The challenge of dense and scalable wireless systems

For many decades wireless designers have focused on range and link performance. Communications systems were assessed in terms of their range and data-carrying capacity. Early cellular system designers and architects were concerned that the towers and base stations were assured to have sufficient range to provide complete coverage. However, the pervasiveness of wireless communications has almost reversed this consideration, and the emphasis in the development of wireless architectures must adapt itself to address fundamentally new technology objectives, since the success of future networks will be driven by network density and scalability.

While radio engineers once strove to achieve as much range as possible from radio systems, in an emerging wireless-dependent society, the need for high device and bandwidth density has begun to drive wireless communications in a reverse direction; stressing spectrum reuse through short-range, highly localized communication architectures. Some of these adaptions are apparent now: low-cost femtocells fill a service gap between taller and higher-power cellular base transceiver station (BTS) and towers, and the lower-power wireless local area network (LAN) installations. Others, such as flexible spectrum sharing, interference management, and heterogeneous networks are only emerging.

Future communications architectures are unlikely to be a choice among current cellular, wireless local area network (WLAN), peer-to-peer (P2P), and fixed modalities, but will be convergent to an integrated framework whose optimizing process will be so dynamic that it will be invisible to users.

AND SO WE HAVE come to the end of our journey through the ‘making sense of data’ landscape. We have seen how machine learning can build models from features for solving tasks involving data. We have seen how models can be predictive or descriptive; learning can be supervised or unsupervised; and models can be logical, geometric, probabilistic or ensembles of such models. Now that I have equipped you with the basic concepts to understand the literature, there is a whole world out there for you to explore. So it is only natural for me to leave you with a few pointers to areas you may want to learn about next.

One thing that we have often assumed in the book is that the data comes in a form suitable for the task at hand. For example, if the task is to label e-mails we conveniently learn a classifier from data in the form of labelled e-mails. For tasks such as class probability estimation I introduced the output space (for the model) as separate from the label space (for the data) because the model outputs (class probability estimates) are not directly observable in the data and have to be reconstructed. An area where the distinction between data and model output is much more pronounced is reinforcement learning. Imagine you want to learn how to be a good chess player. This could be viewed as a classification task, but then you require a teacher to score every move.