The Elephant in your Dataset: Addressing Bias in Machine Learning

Explore the critical issue of bias in machine learning through this 43-minute conference talk recorded at NDC London. Delve into the various forms of bias that can impact machine learning models, from unconscious societal biases to those introduced in datasets and algorithms. Gain insights into different types of bias within the context of machine learning, including overfitting bias, algorithmic bias, and prejudice bias, illustrated through real-world examples. Examine the potential consequences of unchecked bias, such as discrimination and unfair decision-making. Learn about cutting-edge research and approaches to mitigate bias, including data preprocessing, fair representation learning, and bias correction algorithms. Participate in a hands-on example to understand how to apply these techniques in practice. Discover how addressing bias can lead to more responsible and ethical use of machine learning, ultimately contributing to the development of more trustworthy technology.