NOKIA Science and Spritz Vol. 11 ft. summer interns
Summer is in full swing and, shockingly, that means another Science & Spritz! Join us at 4pm for Spritz (the best in Cambridge), soft drinks (better than Uni’s), and Neapolitan pizza (the best in East Anglia). Our interns would love to tell you more about what they are up, and they will do so in the form of Pecha Kuchas (20 slides, 20 seconds each).
Social Psychology at Scale, Rappaz Jérémie. We will discuss the challenges that revolve around building an online portal to conduct social psychology experiments at scale. We will also introduce the first of such experiments — one that aims at capturing the true nature of a social tie.
Psychological Well being Tests, Eva Sharma. To answer important questions about a person’s well being, we have sketched a few tests that are able to determine one’s aspirations, self-esteem and life satisfaction. Test results are then related to Instragram use.
Human and Machine Judgment — The Case of Lie Detection, Sebastian Deri. As computers come to play an increasing role in our everyday lives, human and machine judgments are becoming increasingly intertwined. We report the result of a task in which both humans and machines made over 7,000 judgments about deceitful statements. We found the highest accuracy when human machine judgments were combined — but only when machine “made the final decision,” so the speak.
Social Vitality in Cities, Danaja Maldeniya. We use geo-located social media and map data to characterize cities in terms of human activity and urban design. We then relate spatial variations in these dimensions to the social fabric of cities. We will explore this relationship to test for existing sociological theories and to develop metrics for the social health of neighborhoods.
Explaining and Visualising Urban Emotions: A deep learning approach, Sagar Joglekar. Deep learning systems are black boxes, yet they have performed fairly well in learning abstract/intangible properties like beauty, aesthetics and vitality in urban scenes. This project’s goal is to exploit the upcoming methods that visualise the internals of a deep-learning black box and, in so doing, to visualise, reason about, and explain emotions in urban scenes.