By fito_segrera    http://fii.to 

This workshop is an exploration into the visual patterns and sounds which emerge from abstract systems programmed to learn simple tasks; sculptures which make use of reinforcement learning techniques from artificial intelligence. When many layers of simple learning tasks run simultaneously, the process of trial and error and accumulation of knowledge by each agent in the system reveals visual compositions and auditory textures into which the spectator is immersed. The emergence of this new intelligence is then, in real-time, augmented onto an aesthetic experience, the aesthetics of learning machines. 

Artists using machine learning and artificial neural networks in their practice are mainly focused on pixel-based algorithms, hence the most common product of machine learning based art is purely digital and screen-based. Due to the complexity of most deep learning techniques, this approach tends to enclose the learning process within an “ai-black-box” and focuses purely on the outcomes. In this workshop participants will “open the box” and explore the learning process and the intelligence that emerges from it in different scales as an aesthetic object. 

Participants will learn how to develop a simple reinforcement learning algorithm, commonly used in robotics, and use it to design and simulate a possible sculpture or installation. Although the workshop is not focused on physical computing or robotics, by the end of the course, all participants will demo their developed software with some custom electronic boards developed by the participant. 

Objectives: 

• Exploring the artistic forms which result from emerging behaviors of multiple intelligent agents in an intelligent system.
• Using artificial intelligence in sculptural (physical) artworks.
• Programming a reinforcement learning algorithm of 2 actions and n states, specifically Q-learning.
• Using openframeworks in order to generate a piece of software which works as both, a simulation and a controller of the desired sculpture or system. The simulation will behave based on the actual machine learning algorithm and accurately depict the possible outcomes of the desired outworks.
• Demo a prototype using the developed software in openframeworks, and some custom electronics provided by the facilitator. 

Programme: 

Day 1: Friday 6 July, 6pm – 8pm (M3090)

Introduction to artificial intelligence, machine learning and more specifically reinforcement learning in robotics and software applications. Introduction to some reference artworks and artists using artificial intelligence in various forms. Introduction of the facilitators ongoing research project: “The Aesthetics of Learning Machines”. 

Day 2: Saturday 7 July, 10am – 5pm (M3090)

Programming a Q-Learning application (2 actions and N states) in openframeworks. Conceptualizing and brainstorming forms of applying this technique in a sculptural (or installation) artwork. 

Day 3: Sunday 8 July, 10am – 5pm (M3090)

Applying the programmed algorithm into a simulation of the desired artwork. Demo using custom electronic boards. 

Facilitator


 

fito_segrera is an artist, technologist and Head of Research/Creation at Chronus Art Center, Shanghai. He studied fine arts and audiovisual / Multimedia production at Jorge Tadeo Lozano University of Bogota, Colombia and completed a MFA in Design and Technology with honors at Parsons, The New School, New York, while being a Fulbright Scholar from 2013 until 2015. His current research and creative practice appropriates elements from digital philosophy, artificial intelligence, monism and modern physics while using physical computing, software programming and information/telecommunication technologies to inquiry in fundamental ontological questions regarding the nature of reality and the physicality of the universe. Some of his most recent shows include "I am here to lear: On Machinic Interpretations of the World", Frankfurter Kunstverein, Frankfurt, Germany, 2018, "UNREAL: The Algorithmic Present", House of Electronic Arts (HEK), Basel, Switzerland, 2017 & Chronus Art Center, Shanghai, 2017, "VARIATIONS" Paris, France, 2016, "PLUNC Festival" Lisbon, Portugal, 2016, "Migrating Fronteers", Chronus Art Center, Shanghai, China, 2016/2015, "Shanghai Art and Design Fair", West Bund Art Center, Shanghai, China, 2016, "SIGGRAPH", Vancouver, Canada, 2014, "Collision 20 & 21", Boston Ciberarts Gallery, Boston, U.S.A, 2014, "Huston International Performance Biennale", Huston, U.S.A, 2014, "SXSW", Austin, U.S.A, 2014, "Emoji", EYEBEAM, New York, U.S.A, 2013, "Dorkbot", New York, U.S.A, 2013, Harvestworks, New York, 2013. 

Requirements: 

The technical knowledge that the workshop offers will be all given in openframeworks (c++). If the participant has never used openframeworks but has experience with processing (JAVA) or any other programming language at the level of arrays, vectors and objects, that is also fine.

Participants with no programming experience are welcome to participate. It is important to say that the workshop is 50% technical and 50% conceptual / artistic, it involves thought processes and creative thinking, nonetheless the workshop will give priority to artist/students with programming experience on the level mentioned above. 

A laptop (any OS) with openframeworks installed and tested is required: https://openframeworks.cc/

Enrollment: 

Day one is a lecture (max. attendance: 100, first come first serve). Days 2 and 3 are hands-on so a maximum of 12 participants is allowed. 
 

Admission:

Free 

Registration: 

Please email Damien Charrieras at: dcharrie@cityu.edu.hk

This event is supported by the Public Policy Research Funding Scheme from Policy Innovation and Co-ordination Office of the Hong Kong Special Administrative Region Government.