Nils Schaetti

• PhD
• Scientific collaborator at University of Applied Sciences and Arts Western Switzerland – Geneva

Reservoir Computing is an attractive paradigm of recurrent neural network architecture, due to the ease of training and existing neuromorphic implementations. Successively applied on speech recognition and time series forecasting, few works have so far studied the behavior of such networks on computer vision tasks. Therefore we decided to investiga...

This paper evaluates a type of recurrent neural networks (RNN) named Echo State Network (ESN) on a NLP task referred as author verification. In this case, the model has to identify whether or not a given author has written a specific text. We evaluate these models on a difficult task where the goal is to detect the author in a noisy text stream bei...

Reservoir Computing is a paradigm of recurrent neural network (RNN) models, attractive because of its ease of training and new neuromorphic optoelectronic implementations. Applied with success to time series prediction and speech recognition, few works have so far studied the behavior of these networks on natural language processing (NLP) tasks. Th...

In the last few years, a machine learning field named Deep-Learning (DL) has improved the results of several challenging tasks mainly in the field of computer vision. Deep architectures such as Convolutional Neural Networks (CNN) have been shown as very powerful for computer vision tasks. For those related to language and timeseries the state of th...

Automatic health monitoring and activity recognition systems provide specific information for caregivers and health professionals to prevent injury or disease. With the improvement of sensor technologies, wireless communication and machine learning, systems can now be aware of changes in the user’s state and its environment in order to provide acti...

Contemporary research into recurrent neural networks (RNNs) focus on deep architectures which can discover long-range dependencies in textual data. However , whether this property can help in authorship attribution tasks remains an open question and is dependent on the dataset size, which is intrinsically limited in this field. As a result, this pa...

Dans le domaine du traitement du langage naturel (NLP), l’attribution d'auteur est une tâche bien connue dont le but est de répondre à la question: quel est le véritable auteur d’un document?, fondé sur des modèles et des marqueurs linguistiques. Étant donné un ensemble d'auteurs candidats et un corpus d’échantillons de documents, il s’agit de trou...

This paper describes and evaluates a mixing model for multimodal author profiling using character-based Convolutional Neural Networks (CNN) for tweet classification and ResNet18 for images. We applied theses models to the author profiling task of the PAN18 challenge and show that its architecture allows these model to be applied to any language. Fo...

This paper describes and evaluates a model for cross-domain authorship attribution using Bidirectional Echo State Network-based (ESN) Reservoir Computing. We applied this model to the cross-domain authorship attribution task of the PAN18 challenge and show that it can be applied to this task. This BD-ESN based on a word embedding layer of dimension...

This paper describes and evaluates a model for style change detection using character-based Convolutional Neural Networks (CNN). We applied this model to the style change detection task of the PAN18 challenge and show that its architecture allows this model to be applied to any language. This CNN based on a character-embedding layer, 25 filters and...

This paper describes and evaluates two neural models for gender profiling on the PAN@CLEF 2017 tweet collection. The first model is a character-based Convolutional Neural Network (CNN) and the second an Echo State Network-based (ESN) recurrent neural network with various features. We applied these models to the gender profiling task of the PAN17 ch...

This paper describes and evaluates two neural models for gender profiling on the PAN@CLEF 2017 tweet col- lection. The first model is a character-based Convolutional Neural Network (CNN) and the second an Echo State Network-based (ESN) recurrent neural network with various features. We applied these models to the gender pro- filing task of the PAN1...

We implemented and evaluated a strategy for author profiling using TF-IDF and a Deep-Learning model based on CNN, and applied it to the author profiling task of the PAN17 challenge. We show that it can be applied to different languages (English, Spanish, Portuguese and Arabic)[2]. As features, a simple cleaning method and a matrix of 2-grams of let...

This paper describes and evaluates a strategy for author profiling using TF-IDF and a Deep-Learning model based on Convolutional Neural Networks. We applied this strategy to the author profiling task of the PAN17 challenge and show that it can be applied to different languages (English, Spanish, Portuguese and Arabic). As features, we suggest using...

Le Reservoir Computing est un nouveau paradigme informatique permettant le traitement de l’infor- mation par un système dynamique non-linéaire, dont les réseaux de neurones récurrents (RNN) sont un exemple. Prenant ses bases dans les neurosciences et l’apprentissage automatique, le Reservoir Computing obtient de meilleurs résultats que d’autres app...

Cadaveric temporal bones (TB) have been the traditional starting point for surgical simulation in otolaryngology. However, pediatric cadaveric TBs are a limited and expensive resource. We generated two (one ceramic, one nylon) 3D printed models of a pediatric temporal bone model (including ossicles), generated from CT and MRI images and assessed th...

Professional requests for an alternative to dissection-lead ENT training prompted the creation of an easily accessible anatomically accurate model of the temporal bone (TB). Access to traditional cadaveric models is limited by cost, time consumption, and availability, especially paediatric cadavers. We set out to create and assess an online 3D juve...