How machine learning may change the way the elderly live

Among the different possible applications of machine learning and respective groups of people favored by them, some welcome ones would be for improvement of quality of life of the elderly.

To have insights on how to improve their quality of life, is necessary to listen from the seniors which factors they consider decisive on the way they experience daily life. Papers on this topic ([1],[2]) have shown that the majority of the elderly people evaluate their quality of life on the basis of social contacts (especially with family and children), dependency, health, material circumstances and social comparisons.

The context of health has particularly multiple branches. One of them is rehabilitation and inclusion, since effects of aging combined with possible of events such as strokes, dementia can compromise multiple visual, auditory, motor abilities. For example, here machine learning can be helpful for learning how to compensate electrical stimuli in Parkinson patients, mapping environments for autonomous navigation of wheelchairs.

Still on the improvement of accessibility/inclusion, constant improvements on size, cost of sensors, data-storage and microprocessors allow more and more their use in multiple objects, such as automatic doors, lamps, televisions, air-conditioners, drugs dispensers etc. Machine learning can once again be helpful for learning patterns from multi sensorial inputs and properly activate these kind of actuators.

Monitoring health status of elderly can also benefit a lot from machine learning, e.g. for detecting abnormal patterns on signals as ECG, blood pressure or even their spatiotemporal behavior. As explained in [3], using motion sensors combine with cloud-based software, is possible to learn their daily habits and detect unusual events such as lack of motion during the typical sleep hours or bath hours of an elderly in his/her bedroom or bathroom.

As mentioned in [4] for the case of Toyota, robots to take care of elderly are already being developed. They can play a helpful role not only in monitoring those types of health parameters, but perform a wide range of skills in a delicate way, where machine learning will play an important role in training the robots how to treat elderly people. Returning to the factors listed by elderly as important for quality of live, robots could not only be nurses, but also companions, by learn how to interact with the elderly — e.g. talking, playing games.

One interesting initiative currently taking place is connecting elderly with young people interested in learn foreign languages, using tools as Skype, Hangout etc. Both sides benefit: one side is able to improve its linguistic abilities and learn about experiences of someone who lives this other culture; the other side has the satisfaction of communicating to younger people, feeling useful and integrated to society. While machine learning is not essential in this case, it could be helpful to for example learn patterns that allow putting in contact people who share common interests.

In the year 2026, I believe it is perfectly possible that we will already have some nurse-robots, multiple “intelligent dispositives” in smarthouses and, hopefully, intelligent ways to keep elderlies happily integrated to society.

## SOURCES

[1] Gopalakrishnan Netuveli and David Blane, Quality of life in older ages. Br Med Bull (2008) 85 (1): 113–126 first published online February 15, 2008 doi:10.1093/bmb/ldn003

[2] Xavier, Flávio M F, Ferraz, Marcos P T, Marc, Norton, Escosteguy, Norma U, & Moriguchi, Emílio H. (2003). Elderly people´s definition of quality of life. Revista Brasileira de Psiquiatria, 25(1), 31–39. https://dx.doi.org/10.1590/S1516-44462003000100007

[3] http://cw.com.hk/feature/machine-learning-enables-proactive-home-based-elderly-care

[4] http://www.techinsider.io/toyota-developing-robots-to-care-for-elderly-2016-6