Google and Facebook, two of the world’s biggest tech giants, have a growing interest in AI and have demonstrated this by acquiring several AI startups in the last two years. The reason for this is that AI can better understand their users and facilitate faster product innovation. In this regard, AI research has increased vastly in the search and advertising space.
AI is still finding ways to be relevant in the health and wellness space. AI guarantees more efficient prognoses and diagnosis of conditions, while finding treatments based on a person’s specific biological profile. AI is key in this area because it can go beyond the scope of the set doctrines and be guided into unchartered territory.
One of the trends towards understanding a person’s unique profile is the advent of digital health or “mhealth”. People have been strapping numerous sensors to their bodies and utilizing smartphone apps to collect data about their lifestyles. Although this has considerably promoted a better sense of health awareness among the end users, no sort of herculean knowledge discovery is happening. Regardless of whether or not an end user is able to even decipher that kind of data, the information decaying because no practical AI is being applied to it.
It’s essential to invent technologies that will integrate data collected by digital-health users to give a solution that factors in aspects from all data gathered over time.
The challenge in using AI towards health is engineering features and knowledge bases from which a computer can learn. . The field of AI uses machine learning involving several different techniques, each being unique in terms of the size of dataset and the features needed to make accurate predictions.
Also, the fundamental knowledge bases and features have to be constructed by AI experts with experience in chemistry, biology, advanced mathematics and computer science to analyze, materialize and construct solutions to complex problems.
Our team is comprised of individuals with backgrounds in chemistry, biology, computer science and advanced mathematics engineers. Over the past two years, our team has been working tirelessly on bridging the gap between understanding the vast amounts of gathered user health data and using it to create new innovative products, such as smart materials.
Our first technology NuSilico was constructed to simulate the time it took for materials to reach certain sites in the body. Our end product used a technique called deep learning that also employed a certain amount of feature engineering. Feature engineering is essential in the healthcare/personal care space to account for the sheer complexity of training datasets. When tested with researchers, the technology proved to be a great complement to experimentation that cost millions of dollars over several months.
Next, we put the mathematical prowess of NuSilico to use by testing its ability to combine several datasets a single person to predict outcomes of their lifestyle and the effects on their skin.
From there, two benefits emerge: first, the ability to provide personalized guidance directly to users to improve their skin. Secondly , there’s a clearer understanding of user populations to create more effective personalized products. On a larger scale, this technology can be applied towards other health datasets.
Quantified Skin is partnering with a manufacturer to release a suite of products built around personalized guidance. Look for the first batch to start rolling out later this year.
If this topic interests you, here’s a deep learning presentation we gave at the SF Artificial Intelligence in Health MeetUp: