How AI And Exercises, Together Give You Better Health

Portal takes working out from home to the next level by leveraging AI and ML to provide Hyper-personalization via fitness coaching and mental wellness.

Vishal Chandapeta CTO and Co founder Portl

Post-Covid the world has realised the importance of fitness in our daily life. From the normalisation of work from home, has also come the natural synergy of workouts from home. Portal takes working out from home to the next level by leveraging AI and ML to provide Hyper-personalization via fitness coaching and mental wellness. In this article, we will talk a little about how AI and fitness work together to help people live more robust, healthier, and longer lives, along with some of the innovations Portl is bringing to this new domain of FitTech.

The Modern form of AI that we commonly see being used in most products nowadays got its early form and start in 2011. Although the academic principles were derived nearly 60 years ago, it took half a century for our hardware capabilities to make these concepts into a scalable reality for the general public. Some of the earliest uses of AI were based on human health, and as such we see a lot of development and interest in AI and Health, and as a natural evolution, we now see the application of AI in fitness and wellness.

There are many ways we can go about using AI in fitness, from a visual approach using computer vision and ML, to using quantitative methods, such as taking a user’s health or medical metrics, along with other parameters, to help increase the effectiveness of curating and creating programs for a user.

The way Portl uses AI/ML based fitness is to take user data, results, inputs, and inferences, a large majority of this automatically, and help curate personalised programs, mental wellness suggestions, along with other forms of programming, to help the user reach their end goal, in the most effective way possible, on a daily basis. The advantage of using fitness-based AI/ML is that the system is able to understand a user better than a user may understand themselves, due to a large number of data points taken, and is then able to pre-determine issues a user may face in their journey to reach their goals, and then provides preemptive solutions, addressing the problem before it even arises. This helps a user have a fun, scientific, and personalised fitness regimen that is made just for them. Now that we know some of the applications of fitness with AI, let’s take a brief look at how this is done.

Computer vision is a method of AI/ML that is able to infer data from an image or video. This is the same methodology that is used to automate driving in smart cars. When we apply this method to fitness we can recognize and correct movements, and gauge the difficulty of an exercise for a user based on their speed, facial expressions, quality of movement, and level of “correctness” for that user’s fitness level, among many other things. Based on these results, the system is then able to automatically tell the user how to correct their form, what their calorie burn rate is, as well as the number of actions done per minute (apm) to help gauge a user’s execution rate. These all come under “Live Feedbacks” i.e they provide us with real-time information to help give instant feedback. This gives the user a feeling of “companionship” or “interaction” during a workout, and doesn’t make working out feel one dimensional, such as watching a video.

On the other hand, we have a purely quantitative form of AI. Here, we take metrics such as Age, Weight, pre-existing health conditions, injuries, blood sugar, body fat % , etc. These metrics can be used to help set a user’s baseline to help a user reach their desired fitness goal. This is the heart of any AI program. How do we select what data is important, what affect, or effect more change in helping a user reach their fitness goal? Figuring out these metrics was one of the initial tasks undertaken by Portl, and to perfect this relationship of “key” values and “junk” values we went about a human augmented approach. Relying on our vast datasets of curated workouts given to people across multiple demographics, and taking regular medical, physical, and performance metrics, along with giving them curated programs by experts in the fields of Strength, Kinesiology, Performance, medical etc, we were able to find out which details are needed for which results. One interesting result we came across was the personalization of programming across people of different demographics was a very real thing, and this might have never been realised if it wasn’t for our need to implement AI/ML.

Another quantitative detail we take into consideration is how a user is feeling, the emotional, or mental wellness of a user. A user’s performance is highly influenced by their mentality on that day. From waking up, to right before a workout. Using feedback forms pre and post workouts, we are able to get the data needed to help select the next workout to help increase both motivation and mental wellness of a person. Some examples of this would be us asking the user, “how are you feeling today, do you feel tired, or worn out?”, or “how did you like the exercise, was it too long or short? Did you feel inspired or motivated by the trainer?”.We are better able to curate and create content that can be used among a wide range of users and situations, while still being personalised based on a user’s emotional and mental levels that day, which would be impossible to do for every user individually, if it were not for AI/ML.

What is the holy grail of fitness, wellness and AI? For Portl , it is to automatically and efficiently give user’s health, wellness, and mental success, with the bare minimum of user input. The better we can intelligently gather health metrics without a user’s manual input, the better we can give them Bespoke programs to address current and upcoming issues, and thus increase their quality of life and lifespan.

Facebook Comments