Projects

Perceiving Art: Physics Principles & Research Challenges

A workshop funded by NSF

16-18 October, 2023, Institut Henri Poincaré, Paris

Organizers:

Beata Bajno, (artist, architect)
Sylvie Benzoni, Director, Institut Henri Poincaré, Paris
Andrzej Herczyński, Boston College
Joanna Dreszer, Nicolaus Copernicus University, Toruń
Jacek Rogala, University of Warsaw (Principal Organizer)

The workshop will be held on October 16-18, 2023, at the Institut Henri Poincaré (IHP) in Paris. It is a part of a wider program funded by the National Science Foundation (NSF) on Exploring the Intersection of Science and Art and is supported by the NSF and the IHP.

Founded in 1928, the Institut Henri Poincaré is an international research center for mathematical sciences and their applications and a part of Sorbonne Université and the CNRS. The primary mission of the Institute is to promote and host scientific exchanges and to organize international research programs in mathematics, physics, and related fields. For more information please see: https://www.ihp.fr/en .

The aim of the workshop is to explore how viewers experience visual art by integrating a variety of different research approaches and techniques, such as EEG, fMRI, eye tracking, psychological and behavioral studies, and explainable neural networks. While a number of researchers work on these subjects individually, the workshop will hopefully help start a more cohesive community. The interest, from the scientific point of view, is to identify general principles of the perception of art and promising new approaches. We anticipate that a multidisciplinary character of the workshop, bringing together physicists, mathematicians, biologists, neuroscientists, computer scientists, cognitive scientists, and psychiatrists—as well as practicing artists—will prove conducive to a productive exchange of ideas, and will stimulate new research directions. 

I. Goals

The aim of this workshop is to explore how viewers experience visual art by integrating a variety of different research approaches and techniques, such as EEG, fMRI, eye tracking, psychological and behavioral questionnaires, and explainable neural networks.

To date, research on perception of images has been highly compartmentalized, with experts in various fields working largely independently, and with scant cross-pollination of ideas and results across disciplines. An open-ended discussion of different approaches, and a vigorous debate will generate a more complete understanding and new ideas.

Indeed, new insights may emerge from combining different research techniques. For example EEG provides a high temporal but low spatial resolution, whereas fMRI a high spatial but low temporal resolution. The two methods are complementary and used together would enable a more comprehensive model of image perception.

It is hoped that a multidisciplinary character of the workshop, bringing together physicists, mathematicians, computer scientists, and biologists, will prove conducive to a productive exchange of ideas, and will stimulate new research directions. It may also result in new quantitative methods for advanced artificial neural networks, and for topological data analysis suitable to problems outlined below.

II. Suggested Topics

The workshop will endeavor to address a broad range of problems and methods related to the reception of art. Examples of topics and questions, which outline the scope of the workshop, follow below. Items 1-5 reflect some of the currently active research areas.

1. Intentionality in art

Natural forms, however beautiful, are generally not interpreted as art. Conversely, art rarely aims to simply offer pleasing or naturally looking forms. Is the presence of artistic intent during the creative process a necessary attribute of artworks, or can an object become art a posteriori, for example by a suggestive placement or a frame?

2. Coding-decoding of meaning

Visual art serves as a mode of communicating meaning, from the artist to the viewer. The content conveyed (or the message) must therefore be encoded in the artwork in such a way that, at least in principle, it can be accessible and understandable – that is decoded.

Perception of an artwork may evolve while decoding the message intended by the author (Leder et al., 2004). The decoding may occur in a similar way as for the information transmitted from a sender to a receiver via some channel, as described in the information theory (Shannon 1948). If so, art appreciation would depend on how easily the encoded meaning can be abstracted and decoded. Indeed, these processes could hardly be “error- free.” What role then do personal experiences, predilections, expectations, and the state of mind play in perceiving art? Can repeated viewing, or learning more about the artist, lead to a more “accurate” perception?

3. Figuration versus abstraction

Are there fundamental differences between perceiving figurative art, such as portraits, landscapes, or still lives, and abstract art? Is reception of abstract art inherently more challenging or prone to misinterpretation? Or is the experience of the viewer, and general familiarity with abstract art, a major or the deciding factor?

4. Art and Artificial Intelligence

Artificial neural networks have recently been engaged to model artistic processes and to study the reception of art. This line of research raises many questions and opens perplexing dilemmas.

Can intention, or artistic purpose, be ascribed to generative neural networks when they are trained and then manipulated specifically to create esthetically pleasing images, or to imitate distinctive styles developed by various painters? Are such explorations capable of suggesting future directions or styles in art? Can they shed light on the neural activities in the brain, which are responsible for our evolving understanding of the abstract idiom and our discernment in confronting modern art?

5. Perception and image complexity

What role does complexity of an artwork play in its reception? According to some theories, topological properties of the image are reflected in the topological aspects of the neural processes of the viewer. How can this hypothesis be tested, and what are its implications? Does a scaling regularity, such as the presence of quasi-fractal patterns in the image, lead to a more satisfying or memorable viewer experience? Can complexity, regularity, or symmetry in art serve to enhance or emphasize its meaning?

6. Setting and dimensionality

How does perception of a “flat” image, such as a painting or a photograph, differ from interacting with a textured quasi-2D artwork, a relief, or a fully three-dimensional sculpture, which requires viewing it from various vantage points? Inverting the problem, does interaction with mobiles or evolving art objects require a substantially different engagement of the viewer? Does the setting of an artwork, within a gallery or outdoors, in proximity to other works (by the same or different artists) or in a separate space make a difference, or is it merely a matter of the concentration of the viewer?

7. Timing and memory

What determines how quickly the impact of a single work of art, or an exhibition, fades away? What is the role of memory in experiencing art, during and after viewing it, and to what degree are these experiences conditioned by the viewer’s personality or socio- economic status? When is contemporary art perceived as a commentary on current events as opposed to conveying broader or more universal message?

III. Plans and Logistics

It is envisioned that the workshop would be proximate and possibly even concurrent with the exhibition on Art and Artificial Intelligence planned for the first half of 2023 at the Newton Institute. A major focus of the workshop is the difference in the perception of artist-created paintings versus images generated via machine learning, so the two endeavors would ideally be coordinated.

The workshop is envisioned as a 5-day meeting, primarily in person, with some participants possibly connecting via Zoom. It would also be available as a broadcast, in real time and as a recording on the Project website. The number of participants is expected to be about 20-30, with a judicious mix of mathematicians, physicists neurologists, cognitive scientists, art historians and curators, and hopefully also a few interested artists. A preliminary list of potential participants is being developed.