Chris Insinger talked about why AI is so hot and sensationalised these days but also how misunderstood it still seems to be. Machine learning is not just one thing but rather a broad range of tools, techniques and algorithms that can be applied to a variety of tasks. He demonstrated examples of problems where machine learning is often applied and how decision trees, linear regression, neural networks and other techniques help create insights, recommendations, and powerful visualisations from large and complex data sets.
Bringing machine learning from research into practice is harder than you think: costs can easily spiral out of control, and the quality of data sets often leaves a lot to be desired, hampering the effectiveness of machine learning initiatives. To succeed, it is essential to have a deep understanding of the domain where machine learning is being applied.
As AI plays an increasing role in decision making that impacts human lives, issues of fairness and bias become more important. Even though learning algorithms are unbiased in a technical sense, they’ll learn social and historical biases that are contained within the data sets that they are fed. Applying AI within domains such as healthcare, finance, and surveillance requires strict overview and regulation to avoid introducing or perpetuating unfairness at unprecedented scale.
Chris closed his talk by discussing the impact of AI on our jobs, showing that even though almost all jobs have elements that could be replaced or augmented by AI, the decision to actually do so depends on many factors beyond technical feasibility alone.
Meri Rosich started her talk by referring to some topics that were related to International Women’s Day (8 March): the pay gap between men and women, the longer average working day of women and the lower female labor participation rate in Singapore. Things are improving but progress is still quite slow. She is optimistic though, and believes that big data and machine learning can help us accelerate change.
Big data is the new oil: the world’s most valuable resource today. The most valuable companies these days collect, store, analyze and monetize vast amounts of data about us. These data could be about our companies, our finances, our relationships, our whereabouts, our health, or our voting behavior, collected by connected devices in our homes, our cars, our workplaces – often without us realizing it.
More and more industries are ready for disruption because of the exponential and clever use of big data with the help of AI: manufacturing, transportation, financial services, healthcare. Because of these rapid changes, there’s a strong need for clear data protection regulations. The notion of ‘just trusting’ the black box nature of AI is fortunately becoming increasingly more unacceptable.