Skip Nav Destination
Close Modal
Update search
NARROW
Format
TocHeadingTitle
Date
Availability
1-18 of 18
Josh Bongard
Close
Follow your search
Access your saved searches in your account
Would you like to receive an alert when new items match your search?
Sort by
Proceedings Papers
. isal2024, ALIFE 2024: Proceedings of the 2024 Artificial Life Conference115, (July 22–26, 2024) 10.1162/isal_a_00747
Abstract
View Paper
PDF
Evolution must explain both its ability to produce beneficial innovations as well as preserve organisms’ existing functional adaptedness to their environment. A proposed mechanism which resolves this tension is the concept of neutral networks, wherein mutations are not strictly beneficial or deleterious but neutral in their effect on organisms’ adaptedness. Neutral networks have been shown to be both prevalent and vast at multiple levels of biological organization. Additionally, there is much philosophical debate regarding how information flows between and across these levels of organization in reality. However, how to pragmatically engineer systems with multiscale structure to harness the inherent robustness that neutral networks confer remains largely unexplored. Here we show that, in hierarchical neural cellular automata (HNCA), various inter-scale connectivity architectures support mutational robustness and evolvability through the formation of neutral networks, wherein similar functional outcomes (e.g., morphogenesis, homeostasis) are achievable through diverse pathways of multiscale interactions. These findings can help inform the way we engineer artificial multiscale systems, e.g. hierarchical arrangements of robots. Operationalizing these insights may offer new ways of designing and engineering intelligent, robust, and adaptive machines. Additionally, the connection structures we explore have philsophical implications which may inform discussions of causal emergence in complex systems.
Proceedings Papers
. isal2024, ALIFE 2024: Proceedings of the 2024 Artificial Life Conference72, (July 22–26, 2024) 10.1162/isal_a_00807
Abstract
View Paper
PDF
Unconventional computing seeks to develop new means of acting on and interpreting the world. These emerge when new tools and computational substrates are built or discovered, or when existing artifacts are deployed in novel ways. Prior work designed sheets of vibrating particles to achieve mechanical polycomputation, wherein multiple logical operations were physically executed by the same parts at the same time. This works by exploiting the vibrational superposition of particles induced by external drives acting at multiple frequencies. In this paper, we introduce an idea called refractive computation, in which a sufficiently high density of polycomputed logic gates results in parallelized computations across driving frequencies. Parallelized logic gates are split across external drive frequencies in a single simulation, and emerge in the course of polycomputing sequential logic gates.
Proceedings Papers
. isal2024, ALIFE 2024: Proceedings of the 2024 Artificial Life Conference11, (July 22–26, 2024) 10.1162/isal_a_00724
Abstract
View Paper
PDF
Using soft materials to build robots is becoming increasingly popular as the importance of morphological complexity in robot design becomes apparent. Additionally, physics simulators which are differentiable are increasingly being used to optimize robot morphologies and controllers over e.g. evolutionary algorithms due to the computational efficiency of gradient descent. One of the most commonly used methods to simulate soft materials is the Material Point Method (MPM), and soft roboticists have implemented the MPM in differentiable robotics simulations and successfully transferred their optimized designs to the real world, validating this approach for real-world soft robot design. However, choosing parameters for MPM that render it stable in a differentiable simulator are non-obvious. For this reason, here we introduce for the first time a set of best practices for employing the MPM to design and optimize soft robots using differentiable physics engines. We perform grid searches over many of the parameters involved in MPM to determine simulation stability ranges and performant parameter choices for a displacement task. This will allow newcomers to MPM simulation to rapidly iterate to find parameters for their application.
Proceedings Papers
. isal2023, ALIFE 2023: Ghost in the Machine: Proceedings of the 2023 Artificial Life Conference70, (July 24–28, 2023) 10.1162/isal_a_00683
Abstract
View Paper
PDF
Robots have long been proposed as a solution for jobs that are difficult for humans, but their construction from nonrenewable and pollutant-causing materials presents a problem. The field of bio-robotics was developed, in part, to address this issue. In previous bio-robotic systems, such as Xenobots, AI-generated morphologies have been used to engineer desired behaviors in individual robots. However, this approach cannot be applied to biobots that are mass-fabricated as this limits our ability to control the behaviors of individual bots. While mass fabrication could have significant implications in the development of scalable biobot technologies for use in real-world applications, developing a reliable method to control their behavior remains a significant challenge. In this paper, we use evolutionary algorithms to create biobot swarm compositions that explore environments with varying obstacles efficiently at several scales. We demonstrate here that, while we cannot control the behavior of individual biobots, carefully selected swarm compositions can lead to desired behavior outcomes. This work thus provides one potential option for realizing biotechnology at scale, where mass-produced biobots must be filtered and combined appropriately.
Proceedings Papers
. isal2022, ALIFE 2022: The 2022 Conference on Artificial Life47, (July 18–22, 2022) 10.1162/isal_a_00530
Abstract
View Paper
PDF
Recent work with Lenia, a continuously-valued cellular automata (CA) framework, has yielded ~100s of compelling, bioreminiscent and mobile patterns. Lenia can be viewed as a continuously-valued generalization of the Game of Life, a seminal cellular automaton developed by John Conway that exhibits complex and universal behavior based on simple birth and survival rules. Life’s framework of totalistic CA based on the Moore neighborhood includes many other interesting, Life-like, CA. A simplification introduced in Lenia limits the types of Life-like CA that are expressible in Lenia to a specific subset. This work recovers the ability to easily implement any Life-like CA by splitting Lenia’s growth function into genesis and persistence functions, analogous to Life’s birth and survival rules. We demonstrate the capabilities of this new CA variant by implementing a puffer pattern from Life-like CA Morley/Move, and examine differences between related CA in Lenia and Glaberish frameworks: Hydrogeminium natans and s613, respectively. These CA exhibit marked differences in dynamics and character based on spatial entropy over time, and both support several persistent mobile patterns. The CA s613, implemented in the Glaberish framework, is more dynamic than the Hydrogeminium CA (and likely most Lenia-based CA) in terms of a consistently high variance in spatial entropy over time. These results suggest there may be a wide variety of interesting CA that can be implemented in the Glaberish variant of the Lenia framework, analogous to the many interesting Life-like CA outside of Conway’s Life. Supporting information and resources are open-source 1 .
Proceedings Papers
. isal2022, ALIFE 2022: The 2022 Conference on Artificial Life43, (July 18–22, 2022) 10.1162/isal_a_00526
Proceedings Papers
. isal2020, ALIFE 2020: The 2020 Conference on Artificial Life1, (July 13–18, 2020) 10.1162/isal_e_00358
Proceedings Papers
. isal2020, ALIFE 2020: The 2020 Conference on Artificial Life2, (July 13–18, 2020) 10.1162/isal_a_00357
Abstract
View Paper
PDF
The field of artificial life has a rich history of highly interdisciplinary research investigating methods for simulating and synthesizing living systems, drawing from research in diverse fields such as evolutionary computation, philosophy, and artificial chemistry. This volume contains the proceedings from the 2020 conference on Artificial Life, which was supposed to be held in Montreal, Canada, but was instead held virtually. The theme for this year's conference is “New Frontiers in AI: What can ALife offer AI?”, with selected works exploring both the possibilities and challenges of combining revolutionary ideas originating in artificial life with modern artificial intelligence methods.
Proceedings Papers
. isal2020, ALIFE 2020: The 2020 Conference on Artificial Life148-156, (July 13–18, 2020) 10.1162/isal_a_00317
Abstract
View Paper
PDF
Word embeddings have triggered great advances in natural language processing for non-embodied systems such as scene describers. Embeddings may similarly advance natural language understanding in robots, as long as those robots preserve the semantic structure of an embedding corpus in their actions. That is, a robot must act similarly when it hears ‘jump’ or ‘hop’ and differently when it hears ‘crouch’ or ‘launch’. This could help a robot learn language because it would immediately obey an unknown word such as ‘hop’ if it had been trained to obey ‘jump’. However, ensuring such alignment between semantic and behavioral structure is currently an open problem. In previous work we showed that the choice of a robot's mechanical structure can facilitate or obstruct a machine learning algorithm's ability to induce semantic and behavioral alignment. That work however required the investigator to create a loss function for each natural language command, including those for which formal definitions are elusive, such as ‘be interesting’. A more scalable approach is to bypass loss functions altogether by inviting non-experts to supply their own commands and reward robots that obey them. Here we found that more semantic and behavioral alignment existed among robots reinforced under popular commands than among robots reinforced under less popular commands. This suggests the crowd either chose alignment-inducing commands and/or preferred robots that acted similarly under similar commands. This may pave the way to scalable human-robot interaction by avoiding loss function construction and increasing the probability of zero-shot obedience to previously unheard commands.
Proceedings Papers
. isal2020, ALIFE 2020: The 2020 Conference on Artificial Life52-59, (July 13–18, 2020) 10.1162/isal_a_00243
Abstract
View Paper
PDF
Catastrophic forgetting continues to severely restrict the learnability of controllers suitable for multiple task environments. Efforts to combat catastrophic forgetting reported in the literature to date have focused on how control systems can be updated more rapidly, hastening their adjustment from good initial settings to new environments, or more circumspectly, suppressing their ability to overfit to any one environment. When using robots, the environment includes the robot's own body, its shape and material properties, and how its actuators and sensors are distributed along its mechanical structure. Here we demonstrate for the first time how one such design decision (sensor placement) can alter the landscape of the loss function itself, either expanding or shrinking the weight manifolds containing suitable controllers for each individual task, thus increasing or decreasing their probability of overlap across tasks, and thus reducing or inducing the potential for catastrophic forgetting.
Proceedings Papers
. isal2020, ALIFE 2020: The 2020 Conference on Artificial Life60-68, (July 13–18, 2020) 10.1162/isal_a_00310
Abstract
View Paper
PDF
In domains where measures of utility for automatically-designed artefacts (or agents performing subjective tasks) are difficult or impossible to mathematically describe (such as ‘be interesting’), human interactive search algorithms are an attractive alternative. However, despite notable achievements, they are still designed around a specific search method, resulting in a lack of problem generality: applying a new search algorithm requires an excessive amount of redesign such that an altogether new interactive method is formed in the process. This leads to missed opportunities for human interactive methods to utilize the power of state of the art optimization algorithms. Here, we introduce for the first time a framework for human interactive optimization that is agnostic to both the search method and the application problem. Using 13 different search methods on 24 fitness functions commonly found in evolutionary algorithm benchmarks, we show that our approach works on the majority of tested applications: many of the search methods, provided with access to the fitness functions, performed no better than our framework, which employs surrogate human participants who act as less informed and erroneous representations of the fitness function. In this way, our framework for interactive optimization provides a scalable solution by facilitating the integration of numerous types of current state of the art or future search algorithms. Future work will involve generalizing this method to admit multi-objective optimization methods and validation with human participants.
Proceedings Papers
. alife2018, ALIFE 2018: The 2018 Conference on Artificial Life606-613, (July 23–27, 2018) 10.1162/isal_a_00111
Abstract
View Paper
PDF
Catastrophic interference occurs when an agent improves in one training instance but becomes worse in other instances. Many methods intended to combat interference have been reported in the literature that modulate properties of a neural controller, such as synaptic plasticity or modularity. Here, we demonstrate that adjustments to the body of the agent, or the way its performance is measured, can also reduce catastrophic interference without requiring changes to the controller. Additionally, we introduce new metrics to quantify catastrophic interference. We do not show that our approach outperforms others on benchmark tests. Instead, by more precisely measuring interactions between morphology, fitness, and interference, we demonstrate that embodiment is an important aspect of this problem. Furthermore, considerations into morphology and fitness can combine with, rather than compete with, existing methods for combating catastrophic interference.
Proceedings Papers
. alife2018, ALIFE 2018: The 2018 Conference on Artificial Life461-468, (July 23–27, 2018) 10.1162/isal_a_00086
Abstract
View Paper
PDF
A long time goal of evolutionary roboticists is to create everincreasing lifelike robots which reflect the important aspects of biology in their behavior and form. One way to create such creatures is to use evolutionary algorithms and genotype to phenotype maps which act as proxies for biological development. One such algorithm is HyperNEAT whose use of a substrate which can be viewed as an abstraction of spatial development used by Hox genes. Previous work has looked into answering what effect changing the embedding has on HyperNEAT’s efficiency, however no work has been done on the effect of representing different aspects of the agents morphology within the embeddings. We introduce the term embodied embeddings to capture the idea of using information from the morphology to dictate the locations of neurons in the substrate. We further compare three embodied embeddings, one which uses the physical structure of the robot and two which use abstract information about the robot’s morphology, on an embodied version of the retina task which can be made modular, hierarchical, or a combination of both.
Proceedings Papers
. alife2018, ALIFE 2018: The 2018 Conference on Artificial Life242-249, (July 23–27, 2018) 10.1162/isal_a_00050
Abstract
View Paper
PDF
An ongoing discussion in biology concerns whether intrinsic mortality, or senescence, is programmed or not. The death (i.e. removal) of an individual solution is an inherent feature in evolutionary algorithms that can potentially explain how intrinsic mortality can be beneficial in natural systems. This paper investigates the relationship between mutation rate and mortality rate with a steady state genetic algorithm that has a specific intrinsic mortality rate. Experiments were performed on a predefined deceptive fitness landscape, the hierarchical if-and-only-if function (H-IFF). To test whether the relationship between mutation and mortality rate holds for more complex systems, an agent-based spatial grid model based on the H-IFF function was also investigated. This paper shows that there is a direct correlation between the evolvability of a population and an indiscriminate intrinsic mortality rate to mutation rate ratio. Increased intrinsic mortality or increased mutation rate can cause a random drift that can allow a population to find a global optimum. Thus, mortality in evolutionary algorithms does not only explain evolvability, but might also improve existing algorithms for deceptive/rugged landscapes. Since an intrinsic mortality rate increases the evolvability of our spatial model, we bolster the claim that intrinsic mortality can be beneficial for the evolvability of a population.
Proceedings Papers
. ecal2017, ECAL 2017, the Fourteenth European Conference on Artificial Life60-67, (September 4–8, 2017) 10.1162/isal_a_014
Abstract
View Paper
PDF
In evolutionary robotics, controllers are often represented as networks. Modularity is a desirable trait of such networks because modular networks are resistant to catastrophic forgetting and tend to have less connections than nonmodular ones. However, these advantages can only be realized if the control task is solvable by a modular network, and for any given practical task the control task depends on the choice of the robot’s morphology. Here we provide an example of a task solvable by robots with two different morphologies. We consider the most extreme kind of modularity – disconnectedness – and show that with the first morphology the task can be solved by a disconnected controller with few connections. On the other hand, the second morphology makes the task provably impossible for disconnected controllers and requires about three times more connections. For this morphology, most controllers that partially solve the task constitute local optima, forming an extremely deceptive fitness landscape. We show empirically that in this case a connection cost-based evolutionary algorithm for evolving modular controllers is greatly slowed down compared to the first morphology’s case. Finally, this performance gap increases as the task is scaled up. These results show that the morphology may be a major factor determining the performance of controller optimization. Although in our task the optimal morphology is obvious to a human designer, we hypothesize that as evolutionary robotics is scaled to more sophisticated tasks the optimization of morphology alongside the control might become a requirement for evolving modular controllers.
Proceedings Papers
. alif2016, ALIFE 2016, the Fifteenth International Conference on the Synthesis and Simulation of Living Systems226-233, (July 4–6, 2016) 10.1162/978-0-262-33936-0-ch042
Abstract
View Paper
PDF
The field of evolved virtual creatures has been suspiciously stagnant in terms of complexification of evolved agents since its inception over two decades ago. Many researchers have proposed algorithmic improvements, but none have taken hold and greatly propelled the scalability of early works. This paper suggests a more fundamental problem with co-evolving both the morphology and control of virtual creatures simultaneously one cemented in the theory of embodied cognition. We reproduce and explore in greater detail a previous finding in the literature: premature convergence of the morphology (compared to the convergence point of optimizing controllers), and discuss how this finding fits as a symptom of the proposed problem. We hope that this improved understanding of the fundamental problem domain will open the door for further scalability of evolved agents, and note that early findings from our future work point in that direction.
Proceedings Papers
. alif2016, ALIFE 2016, the Fifteenth International Conference on the Synthesis and Simulation of Living Systems684-691, (July 4–6, 2016) 10.1162/978-0-262-33936-0-ch109
Abstract
View Paper
PDF
The non-embodied approach to teaching machines language is to train them on large text corpora. However, this approach has yielded limited results. The embodied approach, in contrast, involves teaching machines to ground abstract symbols in their sensory-motor experiences, but howor whether humans achieve this remains largely unknown. We posit that one avenue for achieving this is to view language acquisition as a three-way interaction between linguistic, sensorimotor, and social dynamics: when an agent acts in response to a heard word, it is considered to have successfully grounded that symbol if it can predict how observers who understand that word will respond to the action. Here we introduce a methodology for testing this hypothesis: human observers issue arbitrary commands to simulated robots via the web, and provide positive or negative reinforcement in response to the robots resulting action. Then, the robots are trained to predict crowd response to these action-word pairs. We show that robots do learn to ground at least one of these crowd-issued commands: an association between jump, minimization of tactile sensation, and positive crowd response was learned. The automated, open-ended, and crowd-based aspects of this approach suggest it can be scaled up in future to increasingly capable robots and more abstract language.
Proceedings Papers
. alife2014, ALIFE 14: The Fourteenth International Conference on the Synthesis and Simulation of Living Systems973-980, (July 30–August 2, 2014) 10.1162/978-0-262-32621-6-ch158