Hubert Soyer

Learn More
Learning to solve complex sequences of tasks—while both leveraging transfer and avoiding catastrophic forgetting—remains a key obstacle to achieving human-level intelligence. The progressive networks approach represents a step forward in this direction: they are immune to forgetting and can leverage prior knowledge via lateral connections to previously(More)
We present a computationally e cient architecture for image super-resolution that achieves state-of-the-art results on images with large spatial extend. Apart from utilizing Convolutional Neural Networks, our approach leverages recent advances in fast approximate inference for sparse coding. We empirically show that upsampling methods work much better on(More)
Learning to navigate in complex environments with dynamic elements is an important milestone in developing AI agents. In this work we formulate the navigation question as a reinforcement learning problem and show that data efficiency and task performance can be dramatically improved by relying on additional auxiliary tasks. In particular we consider jointly(More)
In recent years deep reinforcement learning (RL) systems have attained superhuman performance in a number of challenging task domains. However, a major limitation of such applications is their demand for massive amounts of training data. A critical present objective is thus to develop deep RL methods that can adapt rapidly to new tasks. In the present work(More)
We are increasingly surrounded by artificially intelligent technology that takes decisions and executes actions on our behalf. This creates a pressing need for general means to communicate with, instruct and guide artificial agents, with human language the most compelling means for such communication. To achieve this in a scalable fashion, agents must be(More)
In this work, we present a novel neural network based architecture for inducing compositional crosslingual word representations. Unlike previously proposed methods, our method fulfills the following three criteria; it constrains the wordlevel representations to be compositional, it is capable of leveraging both bilingual and monolingual data, and it is(More)
The pipeline of modern statistical machine translation (SMT) systems consists of several stages, presenting interesting opportunities to tune it towards improved performance on distant language pairs like Japanese and English. We explore modifications to several parts of this pipeline. We include a preordering method in the preprocessing stage, a neural(More)
For the assessment of physical activity, motion sensors have become increasingly important. To assure a high accuracy of the generated sensor data, the measurement error of these devices needs to be determined. Sensor variability has been assessed with various types of mechanical shakers. We conducted a small feasibility study to explore if a programmable(More)
We present an interactive web-based writing assistance system that is based on recent advances in crosslingual compositional distributed semantics. Given queries in Japanese or English, our system can retrieve semantically related sentences from high quality English corpora. By employing crosslingually constrained vector space models to represent phrases,(More)
Learning to navigate in complex environments with dynamic elements is an important milestone in developing AI agents. In this work we formulate the navigation question as a reinforcement learning problem and show that data efficiency and task performance can be dramatically improved by relying on additional auxiliary tasks leveraging multimodal sensory(More)