An insect‑scale robot that integrates aerial takeoff and open-loop attitude modulation, terrestrial crawling, water-surface gliding, and electrowetting-triggered controlled immersion within a 140-mg platform is presented, addressing the fundamental trade-off between operational diversity and extreme miniaturization.
Abstract
Insects traverse diverse environments by reusing body structures and appendages through adaptable mechanics, an ability that remains a grand challenge for insect-scale robotics under strict mass constraints. Here we present an insect‑scale robot that integrates aerial takeoff and open-loop attitude modulation, terrestrial crawling, water-surface gliding, and electrowetting-triggered controlled immersion within a 140-mg platform, with these capabilities demonstrated as representative open-loop functional tests rather than a single continuous autonomous mission. Moving beyond actuator‑heavy designs, the robot encodes structural intelligence in a single multifunctional morphing frame driven by shape-memory alloy artificial muscles. This morphing body reconfigures to take on distinct roles across diverse environments: it acts as a steering transmission for asymmetric flapping-wing kinematics, a linear oscillator for terrestrial crawling through anisotropic friction, and a deformation actuator that tunes kirigami-inspired footpad gaps for interfacial support and immersion control. Footpad modulation enables active electrowetting, rupturing the water surface at 225 V to achieve controlled immersion and increasing load capacity to 1.7 × body weight. By integrating soft actuation, structural mechanics, and interfacial control into one reconfigurable structure, this work establishes a paradigm of functional recursion, where a minimal physical platform repeatedly reassigns form and function, addressing the fundamental trade-off between operational diversity and extreme miniaturization.
This paper investigates developing real-world perception for active agents, proposes Gibson Environment for this purpose, and showcases a set of perceptual tasks learned therein.
F. Xia, Amir Zamir, Zhi-Yang He et al.· 2018 IEEE/CVF Conference on...· 1,069 citations· ⚡112
GAOKAO-Bench is introduced, an intuitive benchmark that employs questions from the Chinese GAOKAO examination as test samples, including both subjective and objective questions that contribute a robust evaluation benchmark for future large language models and offers valuable insights into the advantages and limitations...
Xiaotian Zhang, Chun-yan Li, Yi Zong et al.· arXiv.org· 216 citations· ⚡17
An algorithm is proposed which leverages disentangled semantic factors to generate adversarial perturbation by altering controlled semantic attributes to fool the learner towards various "adversarial" targets.
Haonan Qiu, Chaowei Xiao, Lei Yang et al.· arXiv.org· 208 citations· ⚡17
This work proposes a HALO that directly maximizes the utility of generations instead of maximizing the log-likelihood of preferences, and suggests that the best loss depends on the inductive biases most appropriate for a given setting, an oft-overlooked consideration.
This work investigates the possibilities of using LLMs in a resume screening setting via a document retrieval framework that simulates job candidate selection and finds that the MTEs are biased, significantly favoring White-associated names in 85% of cases and female-associated names in only 11.1% of cases.
This work shows that orders of magnitude enhancement in performance could be obtained by a combination of hardware improvements and tight quantum-HPC integration and introduces high-performance architectures for quantum-probabilistic computing with custom-designed accelerators to tackle today's industry-scale classical...
Masoud Mohseni, Artur Scherer, K. Johnson et al.· arXiv.org· 121 citations· ⚡9