Thirty Years of Machine Learning: The Road to Pareto-Optimal Wireless Networks
read more
Citations
Pattern Recognition and Machine Learning
Reconfigurable Intelligent Surfaces: Principles and Opportunities
Reconfigurable Intelligent Surfaces: Principles and Opportunities
References
Random Forests
ImageNet Classification with Deep Convolutional Neural Networks
Long short-term memory
Very Deep Convolutional Networks for Large-Scale Image Recognition
Related Papers (5)
Human-level control through deep reinforcement learning
Frequently Asked Questions (15)
Q2. What have the authors stated for future works in "Thirty years of machine learning: the road to pareto-optimal wireless networks" ?
Furthermore, the authors have highlighted the development tendency of wireless network techniques and a variety of representative scenarios for future wireless networks as seen in Fig. 5 and Fig. 8. they also have provided a caseby-case description of numerous compelling applications relying on ML algorithms in wireless networks as shown in Table VII, followed by a pair of detailed application examples relying on their recent research results. In comparison with state-of-the-art survey papers seen in Fig. 1, their paper overviews all the four popular kinds of learning schemes and their applications in future wireless networks, which has a full scope of how ML algorithms bear fruits in the past decades in wireless networks.
Q3. What are the metrics of evaluating the proposed network selection schemes?
the network access cost function and the QoE reward were defined as the metrics of evaluating the proposed network selection schemes.
Q4. What distance is used for calculating the similarity between the object x and the training samples?
the authors use the Euclidean distance or the Manhattan distance [202] for calculating the similarity between the object x and the training samples.
Q5. What is the way to estimate the regression coefficient vector?
Given a set of training samples {yn, xn1, xn2, . . . , xnM}, n = 1, 2, . . . , N , the authors are capable of estimating the regression coefficient vector w = [w0, w1, . . . , wM ] with the aid of the maximum likelihood estimation (MLE) method.
Q6. What is the appropriate tool for adapting the network’s structure to the human behavior observed?
By mimicking human intelligence, ML may be deemed to be the most appropriate tool for adapting the network’s structure to the human behavior observed [22], [23].
Q7. What is the main advantage of the deep reinforcement learning in autonomous systems?
Deep reinforcement learning is eminently suitable for supporting the interaction in autonomous systems in terms of a higher level understanding of the visual world, which can be readily applied to a diverse analytically intractable problems in future wireless networks.
Q8. How did the model learn the real-time capability of wireless sensors?
By carefully considering the realistic capability of wireless sensors, the model relied on the time- and frequencylimited sensing snapshots having the duration of 12.8 µs as well as the bandwidth of 10MHz.
Q9. How does the algorithm arrive at the final cluster segmentation result?
given K initial cluster centroid µk, k = 1, . . . ,K, Lloyd’s algorithm arrives at the final cluster segmentation result by alternating between the following two steps,• Step 1: In the iterative round r, assign each sample to a cluster.
Q10. Why is interference in UDNs more severe than in traditional cellular networks?
the interference encountered in UDNs tends to be more severe and of higher volatility than that in traditional cellular networks because of the dense deployment of BSs and APs.
Q11. What was the first proposed method for detecting a multicell multiuser MIMO system?
A semi-blind received signal detection method based on ICA was proposed by Lei et al. [262], which additionally estimated the channel information of a multicell multiuser massive MIMO system.
Q12. What are the advantages of the deep learning in environment in interactive decision making?
Given the intrinsic advantages of the reinforcement learning in environment in interactive decision making, it may play a significant role in the field of control decision [344], [345].
Q13. How did Zhang et al. construct a four-layer DNN for extracting?
in [318], Zhang et al. constructed a four-layer DNN for extracting reliable high level features from massive WiFi data, which was pre-trained by the stacked denoising auto-encoder.
Q14. What is the main reason why Zhao et al. proposed a K-means?
Zhao et al. [249] conceived an efficient K-means clustering algorithm for optical signal detection in the context of burst-mode data transmission.
Q15. What is the description of the proposed CNN based wireless interference identifier?
The proposed CNN based wireless interference identifier was shown to have a higher identification accuracy than the state-of-the-art schemes in the context of low SNRs, such as −5dB, for example.