Comment report |
|
Lists all the questions in the survey and displays all the comments made to these questions, if applicable. |
Table of contents |
Report info
|
Question 1:
Do you conduct research in, teach, or study machine learning?
|
Question 2:
What do you do?
|
Question 3:
Which department(s) are you based in? (You may select more than one)
|
Question 4:
Which conferences do you regularly attend? (You may pick more than one)
|
Question 5:
Do you have (or have you had) students in machine learning?
|
Question 6:
What is your educational background/training/disciplinary background? (You may select more th...
|
Question 7:
Did your educational background prepare you well for research in machine learning?
|
Question 8:
What educational backgrounds do your students have? (You may select more than one)
|
Question 9:
Did the educational backgrounds of your students prepare them well for studying machine learni...
|
Question 10:
In your opinion, what will constitute good educational backgrounds for a graduate student ente...
|
Question 11:
What area(s) in machine learning do you work in? (You may select more than one)
|
Question 12:
What other area(s) do you work in? (You may select more than one)
|
Question 13:
In your opinion, what are core (foundation) areas of machine learning? (You may select as man...
|
Question 14:
In your opinion, machine learning is a subfield of statistics.
|
Question 15:
In your opinion, machine learning is a subfield of computer science.
|
Question 16:
In your opinion, machine learning is a subfield of cognitive science.
|
Question 17:
In your opinion, machine learning is a subfield of neuroscience.
|
Question 18:
In your opinion, machine learning is a subfield of engineering.
|
Question 19:
In your opinion, machine learning is a subfield of artificial intelligence.
|
Question 20:
In your opinion, machine learning is an interdisciplinary area.
|
Question 21:
In your opinion, machine learning is a a discipline distinct from other disciplines.
|
Question 22:
Any further comments regarding the relationship of machine learning to other disciplines?
|
Question 23:
In your opinion, in the university, machine learning should ideally be a group within computer...
|
Question 24:
In your opinion, in the university, machine learning should ideally be a group within neurosci...
|
Question 25:
In your opinion, in the university, machine learning should ideally be a group within engineer...
|
Question 26:
In your opinion, in the university, machine learning should ideally be a group within statisti...
|
Question 27:
In your opinion, in the university, machine learning should ideally be a group within psycholo...
|
Question 28:
In your opinion, in the university, machine learning should ideally be a spread across multipl...
|
Question 29:
In your opinion, in the university, machine learning should ideally be an interdisciplinary ce...
|
Question 30:
In your opinion, in the university, machine learning should ideally be its own department.
|
Question 31:
In your opinion, in the university, machine learning should be taught at the graduate level.
|
Question 32:
In your opinion, in the university, machine learning should be taught as a specialist track in...
|
Question 33:
In your opinion, in the university, machine learning should be taught in a dedicated undergrad...
|
Question 34:
In your opinion, in which departments should machine learning be taught?
|
Question 35:
Any further comments regarding the place of machine learning within the university?
|
Question 36:
What brought you into machine learning? (You may select more than one)
|
Question 37:
In your opinion, what are the goals of machine learning? (You may select more than one)
|
Question 38:
Do you have further comments regarding any part of this survey?
|
Report info |
|
Report date |
Tuesday, May 17, 2011 2:55:20 PM BST |
|
Start date |
Monday, April 4, 2011 12:47:00 AM BST |
|
Stop date |
Sunday, April 24, 2011 11:59:00 PM BST |
|
Stored responses |
248 |
|
Number of completed responses |
194 |
Question 1 |
|
Do you conduct research in, teach, or study machine learning? |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Yes
|
241 |
97.18% |
97.18% |
No
|
7 |
2.82% |
2.82% |
Sum: |
248 |
100% |
100% |
Total answered: 248 |
|
Question 2 |
|
What do you do? |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Graduate student
|
248 |
100% |
100% |
Sum: |
248 |
100% |
100% |
Total answered: 248 |
|
Question 3 |
|
Which department(s) are you based in? (You may select more than one) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Biology
|
6 |
1.64% |
2.42% |
Cognitive science
|
7 |
1.92% |
2.82% |
Computer science
|
176 |
48.22% |
70.97% |
Electrical engineering
|
28 |
7.67% |
11.29% |
Engineering sciences
|
21 |
5.75% |
8.47% |
Machine Learning
|
73 |
20% |
29.44% |
Mathematics
|
13 |
3.56% |
5.24% |
Neuroscience
|
15 |
4.11% |
6.05% |
Physics
|
1 |
0.27% |
0.4% |
Psychology
|
1 |
0.27% |
0.4% |
Social science
|
1 |
0.27% |
0.4% |
Statistics
|
12 |
3.29% |
4.84% |
Other:
|
11 |
3.01% |
4.44% |
Sum: |
365 |
100% |
100% |
Total answered: 248 |
|
Question 4 |
|
Which conferences do you regularly attend? (You may pick more than one) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
AAAI
|
30 |
6.71% |
18.52% |
ACL
|
16 |
3.58% |
9.88% |
AISTATS
|
32 |
7.16% |
19.75% |
ALT
|
1 |
0.22% |
0.62% |
CogSci
|
3 |
0.67% |
1.85% |
COLT
|
8 |
1.79% |
4.94% |
CoNLL
|
4 |
0.89% |
2.47% |
Cosyne
|
7 |
1.57% |
4.32% |
CVPR
|
21 |
4.7% |
12.96% |
ECCV
|
11 |
2.46% |
6.79% |
ECML
|
30 |
6.71% |
18.52% |
EMNLP
|
13 |
2.91% |
8.02% |
ICCV
|
14 |
3.13% |
8.64% |
ICML
|
73 |
16.33% |
45.06% |
IJCAI
|
28 |
6.26% |
17.28% |
IROS
|
9 |
2.01% |
5.56% |
ISMB
|
3 |
0.67% |
1.85% |
NAACL
|
10 |
2.24% |
6.17% |
NIPS
|
93 |
20.81% |
57.41% |
RECOMB
|
3 |
0.67% |
1.85% |
SIGIR
|
10 |
2.24% |
6.17% |
SIGGRAPH
|
1 |
0.22% |
0.62% |
Snowbird
|
5 |
1.12% |
3.09% |
UAI
|
22 |
4.92% |
13.58% |
Sum: |
447 |
100% |
100% |
Total answered: 162 |
|
Text input |
CHI
|
no regular ones
|
.
|
non
|
OHBM
|
ACM Multimedia
|
International Conference on Inductive Logic Programming (ILP)
|
WWW, KDD
|
ISMIR
|
KDD
|
MLSP, ISIT
|
ICPR, MCS, ICISP
|
KDD, PKDD, ICDM, SDM
|
have just attended AISTATS
|
CHI, Ubicomp, Pervasive, IUI
|
CVPR
(clicked conferences where I have published, no regular attendance as a student)
|
KDD
|
www?kddEMNLP????
|
CDC
|
CoSyNe
|
Asian Conference On Machine Learning
|
KDD
|
RSS, NIPS, ICRA
|
FOCS, STOC
|
ICDAR
ICPR
GREC
|
I attended CICLING 2011, and plan to attend ACL 2011
|
ICRA, HRI
|
PRASA (local South African)
|
ICANN, ESANN, ECAI, ADPRL
|
Gecco, icann, agi, ppsn, cec
|
ESWC, ISWC
|
Dessert
|
I have a attended once ICML 2010
|
UBICOMP
|
CVPR, CBMI, MM, etc.
|
ACML, PAKDD
|
AAMAS, ICRA
|
ICRA, AAMAS
|
I used established ML as opposed to contribute to the leading research.
|
UAI
|
ICRA, IJCNN, ICANN
|
I am very new to this area and once attended to ECTI in Thailand.
|
KDD
|
ICRA, CDC, IFAC
|
ICCBR
|
|
Question 5 |
|
Do you have (or have you had) students in machine learning? |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Yes
|
29 |
11.69% |
11.79% |
No
|
217 |
87.5% |
88.21% |
Sum: |
246 |
100% |
100% |
Total answered: 246 |
|
Question 6 |
|
What is your educational background/training/disciplinary background? (You may select more than one) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Biology
|
7 |
1.72% |
2.94% |
Cognitive science
|
8 |
1.96% |
3.36% |
Computer science
|
181 |
44.36% |
76.05% |
Economics
|
4 |
0.98% |
1.68% |
Electrical engineering
|
34 |
8.33% |
14.29% |
Engineering sciences
|
22 |
5.39% |
9.24% |
Machine Learning
|
43 |
10.54% |
18.07% |
Mathematics
|
48 |
11.76% |
20.17% |
Neuroscience
|
5 |
1.23% |
2.1% |
Operations research
|
3 |
0.74% |
1.26% |
Physics
|
15 |
3.68% |
6.3% |
Social science
|
1 |
0.25% |
0.42% |
Statistics
|
22 |
5.39% |
9.24% |
Other:
|
15 |
3.68% |
6.3% |
Sum: |
408 |
100% |
100% |
Total answered: 238 |
|
Question 7 |
|
Did your educational background prepare you well for research in machine learning? |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Poorly prepared)
|
13 |
5.24% |
5.46% |
2
|
65 |
26.21% |
27.31% |
3
|
115 |
46.37% |
48.32% |
4 (Well prepared)
|
45 |
18.15% |
18.91% |
Not answered: |
10 |
4.03% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 238 |
|
Question 8 |
|
What educational backgrounds do your students have? (You may select more than one) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Cognitive science
|
1 |
2.04% |
3.7% |
Computer science
|
22 |
44.9% |
81.48% |
Economics
|
2 |
4.08% |
7.41% |
Electrical engineering
|
7 |
14.29% |
25.93% |
Engineering sciences
|
5 |
10.2% |
18.52% |
Machine Learning
|
5 |
10.2% |
18.52% |
Mathematics
|
4 |
8.16% |
14.81% |
Physics
|
1 |
2.04% |
3.7% |
Other:
|
2 |
4.08% |
7.41% |
Sum: |
49 |
100% |
100% |
Total answered: 27 |
|
Question 9 |
|
Did the educational backgrounds of your students prepare them well for studying machine learning? |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Poorly prepared)
|
5 |
2.02% |
18.52% |
2
|
9 |
3.63% |
33.33% |
3
|
10 |
4.03% |
37.04% |
4 (Well prepared)
|
3 |
1.21% |
11.11% |
Not answered: |
221 |
89.11% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 27 |
|
Question 10 |
|
In your opinion, what will constitute good educational backgrounds for a graduate student entering machine learning? (You may select as many as you like). |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Biology
|
4 |
0.41% |
1.72% |
Cognitive science
|
34 |
3.46% |
14.59% |
Computer science
|
193 |
19.65% |
82.83% |
Economics
|
7 |
0.71% |
3% |
Electrical engineering
|
42 |
4.28% |
18.03% |
Engineering sciences
|
22 |
2.24% |
9.44% |
Machine Learning
|
172 |
17.52% |
73.82% |
Mathematics
|
185 |
18.84% |
79.4% |
Neuroscience
|
19 |
1.93% |
8.15% |
Operations research
|
50 |
5.09% |
21.46% |
Physics
|
49 |
4.99% |
21.03% |
Psychology
|
10 |
1.02% |
4.29% |
Social science
|
8 |
0.81% |
3.43% |
Statistics
|
183 |
18.64% |
78.54% |
Other:
|
4 |
0.41% |
1.72% |
Sum: |
982 |
100% |
100% |
Total answered: 233 |
|
Text input |
applied mathematics
|
Some basic courses in optimization and Decision Sciences too are significant in Machine Learning
|
Actually, I think many different backgrounds could be beneficial for the community.
This seems to be the most obvious to me, though.
|
I'm inclined to select almost all of them. Some experience in Maths is the only common denominator I see in all of them.
|
It depends on the focus - I answered for more theoretically inclined machine learning but some researchers are interested in biological systems or in the learning done in the human brain, for them other disciplines are relevant.
|
Encountered many people with very weak math preparation due to coming from undergrad computer science background, which offers exceedingly poor math preparation even at top universities. Often run into (otherwise good and motivated) machine learning graduate students with very poor understanding of linear algebra other than the most basic symbol pushing, dodgy understanding of what random variables are, and essentially no capacity to absorb more sophisticated concepts like RKHS, Bayesian nonparametrics, etc.
|
Later on, one has to learn about the application, too. But since applications are widespread, one cannot know everything beforehand. My biology knowledge has helped me partially.
|
Computer science training is essential because MLer need to do coding.
|
From my perspective, a requrement for success is first of all a pure or "semi-pure" mathematic background (incl basic statstics) combined with a more physical aspect such as physics or EE a good basis. Computer Science is to broard to be mentioned, yet of course important as a pratical tool. But the mathematics seems to be more and more important to understand the very complicated models.
|
I guess Machine Learning would prepare a student well for Machine Learning. However, since that is non-informative, I chose disciplines OTHER THAN Machine Learning that would give a student a strong enough background to do well in Machine Learning.
|
A solid mathematical background and the ability to write code are most important in my opinion.
|
Knowledge of: neural networks, statistics, computational neuroscience, computer programming, mathematical optimization, game theory, information theory, artificial intelligence, computational intelligence, machine learning, complexity theory. Some knowledge of engineering and physics might be advantageous.
|
Specifically, a solid understanding of both linear algebra and significant programming experience as a hobby, not as a job, are the right combination. It is hard to find both together. Those who lack the math skills will spend all their time catching up, while the types who choose to use high-level programming languages rarely have drive sufficient to undertake ambitious projects.
|
It's all about math and stats. Neuroscience and biology, etc., offer convenient high-level metaphors for some approaches, but they disappear when you look at the details.
|
practical experiences, e.g. with one of the many ml competitions cannot be overvalued
|
The fields I chose will give a good fundamental background in machine learning, though a background in the other fields listed will give imperative domain knowledge and at least one should be studied in some depth.
|
What I've missed the most is a mathematically rigorous, unified introduction to statistics and the parts of machine learning closest to statistics.
|
Bayesian Statistics
|
statistical physics in particular.
|
Statistics is really important... In Iran it didn't receive much attention in my undergrad school but I studied hard myself.
|
|
Question 11 |
|
What area(s) in machine learning do you work in? (You may select more than one) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Bayesian statistics
|
99 |
10.43% |
46.92% |
Control and planning
|
25 |
2.63% |
11.85% |
Deep learning
|
13 |
1.37% |
6.16% |
Dimensionality reduction
|
53 |
5.58% |
25.12% |
Graphical models
|
91 |
9.59% |
43.13% |
Information theory
|
18 |
1.9% |
8.53% |
Kernel methods
|
46 |
4.85% |
21.8% |
Learning theory
|
21 |
2.21% |
9.95% |
Manifold learning
|
12 |
1.26% |
5.69% |
Model selection
|
32 |
3.37% |
15.17% |
Neural networks
|
31 |
3.27% |
14.69% |
Nonparametric methods
|
46 |
4.85% |
21.8% |
Online learning
|
38 |
4% |
18.01% |
Optimization
|
48 |
5.06% |
22.75% |
Reinforcement learning
|
39 |
4.11% |
18.48% |
Relational learning
|
18 |
1.9% |
8.53% |
Semisupervised learning
|
44 |
4.64% |
20.85% |
Sparse learning
|
36 |
3.79% |
17.06% |
Statistical physics of learning
|
3 |
0.32% |
1.42% |
Structured learning
|
38 |
4% |
18.01% |
Supervised learning
|
85 |
8.96% |
40.28% |
Time series modelling
|
38 |
4% |
18.01% |
Unsupervised learning
|
69 |
7.27% |
32.7% |
Other:
|
6 |
0.63% |
2.84% |
Sum: |
949 |
100% |
100% |
Total answered: 211 |
|
Question 12 |
|
What other area(s) do you work in? (You may select more than one) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
None
|
11 |
1.78% |
5.21% |
Artificial intelligence
|
50 |
8.08% |
23.7% |
Bioinformatics
|
25 |
4.04% |
11.85% |
Brain, medical imaging
|
10 |
1.62% |
4.74% |
Cognitive science
|
6 |
0.97% |
2.84% |
Collaborative filtering
|
23 |
3.72% |
10.9% |
Communications
|
4 |
0.65% |
1.9% |
Computer vision
|
48 |
7.75% |
22.75% |
Computer science
|
59 |
9.53% |
27.96% |
Databases
|
8 |
1.29% |
3.79% |
Data mining
|
45 |
7.27% |
21.33% |
Electrical engineering
|
11 |
1.78% |
5.21% |
Genetics
|
13 |
2.1% |
6.16% |
Hardware technologies
|
1 |
0.16% |
0.47% |
Multimedia processing
|
6 |
0.97% |
2.84% |
Linguistics
|
15 |
2.42% |
7.11% |
Image processing
|
27 |
4.36% |
12.8% |
Information retrieval
|
32 |
5.17% |
15.17% |
Mathematics
|
10 |
1.62% |
4.74% |
Natural language processing
|
34 |
5.49% |
16.11% |
Network analysis
|
13 |
2.1% |
6.16% |
Neural data analysis
|
4 |
0.65% |
1.9% |
Neuroscience
|
11 |
1.78% |
5.21% |
Operations research
|
5 |
0.81% |
2.37% |
Physics
|
2 |
0.32% |
0.95% |
Psychology
|
5 |
0.81% |
2.37% |
Robotics
|
22 |
3.55% |
10.43% |
Signal processing
|
24 |
3.88% |
11.37% |
Social science
|
6 |
0.97% |
2.84% |
Speech processing
|
9 |
1.45% |
4.27% |
Statistics
|
23 |
3.72% |
10.9% |
Systems biology
|
8 |
1.29% |
3.79% |
Text and web analysis
|
40 |
6.46% |
18.96% |
Other:
|
9 |
1.45% |
4.27% |
Sum: |
619 |
100% |
100% |
Total answered: 211 |
|
Question 13 |
|
In your opinion, what are core (foundation) areas of machine learning? (You may select as many as you see fit) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Bayesian statistics
|
141 |
9.66% |
67.14% |
Control and planning
|
22 |
1.51% |
10.48% |
Deep learning
|
20 |
1.37% |
9.52% |
Dimensionality reduction
|
59 |
4.04% |
28.1% |
Graphical models
|
103 |
7.06% |
49.05% |
Information theory
|
88 |
6.03% |
41.9% |
Kernel methods
|
79 |
5.41% |
37.62% |
Learning theory
|
113 |
7.75% |
53.81% |
Manifold learning
|
18 |
1.23% |
8.57% |
Model selection
|
57 |
3.91% |
27.14% |
Neural networks
|
43 |
2.95% |
20.48% |
Nonparametric methods
|
52 |
3.56% |
24.76% |
Online learning
|
44 |
3.02% |
20.95% |
Optimization
|
123 |
8.43% |
58.57% |
Reinforcement learning
|
74 |
5.07% |
35.24% |
Relational learning
|
19 |
1.3% |
9.05% |
Semisupervised learning
|
51 |
3.5% |
24.29% |
Sparse learning
|
29 |
1.99% |
13.81% |
Statistical physics of learning
|
19 |
1.3% |
9.05% |
Structured learning
|
36 |
2.47% |
17.14% |
Supervised learning
|
127 |
8.7% |
60.48% |
Time series modelling
|
24 |
1.64% |
11.43% |
Unsupervised learning
|
115 |
7.88% |
54.76% |
Other:
|
3 |
0.21% |
1.43% |
Sum: |
1459 |
100% |
100% |
Total answered: 210 |
|
Question 14 |
|
In your opinion, machine learning is a subfield of statistics. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
29 |
11.69% |
14.36% |
2
|
45 |
18.15% |
22.28% |
3
|
85 |
34.27% |
42.08% |
4 (Agree)
|
43 |
17.34% |
21.29% |
Not answered: |
46 |
18.55% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 202 |
|
Question 15 |
|
In your opinion, machine learning is a subfield of computer science. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
10 |
4.03% |
4.95% |
2
|
40 |
16.13% |
19.8% |
3
|
89 |
35.89% |
44.06% |
4 (Agree)
|
63 |
25.4% |
31.19% |
Not answered: |
46 |
18.55% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 202 |
|
Question 16 |
|
In your opinion, machine learning is a subfield of cognitive science. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
82 |
33.06% |
40.59% |
2
|
84 |
33.87% |
41.58% |
3
|
28 |
11.29% |
13.86% |
4 (Agree)
|
8 |
3.23% |
3.96% |
Not answered: |
46 |
18.55% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 202 |
|
Question 17 |
|
In your opinion, machine learning is a subfield of neuroscience. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
103 |
41.53% |
50.99% |
2
|
72 |
29.03% |
35.64% |
3
|
20 |
8.06% |
9.9% |
4 (Agree)
|
7 |
2.82% |
3.47% |
Not answered: |
46 |
18.55% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 202 |
|
Question 18 |
|
In your opinion, machine learning is a subfield of engineering. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
57 |
22.98% |
28.22% |
2
|
55 |
22.18% |
27.23% |
3
|
65 |
26.21% |
32.18% |
4 (Agree)
|
25 |
10.08% |
12.38% |
Not answered: |
46 |
18.55% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 202 |
|
Question 19 |
|
In your opinion, machine learning is a subfield of artificial intelligence. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
19 |
7.66% |
9.36% |
2
|
29 |
11.69% |
14.29% |
3
|
77 |
31.05% |
37.93% |
4 (Agree)
|
78 |
31.45% |
38.42% |
Not answered: |
45 |
18.15% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 203 |
|
Question 20 |
|
In your opinion, machine learning is an interdisciplinary area. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
2 |
0.81% |
0.99% |
2
|
9 |
3.63% |
4.46% |
3
|
47 |
18.95% |
23.27% |
4 (Agree)
|
144 |
58.06% |
71.29% |
Not answered: |
46 |
18.55% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 202 |
|
Question 21 |
|
In your opinion, machine learning is a a discipline distinct from other disciplines. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
31 |
12.5% |
15.35% |
2
|
66 |
26.61% |
32.67% |
3
|
64 |
25.81% |
31.68% |
4 (Agree)
|
41 |
16.53% |
20.3% |
Not answered: |
46 |
18.55% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 202 |
|
Question 22 |
|
Any further comments regarding the relationship of machine learning to other disciplines? |
|
Text input |
I see it as a subfield of AI that is an intersection of statistics/information theory with computer science and optimization.
|
Applied mathematics
|
machine learning provides useful tools that can be applied to solve problems in many other disciplines.
|
What i feel or enjoy about Machine Learning is that no one can take ownership to it and there is no boundary. It is much an Electrical Engineer's as well as a Statistician and it is our responsibility to nurture it. We are seeing some interesting work in the current days and i hope to see more and be a part of it.
|
It is a pity that it is not more promoted, e.g. in Maths Departments.
|
it is a tool for other disciplines
|
In our lab, we are specifically digging connections between signal processing and machine learning.
There's a lot of issues both fields try to adress. We use a lot of common tools and methods. Still, we often have pretty different views. That's stimulating.
|
Machine Learning, in general, is tightly associated with statistics, mathematical logic and computer science. However, since Machine Learning is applied into a variety of different domains (e.g natural language processing, medical etc), it has strong relationship with the application domain.
|
There seem to be at least two parts of machine learning. Some people just go by standard mathematics, algorithms, statistics, optimization, and so on. Other people attempt, with varying degrees of success, to involve all kinds of "neurological" and "biological" interpretation. It seems a wide range of disciplines can follow the work in the former, but the work in the latter is less respected in other technical fields, and even by the other more technically-oriented part of machine learning.
|
draws from other disciplines (statistics, optimization, ...)
and yields methods for other fields (applications)
|
Machine learning is a marriage of artificial intelligence, statistics, and computer science.
|
Machine learning is its own discipline and it is a subfield of computer science. It has strong connections to other fields, but it is defined by a distinct collection (although items in the collection may not be unique) of attitudes, problems, techniques, tools, and concepts.
|
Lance fortnow recently lamented the lack of role that theoretical computer science has in modern machine learning. I agree with him that it should play an increased role in both machine learning education AND practice.
|
wow that was confusing choices :)
|
Machine Learning is a useful category, I personally tend to exclude Reinforcement Learning from it and place RL as a sibling of ML under AI. Integrated AI systems are not based solely on ML or RL (but at least on both, that's why usually RL viewed as part of ML). ML can be used on its own in engineering or software systems.
|
An interesting poling question with degree of agree/disagree might be "On average, machine learning in Electrical Engineering is more theoretical and machine learning in computer science is more focused on applications."
|
ML has become so diverse that even veteran experts often lack a sufficiently deep understanding of the motivations and foundations of the areas where I operate. Hence, greater fragmentation and specialization within the top ML conferences and journals would be very welcome. In the past, a rift had started to form between the neural approaches and the more traditional classification models. The community, however, worked to mend the differences and hold together as a unified field. Now, a new divide has started to form between the statisticians/Bayesian/graphical-model approaches and the rest of the community. Unfortunately, it seems we are again working hard to hold it all together. The science would be better served by promoting modularization, rather than creating a bloated and monolithic field in which communication and point-of-view difficulties drag at advancement.
|
I guess you miss the empirical process theory Although some may consider it as a subfield of statistics.
|
I've learned it as part of computer science at two schools, but it fits almost as well under statistics. (AI is also part of CS.)
|
I'm very fond one analogy on Hal Daume's blog (nlpers.blogspot.com):
Machine Learning is computational statistics that doesn't care about the statistics
Computational statistics is machine learning that doesn't care about the computation
|
None
|
From the previous page, the subsetting relation between ML and Stat is much stronger than others.
|
I think that right now, the bread and butter of ML is finding models that can capture most of the important structure in data, while still being tractactable. So any discipline that is used to approximate computation (like computational chemistry) would feel at home in ML.
|
In order to create new models, insight into disciplines such as biology and cognitive science may help to model the success of living organisms, however realistically machine learning is more strongly based in statistics and computer science. Though a physics/ applied mathematics background would be useful as these subjects study the creation of new models!
|
|
Question 23 |
|
In your opinion, in the university, machine learning should ideally be a group within computer science. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
9 |
3.63% |
4.57% |
2
|
34 |
13.71% |
17.26% |
3
|
79 |
31.85% |
40.1% |
4 (Agree)
|
75 |
30.24% |
38.07% |
Not answered: |
51 |
20.56% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 197 |
|
Question 24 |
|
In your opinion, in the university, machine learning should ideally be a group within neuroscience. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
99 |
39.92% |
50.25% |
2
|
75 |
30.24% |
38.07% |
3
|
20 |
8.06% |
10.15% |
4 (Agree)
|
3 |
1.21% |
1.52% |
Not answered: |
51 |
20.56% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 197 |
|
Question 25 |
|
In your opinion, in the university, machine learning should ideally be a group within engineering. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
50 |
20.16% |
25.25% |
2
|
80 |
32.26% |
40.4% |
3
|
46 |
18.55% |
23.23% |
4 (Agree)
|
22 |
8.87% |
11.11% |
Not answered: |
50 |
20.16% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 198 |
|
Question 26 |
|
In your opinion, in the university, machine learning should ideally be a group within statistics. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
22 |
8.87% |
11.11% |
2
|
75 |
30.24% |
37.88% |
3
|
79 |
31.85% |
39.9% |
4 (Agree)
|
22 |
8.87% |
11.11% |
Not answered: |
50 |
20.16% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 198 |
|
Question 27 |
|
In your opinion, in the university, machine learning should ideally be a group within psychology. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
133 |
53.63% |
67.17% |
2
|
58 |
23.39% |
29.29% |
3
|
6 |
2.42% |
3.03% |
4 (Agree)
|
1 |
0.4% |
0.51% |
Not answered: |
50 |
20.16% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 198 |
|
Question 28 |
|
In your opinion, in the university, machine learning should ideally be a spread across multiple departments. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
39 |
15.73% |
19.8% |
2
|
44 |
17.74% |
22.34% |
3
|
59 |
23.79% |
29.95% |
4 (Agree)
|
55 |
22.18% |
27.92% |
Not answered: |
51 |
20.56% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 197 |
|
Question 29 |
|
In your opinion, in the university, machine learning should ideally be an interdisciplinary centre. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
11 |
4.44% |
5.56% |
2
|
17 |
6.85% |
8.59% |
3
|
83 |
33.47% |
41.92% |
4 (Agree)
|
87 |
35.08% |
43.94% |
Not answered: |
50 |
20.16% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 198 |
|
Question 30 |
|
In your opinion, in the university, machine learning should ideally be its own department. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
37 |
14.92% |
18.78% |
2
|
55 |
22.18% |
27.92% |
3
|
50 |
20.16% |
25.38% |
4 (Agree)
|
55 |
22.18% |
27.92% |
Not answered: |
51 |
20.56% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 197 |
|
Question 31 |
|
In your opinion, in the university, machine learning should be taught at the graduate level. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
12 |
4.84% |
6.15% |
2
|
18 |
7.26% |
9.23% |
3
|
44 |
17.74% |
22.56% |
4 (Agree)
|
121 |
48.79% |
62.05% |
Not answered: |
53 |
21.37% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 195 |
|
Question 32 |
|
In your opinion, in the university, machine learning should be taught as a specialist track in a related undergraduate programme. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
12 |
4.84% |
6.09% |
2
|
27 |
10.89% |
13.71% |
3
|
74 |
29.84% |
37.56% |
4 (Agree)
|
84 |
33.87% |
42.64% |
Not answered: |
51 |
20.56% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 197 |
|
Question 33 |
|
In your opinion, in the university, machine learning should be taught in a dedicated undergraduate programme. |
Frequency table |
Levels
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
1 (Disagree)
|
33 |
13.31% |
16.92% |
2
|
75 |
30.24% |
38.46% |
3
|
41 |
16.53% |
21.03% |
4 (Agree)
|
46 |
18.55% |
23.59% |
Not answered: |
53 |
21.37% |
- |
Sum: |
248 |
100% |
100% |
Total answered: 195 |
|
Question 34 |
|
In your opinion, in which departments should machine learning be taught? |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
Biology
|
31 |
3.07% |
15.74% |
Cognitive science
|
64 |
6.34% |
32.49% |
Computer science
|
186 |
18.42% |
94.42% |
Economics
|
38 |
3.76% |
19.29% |
Electrical engineering
|
61 |
6.04% |
30.96% |
Engineering sciences
|
72 |
7.13% |
36.55% |
Machine Learning
|
143 |
14.16% |
72.59% |
Mathematics
|
83 |
8.22% |
42.13% |
Neuroscience
|
57 |
5.64% |
28.93% |
Operations research
|
56 |
5.54% |
28.43% |
Physics
|
25 |
2.48% |
12.69% |
Psychology
|
23 |
2.28% |
11.68% |
Social science
|
19 |
1.88% |
9.64% |
Statistics
|
148 |
14.65% |
75.13% |
Other:
|
4 |
0.4% |
2.03% |
Sum: |
1010 |
100% |
100% |
Total answered: 197 |
|
Question 35 |
|
Any further comments regarding the place of machine learning within the university? |
|
Text input |
I think this will be the last field to have humans replaced by computers, so having a larger proportion of students go into it since early in their academic career is probably a good idea.
|
in my opinion, the core of machine learning is mostly closely related to statistics and computer science, and requiring considerable other domain knowledge from both.
|
I've studied maths as an undergraduate student, not so much after that.
I recently found out machine learning started giving me stronger and deeper insights and understanding on different fields of mathematics (mainly in linear algebra, numerical analysis, statistics) which I had found to be pretty disconnected from any reality.
In fact, I think machine learning could be a very nice way to get into various theoretical of applied fields, in undergraduate programmes.
|
Doing machine learning properly requires background in a number of different topics, so however and wherever it is taught, the students should be expected to have appropriate background in algorithms, complexity, programming, basic statistics, probability, linear algebra, optimization, some real analysis, and so on. Many courses just don't list many of these things as requirements but then end up relying on all these topics anyway, which leads to trouble for many students.
|
It should be in the same department where AI is.
|
I have no experience with any department of "Operations research" so can't tell about its relation to machine learning (although it sounds like a department with many AI applications).
|
Spread across separate departments is okay. There should just be an unprecedented amount of collaboration between the groups. Some schools do this pretty well (UWisc - Madison for one). Some schools fail spectacularly (UWash for one).
|
Computer science should be split into "IT", "software development" and "computational science". Then, ML clearly belongs as a part of computational science. If those three fields must be all lumped together under the banner of "computer science", then ML is no longer such a good fit, and probably belongs in the math department.
|
|
Question 36 |
|
What brought you into machine learning? (You may select more than one) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
I want to use machine learning techniques to analyse data.
|
128 |
27.71% |
66.32% |
I want to understand how the brain works.
|
43 |
9.31% |
22.28% |
I want to understand the fundamental principles of learning and intelligence.
|
95 |
20.56% |
49.22% |
I want to understand existing machine learning techniques and develop new ones.
|
111 |
24.03% |
57.51% |
I want to build intelligent computers or robots.
|
85 |
18.4% |
44.04% |
Sum: |
462 |
100% |
100% |
Total answered: 193 |
|
|
Question 37 |
|
In your opinion, what are the goals of machine learning? (You may select more than one) |
Frequency table |
Choices
|
Absolute frequency |
Relative frequency |
Adjusted relative frequency |
To understand the fundamental principles of learning and intelligence.
|
114 |
20.18% |
58.76% |
To analyze and extract useful information from data.
|
152 |
26.9% |
78.35% |
To understand how the brain works.
|
38 |
6.73% |
19.59% |
To build smarter computers or robots.
|
110 |
19.47% |
56.7% |
To develop algorithms for making accurate predictions based on data.
|
151 |
26.73% |
77.84% |
Sum: |
565 |
100% |
100% |
Total answered: 194 |
|
|
Question 38 |
|
Do you have further comments regarding any part of this survey? |
|
|
|