Design génératif

vignette|Le processus itératif de création par design génératif vignette|Samba, un mobilier créé par Guto Requena par design génératif. Le design génératif est un processus de design itératif qui implique un programme qui va générer un certain nombre de sorties qui répondent à certaines contraintes, et un designer qui affine un espace de recherche par le changement des valeurs minimales et maximales d'un intervalle dans lequel une variable du programme répond à l'ensemble des contraintes, dans le but de réduire ou d'augmenter le nombre de sorties à partir desquelles choisir. Le programme n'a pas besoin d'être exécuté sur une machine comme un ordinateur numérique, il peut être exécuté par un humain, par exemple avec un stylo et du papier. Le concepteur n'a pas besoin d'être un humain, il peut être un programme de test dans un environnement de test ou une intelligence artificielle (voir, par exemple, Réseaux antagonistes génératifs). Le designer apprend à affiner le programme et à chaque

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Publications associées

Chargement

Personnes associées

Chargement

Unités associées

Chargement

Concepts associés

Chargement

Cours associés

Chargement

Séances de cours associées

Chargement

Résumé

Source officielle

À propos de ce résultat

Publications associées (11)

Publications associées

Chargement

Personnes associées

Chargement

Unités associées

Chargement

Concepts associés

Chargement

Cours associés (7)

Cours associés

Chargement

Séances de cours associées

Chargement

Particle Source Localization with a Low-Cost Robotic Sensor System: Algorithmic Design and Performance Evaluation

Xinxing Chen, Jian Huang, Ali Marjovi, Alcherio Martinoli

In this paper we tackle the problem of finding the source of particulate matter with a mobile robot equipped with a low-cost multi-channel optical particle counting sensor. The proposed method is based on the Infotaxis odor source localization algorithm and makes multiple modifications to adapt it to particle plumes. In particular, we propose three simple but efficient ways to fuse multiple probability maps associated with various particle sizes and use the resulted integrated map to guide the robot’s movements. A refined measurement data collection is conducted in a wind tunnel to fit the particle plume model. The method with three proposed integration strategies is evaluated in simulation and in the wind tunnel emulating realistic environmental conditions in a repeatable fashion. In particular, we have investigated the impact of two environmental parameters - the wind speed and source release rate on the algorithm performance. The proposed algorithm with the weighted multi-modality map integration strategy outperforms the original Infotaxis and the other two variants. In high wind speed, the proposed algorithm is able on average to estimate the source location with less than 1 meter error in the 80 m2 wind tunnel arena.

2020

Chargement

Probabilistic models for music

Jean-François Paiement

This thesis proposes to analyse symbolic musical data under a statistical viewpoint, using state-of-the-art machine learning techniques. Our main argument is to show that it is possible to design generative models that are able to predict and to generate music given arbitrary contexts in a genre similar to a training corpus, using a minimal amount of data. For instance, a carefully designed generative model could guess what would be a good accompaniment for a given melody. Conversely, we propose generative models in this thesis that can be sampled to generate realistic melodies given harmonic context. Most computer music research has been devoted so far to the direct modeling of audio data. However, most of the music models today do not consider the musical structure at all. We argue that reliable symbolic music models such a the ones presented in this thesis could dramatically improve the performance of audio algorithms applied in more general contexts. Hence, our main contributions in this thesis are three-fold: We have shown empirically that long term dependencies are present in music data and we provide quantitative measures of such dependencies; We have shown empirically that using domain knowledge allows to capture long term dependencies in music signal better than with standard statistical models for temporal data. We describe many probabilistic models aimed to capture various aspects of symbolic polyphonic music. Such models can be used for music prediction. Moreover, these models can be sampled to generate realistic music sequences; We designed various representations for music that could be used as observations by the proposed probabilistic models.

EPFL2008

Chargement

Probabilistic models for music

Jean-François Paiement

École Polytechnique Fédérale de Lausanne2008

Chargement

AR-401(y): Studio MA1 (Huang)

The studio examines the effects of artificial intelligence and automation on architecture and cities. We explore meaningful form generating processes by the use of algorithmic and parametric tools and introduce the notion of growth typologies in architectural and urban design thinking.

ChE-320: Bioreactor modeling and simulation

The course of Bioreactor modeling and simulation focuses on the principles of algorithmic design and analysis of biochemical reactors. The application of these designed reactors would be in the production line of the of pharmaceutical, biotech and chemical industries.

AR-301(h): Studio BA5 (Weinand)

Le studio Weinand développe une approche polytechnique du projet. Via les différentes présentations et ateliers, les étudiants sont encouragés à repousser les frontières de la discipline pour développer une vision holistique de leur projet et questionner la notion de performance en architecture.