Why Should We Care About TinyML and AutoML?

Today, around 12% of the world’s land is utilized for farming. Yet it is believed an increase of 70% more food will be required, by 2050, to support the growing global population. Farmers continue to face challenges such as:

  • Changes in climate producing more soil erosion, heat, and floods
  • Constantly changing global economy
  • Consumer demands for higher quality foods year around

To achieve these goals, farmers would require a nearly real time understanding of the changes in weather conditions, knowing when is the right time to harvest, and how to produce a higher yield of crops.

Monitoring air, temperature, and barometric pressure are examples of sensor data.  A Sensor is a physical device that collects information about the surrounding environment, converts into data, transmits that data to the cloud, and is interpreted by a machine. Multiple studies have shown at least 50% of this data was lost due to limitations in bandwidth and/or power constraints.

TinyML and AutoML to the Rescue!

Tiny Machine Language (TinyML) is compact and enables the value of machine learning to be applied directly on the hardware of the physical device. This means the interpretation of the data (TinyML) and the automation in predictive analytics (AutoML) would no longer have a dependency on the cloud or other external devices.

In a recent conversation with the  Blair Newman,  Chief Technology Officer of Neuton.ai, the team at Neuton.AI used TinyML to enable farmers to make better informed decisions. Creating value-based outcomes by not only optimizing crop planning but supporting improved resource utilization and crop choices.  Where little or no historical data is available, automated machine learning (AutoML) allowed farmers to develop more accurate models based on small training datasets.

Let’s not stop there. The use cases for TinyML and AutoML can expand across all industry sectors and business challenges. For companies who are just now exploring AI in their organization, there is a desire to see the results more quickly. A few examples include:

  • Marketing – Understanding which customers are at risk of leaving
  • Finance – Connecting customers to the right financial service or product
  • Sentiment Analysis – Reveal trends using aggregated data from text reviews and social media source
  • Procurement – identify opportunities for cost-reduction or detect anomalies in supplier performance

AutoML is a great way to jump start your predictive analytics with smaller training data sets, while the organization is developing the long-term infrastructure to collect and store comprehensive data sets. It enables a consistent and quicker method to developing models, rather than solely relying on data scientist to manually code. Companies like Neuton.AI enable data scientists and AI developers by providing another tool in their toolkit.

Creating a more efficient process and automating where possible, it the underlying goal. Fostering the organization to make better-informed decisions with the right data, at the right time when that decision will have the most impact, has always been considered the holy grail.

Key Take Aways

  • Minimize human-based errors
  • Reduce model development to production cycle times
  • Increase return on investment by achieving value more quickly
  • Improve consistency and standardization
  • Minimizes the skills gap

Want to know more. Click below for the video interview with Blair Newman. You can also hear this interview on your favorite audio podcast from AI Time Journal Podcasts on Apple or Spotify.

About The Author

Scroll to Top
Share via
Copy link
Powered by Social Snap