On July 11, Jonas Christensen (Senior Data and AI Leader) joined me on “What’s the BUZZ?” and discussed how you can establish a data-centric approach to machine learning when working with small datasets. Despite the push for Generative AI, Machine Learning remains a cornerstone for solving business problems, but its success hinges on data quality. That’s why a data-centric approach, team collaboration, and improving your business processes are the keys to success. But where should you start? Here is what we’ve talked about…
Taking a Data-Centric Approach to Machine Learning
Machine Learning has become the backbone of many business solutions. Despite the buzz around Generative AI, traditional Machine Learning remains a core technology for most companies.
Consider the analogy of a chef: no matter how skilled they are or how advanced their kitchen is, they can't create a gourmet meal without high-quality ingredients.
The key to unlocking its full potential lies in adopting a data-centric rather than a model-centric approach:
A model-centric approach focuses on tweaking algorithms and hyperparameters to improve performance. While this can yield results, it often overlooks a more significant factor: data quality.
A data-centric approach emphasizes enhancing data quality for Machine Learning models. By ensuring that data is clean, comprehensive, and accurately represents the problem at hand, businesses can significantly boost the performance of their models.
Thank you for reading The AI MEMO. Feel free to share it with your network.
Level up your AI leadership game with the AI Leadership Handbook
Andreas Welsch uses real-world knowledge and examples from interviews with over 60 leaders and experts in AI to help you both introduce and incorporate AI into your organization, from aligning it with your business strategy to turning new-to-AI employees into passionate multipliers to making sure humans stay at the center of your AI use. After reading this book, you will be able to confidently implement AI in your business, no matter your industry.
For Machine Learning to truly benefit an organization, it must be a team sport. This involves collaboration between data scientists, engineers, business stakeholders, and frontline staff who collect the data.
One of the biggest challenges is ensuring that everyone involved understands the importance of data quality and their role in maintaining it. Those who collect or label data often don't realize how their actions impact downstream processes.
» Most businesses don't have one machine learning use case that are worth fifty million dollars. But many businesses have fifty machine learning use cases that might be worth a million dollars. «
— Jonas Christensen
Additionally, leaders must clarify that proper data is a priority for the organization. This might involve adjusting business processes to ensure that data is captured accurately. When everyone in the organization sees the value of high-quality data, it creates a virtuous cycle: better data leads to better models, delivering more valuable business insights.
Transforming Business Processes for Better Data
Poor data quality is often a symptom of a flawed or outdated process. By improving the source (where data is created), businesses can achieve dual benefits: better data for Machine Learning and more efficient operations.
Take the example of a law firm that needed to classify types of legal cases. Initially, the data collected from client interviews was unstructured and difficult to analyze. By restructuring the interview process and asking specific, quantifiable questions, the firm was able to collect higher-quality data. This improved their Machine Learning models and streamlined the entire intake process.
Summary
While Generative AI continues to dominate the headlines, shifting to a data-centric approach for Machine Learning, fostering team collaboration, and improving business processes can help organizations unlock the true potential of their data. This simultaneously enhances the accuracy and reliability of models and drives broader business improvements. By reframing AI projects to start with data and which data needs to be captured for data science teams to derive meaningful results, AI leaders can increase the success rate of their projects.
Which business processes might need rethinking to enhance data quality?
Join my bi-weekly live stream and podcast for leaders and hands-on practitioners. Each episode features a different guest who shares their AI journey and actionable insights. Learn from your peers how you can lead artificial intelligence, generative AI & automation in business with confidence.
