An official website of the United States government
Biological mechanosensation has been a source of inspiration for advancements in artificial sensory systems. Animals rely on sensory feedback to guide and adapt their behaviors and are equipped with a wide variety of sensors that carry stimulus information from the environment. Hair and hair-like sensors have evolved to support survival behaviors in different ecological niches. Here, we review the diversity of biological hair and hair-like sensors across the animal kingdom and their roles in behaviors, such as locomotion, exploration, navigation, and feeding, which point to shared functional properties of hair and hair-like structures among invertebrates and vertebrates. By reviewing research on the role of biological hair and hair-like sensors in diverse species, we aim to highlight biological sensors that could inspire the engineering community and contribute to the advancement of mechanosensing in artificial systems, such as robotics.
more »
« less
This content will become publicly available on May 18, 2026
Of Mice and Machines: A Comparison of Learning Between Real World Mice and RL Agents
Recent advances in reinforcement learning (RL) have demonstrated impressive capabilities in complex decision-making tasks. This progress raises a natural question: how do these artificial systems compare to biological agents, which have been shaped by millions of years of evolution? To help answer this question, we undertake a comparative study of biological mice and RL agents in a predator-avoidance maze environment. Through this analysis, we identify a striking disparity: RL agents consistently demonstrate a lack of self-preservation instinct, readily risking ``death'' for marginal efficiency gains. These risk-taking strategies are in contrast to biological agents, which exhibit sophisticated risk-assessment and avoidance behaviors. Towards bridging this gap between the biological and artificial, we propose two novel mechanisms that encourage more naturalistic risk-avoidance behaviors in RL agents. Our approach leads to the emergence of naturalistic behaviors, including strategic environment assessment, cautious path planning, and predator avoidance patterns that closely mirror those observed in biological systems.
more »
« less
- Award ID(s):
- 2235451
- PAR ID:
- 10612017
- Publisher / Repository:
- arXiv:2505.12204
- Date Published:
- Journal Name:
- arXivorg
- ISSN:
- 2331-8422
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Cyber defense exercises are an important avenue to understand the technical capacity of organizations when faced with cyber-threats. Information derived from these exercises often leads to finding unseen methods to exploit vulnerabilities in an organization. These often lead to better defense mechanisms that can counter previously unknown exploits. With recent developments in cyber battle simulation platforms, we can generate a defense exercise environment and train reinforcement learning (RL) based autonomous agents to attack the system described by the simulated environment. In this paper, we describe a two-player game-based RL environment that simultaneously improves the performance of both the attacker and defender agents. We further accelerate the convergence of the RL agents by guiding them with expert knowledge from Cybersecurity Knowledge Graphs on attack and mitigation steps. We have implemented and integrated our proposed approaches into the CyberBattleSim system.more » « less
-
null (Ed.)As IT/OT convergence continues to evolve, the traditionally isolated ICS/OT systems are increasingly exposed to a myriad of online and offline threats. Although IIoT enhances the reachability in ICS, im- proved data analytics, ensuring ease of access and decision making, it unwittingly opens the ICS environment to attackers. The design of IIoT introduces multiple entry points to an isolated system, which is used to protect itself via air-gapping and risk avoidance strategies. This study explores a comprehensive mapping of threats and risks for IT/OT convergence. Additionally, we propose IIoT-ARAS - an automated risk assessment system based on OCTAVE Allegro and ISO/IEC 27030 methodologies. The design of IIoT-ARAS is aimed to be agentless, with minimum interruptions to the OT environment. Furthermore, the system performs automated regular asset inventory checks, threshold optimization, probability computation, risk evaluations, and contingency plan configuration.more » « less
-
Yue, Bi-Song (Ed.)Large mammalian herbivores use a diverse array of strategies to survive predator encounters including flight, grouping, vigilance, warning signals, and fitness indicators. While anti-predator strategies appear to be driven by specific predator traits, no prior studies have rigorously evaluated whether predator hunting characteristics predict reactive anti-predator responses. We experimentally investigated behavioral decisions made by free-ranging impala, wildebeest, and zebra during encounters with model predators with different functional traits. We hypothesized that the choice of response would be driven by a predator’s hunting style (i.e., ambush vs. coursing) while the intensity at which the behavior was performed would correlate with predator traits that contribute to the prey’s relative risk (i.e., each predator’s prey preference, prey-specific capture success, and local predator density). We found that the choice and intensity of anti-predator behaviors were both shaped by hunting style and relative risk factors. All prey species directed longer periods of vigilance towards predators with higher capture success. The decision to flee was the only behavior choice driven by predator characteristics (capture success and hunting style) while intensity of vigilance, frequency of alarm-calling, and flight latency were modulated based on predator hunting strategy and relative risk level. Impala regulated only the intensity of their behaviors, while zebra and wildebeest changed both type and intensity of response based on predator traits. Zebra and impala reacted to multiple components of predation threat, while wildebeest responded solely to capture success. Overall, our findings suggest that certain behaviors potentially facilitate survival under specific contexts and that prey responses may reflect the perceived level of predation risk, suggesting that adaptive functions to reactive anti-predator behaviors may reflect potential trade-offs to their use. The strong influence of prey species identity and social and environmental context suggest that these factors may interact with predator traits to determine the optimal response to immediate predation threat.more » « less
-
ABSTRACT Disparate bodies of literature implicate risk avoidance and energy conservation as important drivers of animal movement decisions. Theory posits that these phenomena interact in ecologically consequential ways, but rigorous empirical tests of this hypothesis have been hampered by data limitations. We fuse fluid dynamics, telemetry, and attack data to reconstruct risk and energy landscapes traversed by migrating juvenile salmon and their predators. We find that migrants primarily use midriver microhabitats that facilitate migration at night. During daylight, predators become more aggressive in the midriver, and prey reduce midriver use in favour of nearshore microhabitats, resulting in increased energy expenditure and decreased migration efficiency. Predators attack most when migrants are not prioritising threat avoidance and during ephemeral periods of low lighting. Our findings suggest that predator–prey interactions result from an interplay between landscapes of fear and energy, which can determine the degree to which predators affect prey through mortality or fear.more » « less
