Published Jan 7, 2021

SDS 433: Data Science Trends for 2021 — with Ben Taylor

Join Ben Taylor as he delves into 2021's data science trends, tackling AI ethics and bias, model production challenges, and the future of deep learning frameworks like TensorFlow and PyTorch, while exploring groundbreaking concepts like federated learning and AutoML.
Episode Highlights
Super Data Science: ML & AI Podcast with Jon Krohn logo

Popular Clips

Episode Highlights

  • Federated Learning

    Federated learning is poised to revolutionize data privacy and collaborative AI development by allowing machine learning models to train on decentralized data without accessing individual data points. shares a compelling COVID-related example, highlighting the lack of a national database for patient data in the U.S. due to privacy laws 1. This limitation could be overcome with federated learning, which enables models to learn from data across different networks while maintaining privacy. emphasizes the potential of federated learning to address privacy concerns, especially when data is vast and sensitive 2.

    Federated learning allows you to just come up with a standard, and the machine learning models are actually learning at the individual hospital networks.

    ---

    The approach could significantly improve global data sharing and patient care, provided data standards are established.

       

    Remote Work

    The COVID-19 pandemic has permanently altered remote work practices, with many employees favoring flexible work arrangements. notes that a survey at his company revealed no employees wished to return to full-time office work, highlighting the appeal of remote work's flexibility 3. discusses the challenges of adapting home environments for work, such as noise buffering, and the impact on collaboration, particularly the loss of in-person brainstorming sessions 4.

    I really miss whiteboard collaboration, and I'm sure say that there is a remote work alternative, but I haven't seen the same.

    ---

    Despite these challenges, remote work is expected to continue even after widespread vaccination.

       

    Pandemic Impact

    Reflecting on the pandemic's impact, and discuss the rapid changes in data science and workplace dynamics. shares his experience of the sudden shift to remote work and the challenges it posed, such as missing in-person interactions at conferences 5. The pandemic also accelerated the adoption of remote work technologies, which are likely to remain integral to business operations 6.

    I think a lot of people are missing that. Just the peer to peer interactions you get from data conferences.

    ---

    These reflections underscore the lasting influence of COVID-19 on professional environments and data science trends.

Related Episodes