Colloquium

These days, as more and more people use online applications such as Wikipedia, Stack Overflow or Facebook, social phenomena that originally appeared in the offline world make an appearance online and new social phenomena emerge. This calls for narrowing the gap between computer science and social sciences in general, and sociology in particular. The opportunity here is twofold. First, the algorithmic approach can offer a new perspective on social phenomena previously studied in social sciences.

In distributed storage systems (DSS) data is stored over a large number of storage nodes in such a way that a user can always retrieve the stored data, even if some storage nodes fail. To achieve such resilience against node failures, DSS introduce data redundancy based on different coding techniques. When a single node fails, the system performs node repair, i.e., reconstructs the data stored in the failed node in order to maintain the required level of redundancy.

In recent years, geometric data is gaining increasing interest both in the academia and industry. In computer graphics and vision, this interest is owed to the rapid development of 3D acquisition and printing technologies, as well as the explosive growth of publicly-available 3D shape repositories. In machine learning, there is a gradual understanding that geometric structure plays an important role in high-dimensional complicated datasets.

The Media Innovation Lab at IDC is a research and prototyping lab, focused on the interaction between people and technology in domains that merge between the physical and digital worlds. Dr. Oren Zuckerman will present the "Objects for Change" research projects at the media innovation lab, including research opportunities for CS students.

We present a data-driven method for estimating the 3D shapes of faces viewed in single, unconstrained photos. Our method was designed with an emphasis on robustness and efficiency—with the explicit goal of deployment in real-world applications which reconstruct and display faces in 3D. Our key observation is that for many practical applications, warping the shape of a reference face to match the appearance of a query is enough to produce realistic impressions of the query’s 3D shape.

Curiosity is one of the major human drives. Can we model curiosity in biological agents? Can we implement them in artificial systems? What happens when a curious child meets a curious robot? In this talk I present recent work on the study of curiosity. First, studies of curiosity-driven behaviors in humans and rodents are presented, where we show that biological agents attempt to manage their novelty in a structured manner. A model that captures this structure is presented, wherein emergent exploration behaviors are balanced with novelty-based withdrawal-like actions.

Cold boot attacks are a class of attacks wherein memory remanence eff?ects are exploited to extract data from a computer's memory. The idea is that modern computer memories retain data for periods of time after power is removed, so an attacker with physical access to a machine may be able to recover, for example, cryptographic key information. The time during which data is retained can be increased by cooling the memory chips. However, because the memory gradually degrades over time once power is removed, only a noisy version of the data may be recoverable.

This talk covers positioning & navigation challenges. We first give an introduction on GNSS (e.g. GPS) and focus on cases where GPS navigation is not sufficient (i.e., Indoor, or urban canyons). Then a short survey on “smartphone-sensors” will be given and few sub-problems on sensor fusion will be presented. The main part of the talk will cover a new geometric (bio-inspired) framework for navigation and mapping using off-the-shelf mobile devices followed by few positioning simulation and an actual demonstration.

The nearest-neighbor search (NN) is a classic problem in computer science with vast applications in computer vision. Given a database of points in ${\mathbb R}^d$ and a query point, NN search tries to obtain the data point closest to the query point under some metric distance. However, in many settings such searches are known to suffer from the notorious curse of dimensionality, where running time grows exponentially with $d$ (in optimized solutions). This causes severe performance degradation when working in high-dimensional spaces.

In distributed storage systems (DSS) data is stored over a network of nodes in such a way that a user can retrieve the stored data even if some nodes fail. To achieve such resilience against node failures, DSS introduce data redundancy based on different coding techniques. For example, erasure codes are widely used in such systems. When a single node fails, the system reconstructs the data stored in the failed node to keep the required level of redundancy. This process of data reconstruction for a failed node is called a node repair process.