Thought leaders in the AI space such as Andrew Ng have been advocating for a shift from model-centric to data-centric AI. The idea behind this campaign is that AI models can be only marginally improved through tweaks in the algorithm but considerable change can only be achieved by using high-quality data. However, what does "high-quality data" mean and how do we go about ensuring the quality, diversity, and consistency of our dataset? In this talk, we will discuss the practice of collecting and annotating data for your computer vision models and making sure the dataset you are using is representative and free of harmful biases.
Iva Gumnishka is the founder and CEO of Humans in the Loop, a professional data collection and annotation company focused on building high-quality datasets for computer vision applications. The company is a social enterprise and its mission is to provide dignified work opportunities to refugees and conflict-affected people through annotation projects. Iva holds a degree in Human Rights from Columbia University and she was named Forbes 30 under 30 in 2018.