User OmG

34 votes

4 answers

18k views

Could a neural network detect primes?

Fullk33

453

asked May 26, 2017 at 15:15

31 votes

7 answers

8k views

What are examples of promising AI/ML techniques that are computationally intractable?

machine-learning reference-request computational-complexity intractability

k.c. sayz 'k.c sayz'

1,951

asked Oct 20, 2019 at 17:32

26 votes

8 answers

4k views

Is there any research on the development of attacks against artificial intelligence systems?

image-recognition reference-request voice-recognition ai-security adversarial-ml

Lion Lai

403

asked Oct 9, 2019 at 17:45

20 votes

3 answers

3k views

Why doesn't Q-learning converge when using function approximation?

reinforcement-learning q-learning deep-rl proofs function-approximation

nbro

34.3k

asked Apr 5, 2019 at 18:23

20 votes

1 answer

5k views

When should I use Reinforcement Learning vs PID Control?

reinforcement-learning ai-design control-theory

SeeDerekEngineer

491

asked May 22, 2019 at 15:24

16 votes

2 answers

2k views

When is deep learning overkill?

machine-learning deep-learning classification applications comparison

Alexander

283

asked Mar 16, 2017 at 11:41

15 votes

5 answers

749 views

What is the most general definition of "intelligence"?

definitions intelligence

user79161

349

asked Aug 18, 2018 at 14:17

14 votes

2 answers

776 views

Is there any scientific/mathematical argument that prevents deep learning from ever producing strong AI?

deep-learning agi superintelligence

Graviton

261

asked Jun 15, 2018 at 11:40

14 votes

4 answers

5k views

What is the relevance of AIXI on current artificial intelligence research?

models agi aixi

rcpinto

2,029

asked Aug 2, 2016 at 21:10

14 votes

2 answers

11k views

What are the main differences between skip-gram and continuous bag of words?

natural-language-processing comparison word2vec cbow skip-gram

DRV

1,263

asked Mar 14, 2020 at 6:52

13 votes

6 answers

3k views

Is there actually a lack of fundamental theory on deep learning?

deep-learning computational-learning-theory

heleone

141

asked Mar 15, 2017 at 9:10

13 votes

3 answers

5k views

Why exactly do neural networks require i.i.d. data?

neural-networks reinforcement-learning statistical-ai experience-replay iid

nbro

34.3k

asked Feb 23, 2019 at 13:30

11 votes

1 answer

972 views

What are the implications of the "No Free Lunch" theorem for machine learning?

machine-learning optimization proofs no-free-lunch-theorems

user9947

asked Sep 27, 2019 at 13:52

9 votes

2 answers

895 views

Are there dynamic neural networks?

neural-networks incremental-learning online-learning

TomR

773

asked Sep 14, 2018 at 7:28

9 votes

4 answers

2k views

Should neural nets be deeper the more complex the learning problem is?

neural-networks deep-learning architecture

Gilad Deutsch

589

asked Apr 27, 2020 at 13:21

9 votes

2 answers

1k views

What are the biggest barriers to get RL in production?

reinforcement-learning deep-rl applications

Alexandre Krul

103

asked Jan 28, 2021 at 2:11

7 votes

1 answer

812 views

How to measure sample efficiency of a reinforcement learning algorithm?

reinforcement-learning sample-efficiency

rert588

320

asked Jun 18, 2020 at 9:29

7 votes

1 answer

764 views

How should the neural network deal with unexpected inputs?

neural-networks deep-learning classification bayesian-deep-learning bayesian-neural-networks

Jeff

173

asked Mar 21, 2020 at 20:26

7 votes

0 answers

80 views

Normalizing Normal Distributions in Thompson Sampling for online Reinforcement Learning

reinforcement-learning online-learning normalisation thompson-sampling normal-distribution

Kevin

71

asked Nov 22, 2019 at 12:03

7 votes

1 answer

609 views

How do we know if GPT-2 is a better language model?

natural-language-processing transformer gpt

lcrmorin

232

asked Feb 25, 2019 at 9:51

6 votes

1 answer

974 views

How should we choose the dimensions of the encoding layer in auto-encoders?

neural-networks autoencoders hyperparameter-optimization variational-autoencoder hyper-parameters

Neha soni

101

asked Dec 27, 2018 at 17:26

6 votes

2 answers

442 views

Why is dropout favoured compared to reducing the number of units in hidden layers?

neural-networks machine-learning deep-learning dropout regularization

pascal sautot

231

asked Dec 11, 2019 at 16:26

5 votes

3 answers

313 views

How is neural architecture search performed?

neural-networks reference-request neuroevolution hyperparameter-optimization neural-architecture-search

Adam Geringer

51

asked May 20, 2019 at 15:59

5 votes

1 answer

324 views

Examples of ontologies made with AI

applications ontology

AndriuZ

153

asked Sep 8, 2019 at 18:00

5 votes

0 answers

60 views

It is possible to use deep learning to give approximate solutions to NP-hard graph theory problems?

neural-networks deep-learning tensorflow keras graph-theory

Jake B.

171

asked Mar 10, 2021 at 20:58

5 votes

3 answers

140 views

How widely accepted is the definition of intelligence by Marcus Hutter & Shane Legg?

agi definitions intelligent-agent intelligence

Aether

235

asked Mar 23, 2021 at 6:18

4 votes

0 answers

75 views

When do two identical neural networks have uncorrelated errors?

neural-networks proofs statistics ensemble-learning

EmmanuelMess

177

asked Jan 19, 2021 at 17:28

4 votes

1 answer

144 views

Is there a limit of minimum error for a particular training dataset in artificial Neural Network?

neural-networks math proofs

bipul kalita

79

asked Oct 28, 2018 at 9:37

4 votes

4 answers

415 views

What other kind of AIs exist apart from goal-driven?

agi definitions terminology

nbro

34.3k

asked Nov 29, 2017 at 1:43

4 votes

0 answers

59 views

Given an input $x \in R^{1\times d}$ and a network with $s$ hidden layers, is the time complexity of the forward pass $O(d^{2}s)$? [duplicate]

neural-networks feedforward-neural-networks time-complexity computational-complexity

Jonathan Azpur

103

asked Nov 30, 2019 at 20:50

Stack Exchange Network

current community

your communities

more stack exchange communities

47 Bookmarks

Could a neural network detect primes?

What are examples of promising AI/ML techniques that are computationally intractable?

Is there any research on the development of attacks against artificial intelligence systems?

Why doesn't Q-learning converge when using function approximation?

When should I use Reinforcement Learning vs PID Control?

When is deep learning overkill?

What is the most general definition of "intelligence"?

Is there any scientific/mathematical argument that prevents deep learning from ever producing strong AI?

What is the relevance of AIXI on current artificial intelligence research?

What are the main differences between skip-gram and continuous bag of words?

Is there actually a lack of fundamental theory on deep learning?

Why exactly do neural networks require i.i.d. data?

What are the implications of the "No Free Lunch" theorem for machine learning?

Are there dynamic neural networks?

Should neural nets be deeper the more complex the learning problem is?

What are the biggest barriers to get RL in production?

How to measure sample efficiency of a reinforcement learning algorithm?

How should the neural network deal with unexpected inputs?

Normalizing Normal Distributions in Thompson Sampling for online Reinforcement Learning

How do we know if GPT-2 is a better language model?

How should we choose the dimensions of the encoding layer in auto-encoders?

Why is dropout favoured compared to reducing the number of units in hidden layers?

How is neural architecture search performed?

Examples of ontologies made with AI

It is possible to use deep learning to give approximate solutions to NP-hard graph theory problems?

How widely accepted is the definition of intelligence by Marcus Hutter & Shane Legg?

When do two identical neural networks have uncorrelated errors?

Is there a limit of minimum error for a particular training dataset in artificial Neural Network?

What other kind of AIs exist apart from goal-driven?

Given an input $x \in R^{1\times d}$ and a network with $s$ hidden layers, is the time complexity of the forward pass $O(d^{2}s)$? [duplicate]