current version: v1.0
Description
Synthetic Memories is an interactive installation that explores the notion of return through the construction of collective, artificial memories.
At the core of the work lies the oral transmission of memory and the processes of distortion and re-signification of lived experience.
The viewer is invited to select a photograph and place it on an RFID-enabled surface, the image album.
This action activates a projection that presents the algorithmically generated synthetic image, the archival photographs that constitute its layers, and a computationally generated narrative based on testimonies from displaced communities.
A total of 1,128 synthetic images were produced using a procedural algorithm, employing multi-layered visual data composition and controlled randomness techniques.
The accompanying narratives are generated by an artificial intelligence language model trained on oral refugee testimony.
The work functions as an artificial system of memory retrieval, a temporally dislocated environment of return, not to a physical place, but to the trace of experience.
Ultimately, it raises the question:
How do refugee narratives and experiences transmitted across generations shape collective identities, behaviors, and psychological trauma?
Human behavior and the perception of reality are shaped by multiple interrelated factors, with geography often playing a decisive role, as a site of loss, displacement, or potential return.
Methodology
The installation’s technical framework combines image synthesis and computational storytelling to simulate a layered process of memory retrieval.
Two systems were developed in parallel: one for visual generation and one for narrative construction.
Procedural Image Generation
A total of 1,128 synthetic images were generated using a custom-designed procedural algorithm.
This system included:
- Combinatorial layering of diverse visual data, including backgrounds, textures, and archival refugee photographs,
- Algorithmic selection and placement of digital elements, employing stochastic distribution to ensure diversity,
- Collision detection mechanisms to prevent overlap,
- Metadata generation accompanying each image, enhancing both documentation and research value.
Each synthetic image reflects a deliberate interplay between artificial construction and archival grounding, functioning as a computational artifact of memory transmission.
Computational Narrative Generation
Parallel to the image system, a Creative Narrative Algorithm of Collective Memory was developed.
This language model–based system was trained on oral testimonies from displaced communities and structured archival sources.
Its methodology included:
- Historical metadata input, drawn from authentic refugee photography and associated records;
- A carefully constructed linguistic prompt that frames the storytelling within a collective, intergenerational, and emotionally neutral register;
- A large language model enriched with oral testimonies, narratives, and a neutral framing, devoid of references to specific nation-states, individuals, or political events.
The resulting synthetic images and their computational narratives serve as visual metaphors for the processes of memory transmission, transformation, and re-signification.
Conceptual Framework
The work functions as an artificial system of memory retrieval, a temporally dislocated environment of return, not to a physical place, but to the trace of experience.
By integrating algorithmically composed images and narratives derived from real testimonies, Synthetic Memories engages with the poetics of loss, transformation, and cultural memory.
It suggests that memory is not a fixed repository of facts, but a dynamic, recombinatory process, subject to fragmentation, mediation, and reinvention.
Rather than replicating the past, the installation opens a space in which memory is encountered as a speculative construction, layered, plural, and emotionally resonant.
The installation thus functions as a simulated memory retrieval mechanism, engaging viewers in an active return.
It offers a temporally dislocated reflection on foundational experiences, not as a direct recollection, but as a computationally synthesized, multilayered memory.
Mediums
Title: Synthetic Memories
Year of Creation: 2025
Technologies: RFID technology, algorithmically generated images, computational narrative, microcontrollers, electronic components & sensors, programming in React.js & JavaScript, TouchDesigner, 3D printing, archival material.
*This Installation developed under Inspire Project 2025 at MOMus Museum of Contemporary Art.
References
The following references provided a foundational basis for the conceptual development of this work.
- Varghese, A., & K., D. (2024). A Recreation Mechanism for Refugee Memories to Get through the Betrayal and Trauma in The Hungry Tides by Amitav Ghosh. World Journal of English Language, 14(1), 197-205. https://doi.org/10.5430/wjel.v14n1p197
- Kühnapfel, C., Fingerhut, J., & Pelowski, M. (2023). The role of the body in the experience of installation art: A case study of visitors’ bodily, emotional, and transformative experiences in Tomás Saraceno’s “in orbit”. Frontiers in Psychology, 14, 1192689. https://doi.org/10.3389/fpsyg.2023.1192689
- Chen, X., & Ibrahim, Z. A. (2023). A Comprehensive Study of Emotional Responses in AI-Enhanced Interactive Installation Art. Sustainability, 15, 15830. https://doi.org/10.3390/su152215830
- Zhou, E., & Lee, D. (2024). Generative artificial intelligence, human creativity, and art. PNAS Nexus, 3(3), pgae052. https://doi.org/10.1093/pnasnexus/pgae052
