MEMORY is the second generation of D.O.U.G., the robotic drawing system Chung has been developing since 2015. Where MIMICRY (Generation 1) tracked the artist's gesture in real time, MEMORY introduces a neural network trained on two decades of Chung's archived drawings. The first-generation system watched; MEMORY remembers.

In collaborative performance, the robotic arm draws from this memory bank while Chung draws alongside, producing a two-handed composition in which the machine's mark carries the learned trace of Chung's own hand.

MEMORY was acquired into the permanent collection of the Victoria and Albert Museum in 2022, the first AI model collected by a major cultural institution. It links the 2015 origin of the practice (MIMICRY) to the current body of work (RECURSIONS, 2026).

It is currently on view in Why We Make at V&A East, London, through 2026.

Drawing Operations Unit: Generation 2 involves robotic memory.

It is an initial exploration into the machine learning the drawing style of the artists hand.

The robotic arm’s behaviour is generated from neural nets trained on the artist’s drawing gestures.

In a sense, the robotic arm has learned from the visual style of the artists previous drawings and outputs a machine interpretation during the human / robot drawing duet.


Processed with VSCO with hb1 preset
sougwen-2016_drawingOperationsGeneration2_02-th_1600_c

Drawing Operations Unit: Generation 2 — MEMORY

Gestures from previous drawings are collected and saved, existing as a memory bank for D.O.U.G._2. Analysis of visual style of historic artists to translate into gesture, as well as colour palette as a collective memory bank from which robotic arm will be able to select.

绘图操作单元第2代 (Drawing Operations Unit 2):这一代的机器人涉及机器人的记忆。这是对机器人学习艺术家手绘风格的初步探索。机器人的手臂通过在艺术家绘画手势上的训练而产生神经网络的行为。从某种意义上说,机器人手臂已经从艺术家的图画视觉风格中学习,并在人类/机器人绘画二重奏期间呈现出机器的表达。

 

Sougwen_Chung_Drawing_Operations_Unit_Generation_2_Memory_Collaboration_2019

Research

READING LIST

Avanessian, Armen, et al. Speculative Drawing: 2011-2014. Sternberg, 2014.

~

Galloway, Alexander R. Protocol, or, How Control Exists after Decentralization. 2001.
 


~

Bad New Days: Art, Criticism, Emergency. Verso, 2015.

~

Clark, Andy. Supersizing the Mind: Embodiment, Action, and Cognitive Extension. Oxford University Press, 2011.

CREDITS

Technical Direction
Sougwen Chung

Prototypes:
Mary Franck
Francis Tseng


SUPPORTED BY

Supported By
CG Arts
Bell Labs

NTT InterCommunication Center (ICC), Tokyo

First Generative Art cover for Art in America, 2019
Process documentation of MEMORY Drawing Dataset
Sougwen Chung, Drawing Operations, 2016
On view at WHY WE MAKE, V&A 2026, Sougwen Chung, Drawing Operations, 2016 Pencil on paper, 48.5 x 61 cm
Sougwen Chung, 2016, Memory Study
Sougwen Chung, 2016, Memory Study
Processed with VSCO with hb2 preset
Install Image, 2017 NTT InterCommunication Center (ICC), Tokyo
Install Image, 2017 NTT InterCommunication Center (ICC), Tokyo
Install Image, 2017 NTT InterCommunication Center (ICC), Tokyo
IMG_20191025_171225_209
Winner of the Lumen Art Prize

SOUGWEN CHUNG:
DRAWING OPERATIONS UNIT:
GENERATIONAL TIMELINE

YEAR
TECHNIQUE
THEME
TIME
LOCATION
TECHNOLOGY

Generation 1
MIMICRY ⇢

sougwen-2015_drawingOperations_1

2014 - 2016
Colour Tracking
Simultaneous, Mimicry
Immediate
Site Specific
4 axis Robotic Arm 
Camera
Jitter, 
Max MSP

Generation 2
MEMORY ⇢

Processed with VSCO with hb1 preset

2016 - 2018
Infra Red
Input / Output
 Memory
Site Non-Specific
4 axis Robotic Arm
Camera 
RNN, Python, Javascript, C++

Generation 3
COLLECTIVITY ⇢

Photo by Rain Embuscado

2018 - Ongoing
Optical Flow
Multi-agent
Site Interpretive
New York City
Swarm Robotics 
C++, CAD, Optical Flow, Global / Local 
Positioning

Generation 4
SPECTRALITY ⇢

sougwen-2021_mutations_research_pose

2020 - Ongoing
Form Tracking
Electro-physiological EEGBio Interpretive
Basel / London
6 axis Robotic Arm
Python, Custom Software & CAD, 
Biometric Headset

Generation 5
ASSEMBLY ⇢

Nykyaikaa etsimässä 2022 -näyttelyn ripustus. Kuva: Paula Virta / EMMA

2022-Ongoing
Multi-Form Tracking
EEG
Multi-Positional
London / Helsinki
Multi-Robotic System
20 Year Training Data, Bio-feeback, Spatial Kinematics 

Generation 6
SPATIALITY ⇢

sougwen-2026_doug_spatiality2-th

2025-Ongoing 
Spatial Gesture 
Motion Capture, VR Controllers
London / Heilbronn
Robotic Simulation
Transformer Networks, Python, MERIDIAN training dataset

Generation 7
RECURSIONS ⇢

recursions

2026 - Ongoing
Gesture, Machine Motion, 頓 (dùn)
EEG alpha-wave 
Hong Kong 
Multi-Robotic System
2/3d gesture, Neural network, EEG threshold Data

© 2026 Sougwen Chung / SCILICET.
𝘖𝘱𝘦𝘳𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘈𝘳𝘵
drawing alongside robotic systems since 2015
For acquisition, commissioning, & press: ⇢office@scilicet.xyz

© 2025 Sougwen Chung / SCILICET. 
embodied collaboration; metamorphic ecologies.
For press and collaborations: ⇢office@scilicet.xyz

© 2025 Sougwen Chung / SCILICET.
embodied collaboration; metamorphic ecologies. For press and collaborations: ⇢office@scilicet.xyz

© 2025 Sougwen Chung / SCILICET.  
embodied collaboration; metamorphic ecologies.
For press and collaborations: ⇢office@scilicet.xyz

© 2025 Sougwen Chung / SCILICET.  
embodied collaboration;
metamorphic ecologies.

For press and collaborations: ⇢office@scilicet.xyz

Newsletter Sign-up