Workshop Topics and Call for Paper

Despite its remarkable achievements, AI continues to yearn for further extensions in basic and applied research. In the sub-divisions of AI still dedicated to the synthetic modeling of natural cognition, front-line research targets patterns that are increasingly autonomous, adaptive and integrative, and intends to ground them in the structural features and the form(s) of organization characterizing biological systems – from unicellular systems to complex organisms and their (social) aggregates. Currently, this research horizon includes a multitude of bio-inspired lines of inquiry, sharing a ‘bottom-up’ approach – or, in other theoretical words, an ‘emergence by design’ approach – to the construction of artificial or synthetic systems that function and behave like biological systems.

The workshop Synthetic Approaches to Biology and Artificial Intelligence intends to select a series of original contributions related to this bio-inspired modeling of life and cognition, in its hard/soft/wet-ware or hybrid expressions. The primary aim is to promote the emergence of novel paradigms for AI, firmly grounded in SB and AL research, and capable of producing qualitative leaps into next generation of technological artifacts expressing adaptive, communicational, and integrative – “cognitive” – biological-like functions and behaviors.

• How the synthetic approach to biology, in its software, wetware and hardware forms, provide AI with new, relevant insights for the advancement in the scientific understanding of natural cognition? In which conditions, in what ways, in which domains could these synthetic explorations positively contribute to AI-based technologies? That is: What can the synthetic approach to biology, in its different expression, offer to AI?

• What’s the complexity threshold, the links, the rules, the topology, the forms, the boundaries of Artificial Life systems (chemical networks, synthetic cells and droplets, neural networks, swarm robotics, evolutionary agents, …) in order to display minimal adaptive, communicational, integrative – “cognitive” – functions/behaviors?

• How to design Artificial Life systems in robotic, computer science, synthetic biology domains that allow AI emergence, and how the latter compares with the biological counterparts?

This workshop aims at bringing together contributions related to these and related questions, and investigating one or more aspects of the (possible/actual) relationships between the synthetic approach to biology and AI.

The workshop will start with individual contributions, followed by a structured discussion, in the form of a round-table. The discussion aims at constituting a productive dialogue among the different specialists, a highly cross-disciplinary forum, engaging participants in generating answers to the many open questions in the field, and cooperate in the critical improvements of the projects and the ideas there presented.

• Which are the groundings, the procedures, the expected results and the impacts of current research programs involving the synthetic approach to biology in AI?

• Can we, at the present time, plan concrete collaborations between computer science, robotics and synthetic biology in the scientific study of natural forms of cognition and intelligence? How?

• Can these cross-fertilizations contribute to the development of new forms of cognition and intelligence, alternative to the natural ones and to the existing forms of artificial cognition and intelligence?

We expect keynote and selected talks focused on conceptual and experimental issues related to the synthetic modeling of life and cognition, in any respect and in any domain. This way, we intend to put into focus the most advanced research relevant for the topic of Synthetic Approaches to Biology and Artificial Intelligence.

Final Program (time refers to Central European Summer Time)

10,00-10,20 Welcome and Introduction. L. Damiano. Introduction to SB-AI 7

10,20-10,50 F. Bianchini. The organization from life to cognition and the current approaches in AI

10.50-11,20 T. Ikegami. Collective Memory of Tetrahymena Populations

11,20-11,30 Break

11,30-12,00 S. Holler and M. M. Hanczyc. Artificial cells as biological cell transporters

12,00-12,45 J. C. Letelier. Invited Lecture. Part A: Structural Coupling and Bayesian Inference; Part B: CYBERNETICS in CHILE: The fundamental friendship between Heinz von Foerster and Humberto Maturana

12,45-13,00 General discussion and Conclusion

Abstracts

• The organization from life to cognition and the current approaches in AI

  • Francesco Bianchini, University of Bologna

  • In the seminal article on autopoiesis by Maturana e Varela (1974), the authors states in the very introduction the fundamental distance between processes that are history dependent and history independent in the biological world, the former being especially connected to evolution ad ontogenesis and the latter being related to the organizational features of biological individuals. Moreover, the authors make an even more significant statement, concerning the fact that organization of biological systems, and their unity, precede other features of their living nature as evolution and reproduction. Maturana and Varela claims give rise to methodological issues on explanation of systems from the point of view of biological as well as cognitive nature and such issues are particularly clear in the sciences of the artificial where they are reflected by some major or minor trends of research. On one hand, computational architectures can be seen as an example of organizational approach that is very useful to explain cognition, even though they are not fully committed with the coupling between the cognitive systems and the environment. On the other hand, the big development of machine learning has led to the Explainable AI problem, that is the problem to explain the outcomes and the behavior of neural networks and deep learning systems. In this case, we have systems that are very able to perform some tasks because of their unitarian (mathematical and computational) behavior, but they are not understandable regarding the contribution of their parts to the results given by the system understood as a unity. This is another case in which the system has an autonomy, in the very sense of autopoiesis, and so an autonomous “life” as an organizational unity with an emergent performance. But, what about cognition? In my talk, I would like to explore the connection between biological systems and cognition in light of some of the most recent developments of artificial systems and AI trends that could be traced back to the notions of autonomy and organization and try to identify the “chemical” part of AI in synthetic artificial systems

• Collective Memory of Tetrahymena Populations

  • Takashi Ikegami, University of Tokyo

  • Biological evolution is fundamentally concerned with "populations" in which behaviors emerge that are not predictable at the individual level. For example, the key to this is phenotypic plasticity. Here, we discuss the inheritance of traits by populations based on experiments using Tetrahymena and the possible emergence of AI.

• Artificial cells as biological cell transporters

  • Silvia Holler and Martin M. Hanczyc, University of Trento

  • Soft matter systems can be driven out of equilibrium and can respond to externally imposed stimuli. The transition from equilibrium to non-equilibrium can be driven through for example chemical potential, external fluid flow or static external fields. An example of soft matter systems that show response to these kinds of transitions are liquid droplets. Liquid droplets can be created mixing two immiscible fluids (oil and water) and can exhibit behaviors such as fission, fusion and movement. Movement is normally shown by living cells and organisms and can be defined as a ‘life-like’ behavior. We were able to create systems of oil-in-water droplets that self-organize and move in response to external stimuli. Decanol droplets move for example in response to pH or salt gradients. We exploited droplets to transport objects such as cells. Specifically, we developed chemotactic droplets able to move light cargos such as hydrogel alginate capsules embedded with living cells. Using this system, we demonstrate efficient and sterile transport of a few types of bacteria and yeast, and even human cell lines. We recently discovered that some eukaryotic cell lines, only when placed in capsules, secrete compounds that act as surfactants, thereby reinforcing the interface between the artificial and living systems. This is an example of not only how the interface between artificial life and biological life could be designed but how the one system can augment the other.

• Invited lecture. Part A: Structural Coupling and Bayesian Inference; Part B: CYBERNETICS in CHILE: The fundamental friendship between Heinz von Foerster and Humberto Maturana

  • Juan Carlos Letelier, University of Chile

  • Structural Coupling and Bayesian Inference. How Autopoietic systems construct their umwelt? This question can be answered if we focus on the following premises; a) Autopoietic systems are full of plastic and regulatory mechanisms, b)Autopoietic systems, like organisms, act upon their surroundings, c)The essential problem confronted by and Autopoietic system is to have a coherent loop linking Perception to Action and d) this link is a probabilistic mechanism that can be interpreted in the context of Bayesian Inference.

  • CYBERNETICS in CHILE: The fundamental friendship between Heinz von Foerster and Humberto Maturana. To understand the origin of the ideas of AUTOPOIESIS (and Biology of Cognition) one overlooked aspect is the deep friendship between Heinz von Foerster and Humberto Maturana. This talk will survey many little known aspects as well as some important political ramifications like the SYNCO project where another giant of Cybernetics (Stafford Beer) played a fundamental role in trying to implement a digital network connecting all manufacturing plants in the Socialist Chile of 1971-1973.