Aalto University

About: Aalto University is a education organization based out in Espoo, Finland. It is known for research contribution in the topics: Computer science & Context (language use). The organization has 9969 authors who have published 32648 publications receiving 829626 citations. The organization is also known as: TKK & Aalto-korkeakoulu.

PRADA: Protecting Against DNN Model Stealing Attacks

Abstract: Machine learning (ML) applications are increasingly prevalent. Protecting the confidentiality of ML models becomes paramount for two reasons: (a) a model can be a business advantage to its owner, and (b) an adversary may use a stolen model to find transferable adversarial examples that can evade classification by the original model. Access to the model can be restricted to be only via well-defined prediction APIs. Nevertheless, prediction APIs still provide enough information to allow an adversary to mount model extraction attacks by sending repeated queries via the prediction API. In this paper, we describe new model extraction attacks using novel approaches for generating synthetic queries, and optimizing training hyperparameters. Our attacks outperform state-of-the-art model extraction in terms of transferability of both targeted and non-targeted adversarial examples (up to +29-44 percentage points, pp), and prediction accuracy (up to +46 pp) on two datasets. We provide take-aways on how to perform effective model extraction attacks. We then propose PRADA, the first step towards generic and effective detection of DNN model extraction attacks. It analyzes the distribution of consecutive API queries and raises an alarm when this distribution deviates from benign behavior. We show that PRADA can detect all prior model extraction attacks with no false positives.

Fast and robust hand tracking using detection-guided optimization

Abstract: Markerless tracking of hands and fingers is a promising enabler for human-computer interaction. However, adoption has been limited because of tracking inaccuracies, incomplete coverage of motions, low framerate, complex camera setups, and high computational requirements. In this paper, we present a fast method for accurately tracking rapid and complex articulations of the hand using a single depth camera. Our algorithm uses a novel detectionguided optimization strategy that increases the robustness and speed of pose estimation. In the detection step, a randomized decision forest classifies pixels into parts of the hand. In the optimization step, a novel objective function combines the detected part labels and a Gaussian mixture representation of the depth to estimate a pose that best fits the depth. Our approach needs comparably less computational resources which makes it extremely fast (50 fps without GPU support). The approach also supports varying static, or moving, camera-to-scene arrangements. We show the benefits of our method by evaluating on public datasets and comparing against previous work.

Do Investors Overweight Personal Experience? Evidence from IPO Subscriptions

Abstract: We find a strong positive link between past IPO returns and future subscriptions at the investor level in Finland. Our setting allows us to trace this effect to the returns personally experienced by investors; the effect is not explained by patterns related to the IPO cycle, or wealth effects. This behavior is consistent with reinforcement learning, where personally experienced outcomes are overweighted compared to rational Bayesian learning. The results provide a microfoundation for the argument that investor sentiment drives IPO demand. The paper also contributes to understanding how popular investment styles develop, and has implications for the marketing of financial products.