physics, without those three letters next to my surname my CV get automatically deleted. I cannot prove what I know and/or can do without meeting them first and I won't meet them because my CV won't make it through the first stage...
0
reply
TonnyM
Badges: 7
Rep:
?
#14
Report Thread starter 9 months ago
#14
(Original post by DFranklin)
FWIW, here's my (20+ year old) experience:

I was always interested in both maths and computing. I eventually decided to do maths because I felt "I'm going to do computing anyhow, but I need to be taught maths". My first degree was a first in maths from Cambridge. I was a pretty good coder and had done some paid programming work, but I wanted to get a job in computer graphics (which was pretty niche back then), and I wasn't successful in finding anything.

I then did a MSc in computer science at a "mid-level" university. It honestly didn't teach me very much - I was somewhat shocked at the lack of intellectual rigour compared with maths (and I think this would have been true even if I'd done maths at a more mid-level place), There were two of us with maths/physics degrees (everyone else had a BSc in comp-sci) and we 2 came top of the year by a fairly comfortable margin.

Subsequent to getting the MSc, I got my first job in "the industry", which I've now worked in for over 20 years. Although I don't use that much of what I learned mathematically, there's certainly a fair chunk of stuff at work which falls in the "we need DFranklin to work out how to do that". It's as much the training in analytical thought as the actual material knowledge.

Overall, my feeling is that although in many ways what I got out of the MSc was "a piece of paper", it was an important piece of paper in terms of getting that first job. So it may well be best to "suck it up" and do the degree that most visibly "ticks the right boxes", rather than the one you think would work best.

Possibly an important caveat: my MSc was a one year course, and I still frequently found myself frustrated by lecturers with no "real world" experience etc. I could see a 3-4 year course being very difficult to bear.
I have already figured out that I have to "suck it up" and pick one. Currently, I am looking for an undergraduate degree from which I could extract the most value. Does it matter which one? I went through quite a few job ads, and it seems that any STEM degree with a decent Git repo would do.
I was thinking about doing Maths & Stats at Open University, it passes as a legit degree, and I don't have to deal with teenagers. They also happen to offer a degree in Engineering Desing/Modeling which would land me a better job in my company and it feels like it would be "good enough" for more AI-ish job applications.
Would it work?
.
0
reply
DFranklin
Badges: 18
Rep:
?
#15
Report 9 months ago
#15
(Original post by TonnyM)
I have already figured out that I have to "suck it up" and pick one.
What I meant is that I feel that you have a clear preference for a largely mathematical degree, but my own experience is that employers can be scared to find your degree didn't have a substantial computing component, and so you might be best off compromising on a degree that's more clearly relevant to getting a job in the future. (So "suck it up" meant "compromise on the the more commercially viable course").

Currently, I am looking for an undergraduate degree from which I could extract the most value. Does it matter which one? I went through quite a few job ads, and it seems that any STEM degree with a decent Git repo would do.
I was thinking about doing Maths & Stats at Open University, it passes as a legit degree, and I don't have to deal with teenagers. They also happen to offer a degree in Engineering Desing/Modeling which would land me a better job in my company and it feels like it would be "good enough" for more AI-ish job applications.
I don't really know very much about OU degrees to be honest - a big plus point is that I think they are very used to people doing a part-time degree while working a real full-time job - I know people doing part-time degrees in more standard institutions who have found the course really requires a "full-time" number of hours and have found that incompatible with their work.

I had a quick look trying to find which exact course you were talking about but I wasn't sure I was looking at the right one; if you could give an exact link I'll try to give a bit more assessment on content.

You've said that you're not scared of math, but have you done maths A-levels or similar?
0
reply
TonnyM
Badges: 7
Rep:
?
#16
Report Thread starter 9 months ago
#16
BSc Mathematics & Statistics http://www.open.ac.uk/courses/qualifications/q36 - should take about 3 to 4 years
MEng/BEng Engineering (Modeling and applications) http://www.open.ac.uk/courses/qualifications/m04 can take up to 8 years...

I have some Cisco and RedHat certs I used to work as a Network Engineer but to be honest I am not good with customers or random strangers in general. Also, by 2024 I will have a decent portfolio and a git repo full of my creations.

Yes, I have something that passes as A Level in Maths, and I spent many hours here https://ocw.mit.edu/courses/mathematics/ or on the Edx but my problem is that I don't have any structure to follow. All I learn is sort of ad-hoc, as I need it..
0
reply
Acsel
Badges: 20
Rep:
?
#17
Report 9 months ago
#17
(Original post by TonnyM)
Does it matter which one? I went through quite a few job ads, and it seems that any STEM degree with a decent Git repo would do.
In a word, no. Few STEM degrees, even the most relevant ones will make someone perfectly suited to go into AI. Employers understand this and expect a degree to be combined with experience and further research. They appreciate that doing a Maths or CompSci degree, or even a dedicated AI Masters doesn't give someone all the tools they need. Balance "extracting value from your degree" with "just getting the piece of paper an employer wants to see". Because even if it is a large financial commitment, you don't fit the traditional target audience for most degrees and as a result there's going to be a limit to what you'll get out of it compared to an 18 year that's just left college.
1
reply
TonnyM
Badges: 7
Rep:
?
#18
Report Thread starter 9 months ago
#18
I guess that I will go with BSc Mathematics & Statistics from Open University...
0
reply
X

Quick Reply

Attached files
Write a reply...
Reply
new posts
Back
to top
Latest
My Feed

See more of what you like on
The Student Room

You can personalise what you see on TSR. Tell us a little about yourself to get started.

Personalise

University open days

  • University of Surrey
    Postgraduate Open Afternoon Postgraduate
    Wed, 23 Oct '19
  • University of Bristol
    Undergraduate Open Afternoon Undergraduate
    Wed, 23 Oct '19
  • University of Exeter
    Undergraduate Open Day - Penryn Campus Undergraduate
    Wed, 23 Oct '19

Would you turn to a teacher if you were being bullied?

Yes (22)
24.72%
No (67)
75.28%

Watched Threads

View All