Search for: All records

Abstract This contribution is concerned with the following issue: can pretrained large language models (LLMs) be refined and customized to the point where they become virtual assistants helping users with the effective use of a simulation tool? In this case study, the “simulation tool” considered is PyChrono, an open source multi-physics dynamics engine for multibody systems. We present a framework for refining and customizing both open- and closed-source LLMs to harness the power of AI in generating scripts that perform PyChrono virtual experiments. We refine and customize several classes of LLMs through a process that leads to a quantifiable improvement in the quality of the generated PyChrono simulation scripts. These scripts can range from simple single-pendulum simulations to complex virtual experiments involving full vehicles on deformable terrain. While the generated scripts are rarely perfect, they often serve as strong starting points for the user to modify and improve on. Additionally, the LLM can answer specific API questions about the simulator, or recommend modeling approaches. The framework discussed is general and can be applied to lower the entry barrier for simulation tools associated with other application domains. 

The failure resistance of polymer networks dictates their utility as material candidates across industries. However, relating the key length scales driving crack growth to molecular mechanisms remains a key bottleneck in predicting and designing against fracture. The fractocohesive length—defined in terms of the ratio of fracture energy to the specific work to rupture—of a material correlates with the length scale of energy dissipation and controls fracture resistance. Although the Lake–Thomas model predicts the fractocohesive length of a perfect polymer network to match the undeformed mesh size, real soft materials exhibit values that far exceed this prediction. Here, we report extraordinarily high fractocohesive lengths in polymer-like networks with and without defects. We find that even perfect networks can have fractocohesive lengths orders of magnitude higher than the undeformed mesh size due to highly nonlinear chain behavior giving rise to nonlocal effects during fracture. Introducing defects further increases the fractocohesive length. We identify quantitative relations between nonlinear chain mechanics, defect length, defect density, and fractocohesive length. Overall, strain-stiffening chain behavior, defect density, and defect size independently correlate with larger fractocohesive lengths in polymer-like networks, and their individual effects can be collapsed into a single power law scaling. These outcomes point the way towards improved physics-informed design of soft yet tough polymers and metamaterials. 

Abstract The virial coefficient (f), which is meant to encapsulate broad-line region (BLR) geometry and kinematics, remains one of the largest sources of systematic uncertainty in black hole mass estimates for active galactic nuclei (AGNs). While the use of a sample average 〈f〉 enables black hole mass estimates across large samples and cosmological distances, individual AGNs may deviate from this average due to differences in BLR structure and viewing angle. In previous work, we reported marginal evidence for a correlation betweenfand the shape of the broad Hβemission line, ${\log }_{10}(\mathrm{FWHM}/\sigma )$ . In this work, we update our sample to include 10 new sources withCARAMELBLR dynamical modeling, increasing both the black hole mass range and statistical power of our analysis. We find marginal evidence for a correlation betweenfand ${\log }_{10}(\mathrm{FWHM}/\sigma )$ , with a slope and intrinsic scatter consistent with previous results. The confirmation of this trend across a larger sample further supports the idea that line profile shape may reflect BLR properties in a way that directly impactsf. If confirmed with future BLR dynamical modeling of sources within a wider range of ${\log }_{10}(\mathrm{FWHM}/\sigma )$ , this relationship could enable empirical estimates of the virial coefficient and improve single-epoch black hole mass estimates across cosmic time. 

The human hand is dexterous and versatile, allowing it to interact with physical and virtual environments. The ability to track hand movements during daily activities could be of use in the development of spatial computing, virtual and augmented reality, robotics and prosthetics. However, current techniques based on cameras, strain and inertial sensors, and electromyography sensors have limited view angles and hand positions, have constrained hand activities and sensations, and can track only discrete hand gestures, respectively. Here we report a fully integrated, wireless and wearable ultrasound imaging wristband that is combined with an artificial intelligence algorithm. The wristband can continuously track arbitrary hand configurations of the five fingers and the palm in real time during daily activities with a delay of less than 120 ms. We show that the wristband can be used for intuitive and versatile controls in virtual-reality and robotic-hand applications.