Queensland University of Technology
Question
Asked 6th Mar, 2016
What is the pros and cons of Convolutional neural networks?
Hi researchers! I am a learner of statistics learing and machine learning. After applying the Convolutional neural networks into image recognition and text mining, I think this method is powerful in classification. So I want to apply them into statistics field and want to know the advantages and disadvantages of CNNs. Can I view it as a specia kind of "regression", which is blessed with flexible model form and interaction form？ What is the disadvantage of it? Thank you very much!
Most recent answer
Recommend you to read the review paper which published in 2021
1 Recommendation
Popular answers (1)
Sichuan University
Recommend you to read the review paper "Deep Learning" in nature 2015.
7 Recommendations
All Answers (10)
Sichuan University
Recommend you to read the review paper "Deep Learning" in nature 2015.
7 Recommendations
Treelogic
The main adventage is their accuracy in image recognition problems.
They have some disadventages:
-High computational cost.
- If you don't have a good GPU they are quite slow to train (for complex tasks).
-They use to need a lot of training data.
2 Recommendations
University of Management and Technology (Pakistan)
Conference Paper Twitter Sentiment Analysis with Deep Convolutional Neural Networks
One of the weaknesses was pointed out by Severyn et al. (link above).
It was the dependence of CNNs on the initial parameter tuning (for a good point) to avoid local optima. Thus, a weakness of CNNs is the considerable amount of work they require to initialize according to the problem at hand. This would require some expert knowledge in the domain.
1 Recommendation
Nanyang Technological University , Singapore
hyper-parameter tuning is non-trivial, need big dataset for proper training, still black box, comparatively slow
University of Strasbourg
1. Data requirements leading to overfitting & underfitting
2. Parameter-to-memory requirements
3. Non-expressive learning
4. Non-expressive logics
5. Parameter tuning requirements
6. Computationally expensive
Duhok Polytechnic University
The main advantage of CNN compared to its predecessors is that it automatically detects the important features without any human supervision.
Similar questions and discussions
Related Publications
Die Computergraphik beschäftigt sich neben der Herstellung 'echt' wirkender Bilder neuerdings auch mit sog. nichtrealistischen Darstellungen.
Die gesamte Diplomarbeit ist ein Demonstrationsmodell. Es gelang Algorithmen zu finden, mit denen man auf optischem Wege Unterschiede zwischen einer Werbesendung und dem normalen Fernsehprogramm erkennen konnte.
It is shown that the standard relations of P.A. Smith theory between the Betti numbers of a space and those of its fixed point set can be improved in the presents of Poincaré duality.