Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

Excited to share with Inria Flowers team ADTool, an open-source Python toolkit using curiosity-driven AI to explore behavior diversity of complex systems (cellular automata, physics models, ...)!🐍 Serving off-the-shelf strategies like IMGEP You said IMGEP? 👇

13,752 görüntüleme • 2 yıl önce •via X (Twitter)

10 Yorum

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

Intrinsically Motivated Goal Exploration Processes are a family of diversity search algorithms, first developed to model curiosity-driven exploration in humans (like autotelic exploration, where an agent samples its own goals in an open-ended exploration and discovery loop) 2/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

Schematically during the i'th discovery loop, IMGEP will: -define by itself a new goal g_i -sample p_i trying to have system's response S(p_i) as close as g_i -compute S(p_i) -store (p_i, S(p_i)) to improve future predictions of the parameters required to reach a given goal 3/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

Such algorithms can be highly relevant for exploring parameterized complex systems, where system behaviors are diverse, mapping with input parameters is non-trivial, and interesting parameters over the set of possible parameters are difficult to find by random search. 4/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

In particular IMGEP has recently been shown to be very effective in discovering new behaviors in Biological Networks (e.g. GRNs) 🧬 when exploring possible external perturbations. 5/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

ADTool's modular architecture makes it easy to add your own system, parameter space and observed behavior space, so you can launch your own new discovery processes. 6/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

Explorations can be launched by command line, and discoveries are organized on disk for easy saving, manipulation, sharing and reloading from past discoveries. 7/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

Several examples are provided to illustrate the diversity of possible applications, ranging from cellular automata to physics simulation engines and more to come! 8/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

ADTool also incorporates an interactive visualization aid, so you can see the arrival of discoveries in real time 9/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

Join the associated Discord to exchange ideas, questions, bugs and share your discoveries, with already some original ones on a class of continuous reversible cellular automata :) 10/10

Inria Flowers team profil fotoğrafı
Inria Flowers team2 yıl önce

@BertChakovsky @randazzoettore @eyvindn @zzznah @nataliemaphd @ThomasCabaret84 @mayalen_etc @forestier_seb @jesselinsane

Benzer Videolar

Today is a good day for open science. As part of our continued commitment to the growth and development of an open ecosystem, today at Meta FAIR we’re announcing four new publicly available AI models and additional research artifacts to inspire innovation in the community and help advance AI in a responsible way. More in the video from Joelle Pineau. What we’re releasing: 🦎 Meta Chameleon 7B & 34B language models that support mixed-modal input and text-only outputs. 🪙 Meta Multi-Token Prediction Pretrained Language Models for code completion using Multi-Token Prediction. 🎼 Meta JASCO Generative text-to-music models capable of accepting various conditioning inputs for greater controllability. Paper available today with a pretrained model coming soon. 🗣️ Meta AudioSeal An audio watermarking model that we believe is the first designed specifically for the localized detection of AI-generated speech, available under a commercial license. 📝 Additional RAI artifacts Including research, data and code to measure and improve the representation of geographical and cultural preferences and diversity in AI systems. We believe that access to state-of-the-art AI creates opportunities for everyone – not just a small handful of Big Tech companies. We’re excited to share this work and to see how the community learns, iterates and builds using this technology. Details and access to everything released by FAIR today ➡️

AI at Meta

381,248 görüntüleme • 2 yıl önce