ERC FUNDED PROJECTS

Even after three decades of research on human-computer interaction (HCI), current general-purpose user interfaces (UI) still lack the ability to attribute mental states to their users, i.e. they fail to understand users' intentions and needs and to anticipate their actions. This drastically restricts their interactive capabilities. ANTICIPATE aims to establish the scientific foundations for a new generation of user interfaces that pro-actively adapt to users' future input actions by monitoring their attention and predicting their interaction intentions - thereby significantly improving the naturalness, efficiency, and user experience of the interactions. Realising this vision of anticipatory human-computer interaction requires groundbreaking advances in everyday sensing of user attention from eye and brain activity. We will further pioneer methods to predict entangled user intentions and forecast interactive behaviour with fine temporal granularity during interactions in everyday stationary and mobile settings. Finally, we will develop fundamental interaction paradigms that enable anticipatory UIs to pro-actively adapt to users' attention and intentions in a mindful way. The new capabilities will be demonstrated in four challenging cases: 1) mobile information retrieval, 2) intelligent notification management, 3) Autism diagnosis and monitoring, and 4) computer-based training. Anticipatory human-computer interaction offers a strong complement to existing UI paradigms that only react to user input post-hoc. If successful, ANTICIPATE will deliver the first important building blocks for implementing Theory of Mind in general-purpose UIs. As such, the project has the potential to drastically improve the billions of interactions we perform with computers every day, to trigger a wide range of follow-up research in HCI as well as adjacent areas within and outside computer science, and to act as a key technical enabler for new applications, e.g. in healthcare and education.

1 499 625 €

Start date: 2019-02-01, End date: 2024-01-31

The ongoing miniaturization in nanotechnology and functional materials puts an ever increasing focus on the development of three-dimensional (3D) nanostructures, such as quantum dot arrays, structured nanowires, or non-trivial topological magnetic textures such as skyrmions, which permit a better performance of logical or memory devices in terms of speed and energy efficiency. To develop and advance such technologies and to improve the understanding of the underlying fundamental solid state physics effects, the nondestructive and quantitative 3D characterization of physical, e.g., electric or magnetic, fields down to atomic resolution is indispensable. Current nanoscale metrology methods only inadequately convey this information, e.g., because they probe surfaces, record projections, or lack resolution. AToM will provide a ground-breaking tomographic methodology for current nanotechnology by mapping electric and magnetic fields as well as crucial properties of the underlying atomic structure in solids, such as the chemical composition, mechanical strain or spin configuration in 3D down to atomic resolution. To achieve that goal, advanced holographic and tomographic setups in the Transmission Electron Microscope (TEM) are combined with novel computational methods, e.g., taking into account the ramifications of electron diffraction. Moreover, fundamental application limits are overcome (A) by extending the holographic principle, requiring coherent electron beams, to quantum state reconstructions applicable to electrons of any (in)coherence; and (B) by adapting a unique in-situ TEM with a very large sample chamber to facilitate holographic field sensing down to very low temperatures (6 K) under application of external, e.g., electric, stimuli. The joint development of AToM in response to current problems of nanotechnology, including the previously mentioned ones, is anticipated to immediately and sustainably advance nanotechnology in its various aspects.

1 499 602 €

Start date: 2017-01-01, End date: 2021-12-31

String is among the most fundamental and commonly used data types in virtually all modern programming languages, especially with the rapidly growing popularity of scripting languages (e.g. JavaScript and Python). Programs written in such languages tend to perform heavy string manipulations, which are complex to reason about and could easily lead to programming mistakes. In some cases, such mistakes could have serious consequences, e.g., in the case of client-side web applications, cross-site scripting (XSS) attacks that could lead to a security breach by a malicious user. The central objective of the proposed project is to develop novel verification algorithms for analysing the correctness (esp. with respect to safety and termination properties) of programs with string variables, and transform them into robust verification tools. To meet this key objective, we will make fundamental breakthroughs on both theoretical and tool implementation challenges. On the theoretical side, we address two important problems: (1) design expressive constraint languages over strings (in combination with other data types like integers) that permit decidability with good complexity, and (2) design generic semi-algorithms for verifying string programs that have strong theoretical performance guarantee. On the implementation side, we will address the challenging problem of designing novel implementation methods that can substantially speed up the basic string analysis procedures in practice. Finally, as a proof of concept, we will apply our technologies to two key application domains: (1) automatic detection of XSS vulnerabilities in web applications, and (2) automatic grading systems for a programming course. The project will not only make fundamental theoretical contributions — potentially solving long-standing open problems in the area — but also yield powerful methods that can be used in various applications.

1 496 687 €

Start date: 2017-11-01, End date: 2022-10-31

Deep neural networks (DNNs) have led to dramatic improvements of the state-of-the-art for many important classification problems, such as object recognition from images or speech recognition from audio data. However, DNNs are also notoriously dependent on the tuning of their hyperparameters. Since their manual tuning is time-consuming and requires expert knowledge, recent years have seen the rise of Bayesian optimization methods for automating this task. While these methods have had substantial successes, their treatment of DNN performance as a black box poses fundamental limitations, allowing manual tuning to be more effective for large and computationally expensive data sets: humans can (1) exploit prior knowledge and extrapolate performance from data subsets, (2) monitor the DNN's internal weight optimization by stochastic gradient descent over time, and (3) reactively change hyperparameters at runtime. We therefore propose to model DNN performance beyond a blackbox level and to use these models to develop for the first time: 1. Next-generation Bayesian optimization methods that exploit data-driven priors to optimize performance orders of magnitude faster than currently possible; 2. Graybox Bayesian optimization methods that have access to -- and exploit -- performance and state information of algorithm runs over time; and 3. Hyperparameter control strategies that learn across different datasets to adapt hyperparameters reactively to the characteristics of any given situation. DNNs play into our project in two ways. First, in all our methods we will use (Bayesian) DNNs to model and exploit the large amounts of performance data we will collect on various datasets. Second, our application goal is to optimize and control DNN hyperparameters far better than human experts and to obtain: 4. Computationally inexpensive auto-tuned deep neural networks, even for large datasets, enabling the widespread use of deep learning by non-experts.

1 495 000 €

Start date: 2017-01-01, End date: 2021-12-31