What is "Machine-Learning" and how does Sparta utilize it?

“Machine-Learning”, “Data Science”, & “Artificial Intelligence” and the Sparta Platform

What is Machine Learning?


Machine Learning, Data Science, and Artificial Intelligence are commonplace terms used in the marketing of products and services - but what do they actually mean and how do they relate to the Sparta Movement Health Platform?

Though these terms see frequent use, they defy precise definitions. Depending on who is talking, in what context, and when, any one of these terms might be used to describe a particular technology use case. The common thread is that all three terms concern the use of computational tools to process and extract meaning from data. The Venn diagram below does a reasonable job of capturing a common (though by no means universal) opinion on how these concepts relate to each other.

AI-ML Venn Diagram

An Internet search will yield a vibrant discussion of these three subject areas, but for our purposes, the following synopsis is deemed sufficient:
  • Artificial Intelligence: the use of computers to emulate human intelligence capabilities
  • Machine Learning: the use of computation to automatically learn from experience captured in data
  • Data Science: a set of tools used to process, describe, and transform data in order to extract meaning

In describing what we do at Sparta, we most frequently use Machine Learning terminology, because the software tools and algorithms we use are most frequently characterized in that way. These tools include the use of neural networks (referred to as Deep Learning) in addition to several others.


Why is ML needed?


Artificial intelligence and machine learning tools have existed for decades, but the use of machine learning has grown exponentially in recent years. This is largely due to a combination of exponential improvements in computer processing capabilities and an exponential increase in the amount of data being generated. We now have more data than ever, but luckily we also have more tools and computing power to analyze that data.

The key reason to utilize machine learning is to leverage computers to process the vast amount of quantities of data that now exist. For example, our force plates collect 1000 data points per second from four different sensors. In a single 2 minute assessment, this would provide 480,000 data points. Now multiply that by 100 individuals performing the assessment and we can quickly see the need to leverage computers and machine learning to analyze the 48 million data points we’ve generated. For a bit of history, one of the first computers was able to perform a task in 30 seconds that previously took 20 hours of human effort... that was in 1945. Technology can be leveraged to more efficiently and more accurately analyze the data we are collecting to provide meaningful information.


How does Sparta use Machine Learning?


The specific terminology matters far less than what we do. In the case of the Sparta Movement Health Platform our Force Plate Machine Learning (FPML) engine has two distinct components:
  • Prediction Models: take sets of features (or metrics) as inputs and transform them into output assessments (e.g. injury risk or performance capabilities) or recommendations (e.g. activity programs or things to avoid)
  • Feature Learning: transforms high-frequency time series data into differentiating features (or metrics) that characterize a person’s movement health at a particular time

Prediction Models

If one of the key components of machine learning is to use incoming data to predict results. It is a prerequisite to have large amounts of data to ensure accurate predictions can be made. A key differentiator of the Sparta Movement Health Platform is the aggregation of input data (Force Plate) and output data (injury, performance, testing, training, exercise) that began in 2008.

The Sparta Movement Health Platform uses extracted features (that we commonly refer to as “metrics” in our application) to build and use Prediction Models. These models are built using supervised and semi-supervised learning techniques that combine feature data with injury, performance, and activity data. The goal of these models is to transform the metrics extracted from scan results into assessments of injury risk and performance and to recommendations that guide actions that users can take to improve their movement health.


Feature Learning


When a user executes a single scan test such as a Balance or Jump Scan, the Sparta Platform collects on the order of 1 million measurements. The goal of Feature Learning is to transform those million measurements into a compact form that captures the essential pattern or signature in that user’s movement.

In the Movement Health context, there are two approaches to extracting features from the time series data:


  • Physics-driven: Use biomechanics concepts to identify specific characteristics of force-time series data that could be useful features, and then use numerical analysis to calculate these features from the time series data.
  • Data-driven: Use machine learning techniques to extract features directly from the 1000 Hz force-time series data through pattern recognition

Sparta uses both approaches in combination. While force-time data is often simplified into different phases (eccentric) or distinct time points (peak), these definitions are based on biomechanical theories that are likely to oversimplify human movement. This simplification is important and helpful but can lead to challenges as well. For example, different force plate data acquisition software solutions define different calculated metrics or phases differently, thus leading to confusing and conflicting research and understanding in the field. Data-driven features could include complexity measures derived from entropy measurements or features found through unsupervised learning techniques.

Other Key Benefits of Machine Learning


Accuracy and Reliability of Data Collection


Another critical use of ML in the context of feature learning is in ensuring the quality of the collected data. Though the scan tests are relatively simple and force plates are generally reliable, things can go wrong in ways that can invalidate the data, e.g.:
User can incorrectly execute a scan test - step off plate prematurely, remove clothing mid-test, fail to follow an instruction
Hardware/setup issues - uneven surfaces where a foot is off the ground, weight calibration issues, potential sensor errors

Sparta uses ML techniques to validate collected data, automatically identify issues that could jeopardize data integrity, and to guide corrective action.

Efficiency, Simplicity, and Practicality


Machine-learning models are utilized consistently in our Product Research and Development stages to ensure that the data collection process is practical and quick, and the visualization and interpretation of data are simple and intuitive. For example, different unsupervised clustering models are utilized to help identify natural groupings of data as well as data redundancies that allow our Movement Health Platform and Customer Success team to surface only the most critical information. The increase in data collection capabilities can be overwhelming for leading to paralysis analysis because of the volume of data that exists. Our modeling allows the ability to surface the minimum set of meaningful data points and references that can provide insight and information and drive intervention or action. Additionally, these models are deployed directly into our software and run in real-time on our devices. This enables there to be zero lag-time from data collection, through data analysis, and into data interpretation.


Continuous Future Improvement


The Sparta approach facilitates continuous improvement and expansion of ML-based capabilities over time as more data is collected. The Sparta Movement Health Platform allows for frequent updates and the addition of models over time. Update of models occurs when the platform determines that new models that incorporate more recent data outperform their predecessors (relevant to both Feature Learning and Prediction models). The addition of models can take several forms, including:
  • New feature/metric discovery
  • Expanded Prediction Model functional granularity:
  • Injury type-specific models
  • Performance characteristic-specific models
  • Activity-type specific recommendations
  • Population/Group-specific models: As sufficient quantities of data is collected for specific groups, the possibility exists to create models specific to that group that may provide additional insights.
  • Longitudinal Prediction Models: As data is collected over time for individuals, the identification of patterns of change over time frames on the order of weeks becomes possible.