Search
Results
-
Learning Functionally Decomposed Hierarchies for Continuous Control Tasks with Path Planning
(2020)We present HiDe, a novel hierarchical reinforcement learning architecture that successfully solves long horizon control tasks and generalizes to unseen test scenarios. Functional decomposition between planning and low-level control is achieved by explicitly separating the state-action spaces across the hierarchy, which allows the integration of task-relevant knowledge per layer. We propose an RL-based planner to efficiently leverage the ...Conference Paper -
Large-Scale Flipped Classroom with a Single Lecturer
(2020)ETH Learning and Teaching Journal ~ ICED 2020 ProceedingsThe flipped classroom is a setting for blended-learning courses which is generally considered to be effective at a small scale involving many teaching assistants. This paper describes a flipped classroom concept for large classes which can be conducted by one single lecturer while keeping the quality of teaching high. A core pillar of this approach comprises well designed threads of single and multiple-choice questions which have the ...Conference Paper -
Integrating transferable skills into an existing curriculum: The example of Geospatial Engineering at ETH Zurich
(2020)ETH Learning and Teaching JournalIn this contribution we present an approach to integrating improved formation and consolidation of transferable skills into the Bachelor’s program Geospatial Engineering at ETH Zurich. We report how the development of competences in argumentation, critical thinking, technical/scientific writing, visualisation, presentation, learning management, teamwork, and project management was supported without changing the structure of the existing ...Conference Paper -
New conceptual approach combining the probabilistic nature of localised rebar corrosion and the load-deformation behaviour
(2020)There is a need for sound engineering models and concepts taking into account the damage mechanisms of chloride-induced corrosion with respect to the load-bearing behaviour of reinforced concrete structures. In this paper, we present a novel conceptual approach combining the physical-electrochemical processes of chloride-induced corrosion initiation/propagation with the mechanical aspects of load-deformation behaviour. A particular focus ...Conference Paper -
Rumble: Data Independence for Large Messy Data Sets
(2020)Proceedings of the VLDB EndowmentThis paper introduces Rumble, a query execution engine for large, heterogeneous, and nested collections of JSON objects built on top of Apache Spark. While data sets of this type are more and more wide-spread, most existing tools are built around a tabular data model, creating an impedance mismatch for both the engine and the query interface. In contrast, Rumble uses JSONiq, a standardized language specifically designed for querying JSON ...Conference Paper -
Cryptographic Group Actions and Applications
(2020)Lecture Notes in Computer Science ~ Advances in Cryptology – ASIACRYPT 2020 26th International Conference on the Theory and Application of Cryptology and Information Security, Daejeon, South Korea, December 7–11, 2020, Proceedings, Part IIConference Paper -
Incentivizing stable path selection in future Internet architectures
(2020)Performance evaluationConference Paper -
Structural Balance and Interpersonal Appraisals Dynamics: Beyond All-to-All and Two-Faction Networks
(2020)IFAC-PapersOnLine ~ 3rd IFAC Workshop on Cyber-Physical & Human Systems, CPHS 2020. ProceedingsStructural balance theory describes stable configurations of topologies of signed interpersonal appraisal networks. Various mathematical models have been proposed to explain how initially unbalanced appraisal networks evolve to structural balance. However, the existing models either diverge in finite time, or could get stuck in jammed states, or converge to only non-all-to-all graphs starting from certain sets of initial conditions. It ...Conference Paper -
Meta-Learning via Hypernetworks
(2020)Recent developments in few-shot learning have shown that during fast adaption, gradient-based meta-learners mostly rely on embedding features of powerful pretrained networks. This leads us to research ways to effectively adapt features and utilize the meta-learner's full potential. Here, we demonstrate the effectiveness of hypernetworks in this context. We propose a soft row-sharing hypernetwork architecture and show that training the ...Conference Paper -
Interpretable Models for Granger Causality Using Self-explaining Neural Networks
(2020)Conference Paper