Beranda / The Deep Learning Remedy for Netflix Award

The Deep Learning Remedy for Netflix Award

https karthkk.wordpress.com 2016 03 22 deep-learning-solution-for-netflix-prize
https karthkk.wordpress.com 2016 03 22 deep-learning-solution-for-netflix-prize

Deep Learning Solution for Netflix Prize

Introduction Typically the Netflix Prize was basically a competition placed by Netflix through 2006 to 2009. The goal regarding the competition was to improve typically the accuracy of Netflix's movie recommendation criteria by 10%. This winning team, BellKor's Pragmatic Chaos, applied a combination associated with collaborative filtering and even machine learning strategies to achieve an improvement of twelve. 06%.

Deep Learning Deep understanding is a subfield of machine studying that uses artificial neural networks with multiple hidden cellular levels to learn compound relationships in information. Deep learning has got been used effectively in a broad variety of programs, including image identification, natural language running, and speech recognition.

Deep Learning Option for Netflix Valuable A deep learning solution intended for the Netflix Valuable was proposed by simply Karthik Raman within 2016. Raman's solution used a strong neural network together with multiple hidden levels to learn typically the relationship between users and movies. Typically the deep neural system was trained on a dataset regarding 100 million movie ratings.

Raman's solution accomplished an improvement of 10. 2% around the baseline protocol, which was typically the best performing criteria at the period. This improvement has been statistically significant, plus it showed the fact that deep learning may possibly be used in order to improve the precision of movie suggestion algorithms.

Implementation Raman's deep learning solution was executed using the TensorFlow deep learning catalogue. The deep neural network was trained on a bunch of GPUs. Typically the training process got several days to complete.

Once the strong neural network has been trained, it has been used to make movie recommendations regarding users. The full neural network was basically able to create more accurate tips than the primary algorithm.

Conclusion Raman's deep understanding solution for the particular Netflix Prize showed that deep studying could be utilized to improve typically the accuracy of film recommendation algorithms. This specific solution was one particular of the first of all successful applications of deep learning in order to a real-world problem.

References