We kicked off the first AI in Action Dublin meetup 2020 with a packed-house at the Dublin Chamber of Commerce.
It was the first meetup since our rebranding to AI in Action Dublin but it was still all the usual great speakers from the likes of Accenture and IBM telling us all about their exciting work within Data Science, Machine Learning and AI.
First up, IBM Tech Lead Stefano Braghin presented about his work within Next-Generation Data Privacy Management. Stefano discussed how data privacy must always be considered in the context of a legal/compliance framework and therefore we need to ask if the data privacy processes ready to support the compliance frameworks in time and at scale
Stefano discussed how to solve these problems in detail along with telling us about IBM's data privacy toolkit. Their toolkit allows for Reduced input for supervised risk assessment configuration, a (Semi-)supervised data masking/anonymization configuration generation, Augmented and prioritized reports from risk assessment along with Visual guidance for risk assessment.
Next, Analytics Innovation Analyst Andreea Roxana Miu and Analytics & AI Associate Manager Laura Alvarez Jubete from Accenture discussed their work into measuring algorithmic fairness. They focused on some of the key gaps between academia and industry that exist, as well as presenting innovative solutions to overcome them.
Accenture’s Algorithmic Fairness Tool allows the measurement and correction of potential bias in automated decision-making. Working with structured data and binary classification models, the tool is a user-centric solution built to easily integrate with real data, providing real value to the data science teams by allowing them to explore assess and mitigate for potential bias.
Last but not least, Jack Fitzsimons - Principal Data Scientist at Electroroute - gave a brief history of Statistics and Machine Learning, as well as giving us an insight into his research while studying at the University of Oxford in creating general solutions to machine learning challenges, particularly focusing on fairness in Machine Learning models.
Jack also outlined how Group fairness does not consider the individual merits and may result in choosing the less qualified members of a group, whereas individual fairness assumes a similarity metric of the individuals for the classification task at hand that is generally hard to find. When it comes to regression tasks, we can define fairness by a function evaluation under two generative distributions, one for each subpopulation.
A big thank you to Laura, Andreea, Jack and Stefano for the insightful presentations into their amazing work and thanks to everyone who also came along on the evening. We hope you all gained a lot of information to help you with your AI and Data Science aspirations.
Next month, we celebrate AI in Action Dublin's 2nd year anniversary and we have another fantastic line-up speakers set to take to the stage on Wednesday 4 March. Stay tuned on our meetup page for all the latest details.