Resource-rational analysis is used to develop models that assume that people behave optimally given the structure of the task environment and the cost of cognitive operations. We argue in favor of a tight resource-rational analysis, an extension in which model parameters are independently constrained. As a case in point, we demonstrate how to develop a tight resource-rational model of the video game Space Track. Our approach consists of four steps. First, we measure performance-critical parameters in independent micro-tasks, which we input into mathematical models of cognitive processes. Second, we validate these models in other process-specific micro-tasks. Third, we rely on a theory of the cognitive architecture (i.e., ACT-R) to derive estimates of the time costs of these processes. Finally, we generate predictions for the main task, Space Track, by assuming that subjects are doing their best given their abilities. The generated individualized predictions were close to observed subject asymptotic performance, which demonstrated the viability of our approach, even in tasks of similar complexity to that of Space Track.

Open-ended tasks can be decomposed into the three levels of Newell's Cognitive Band: the Unit-Task level, the Operation level, and the Deliberate-Act level. We analyzed the video game Co-op Space Fortress at these levels, reporting both the match of a cognitive model to subject behavior and the use of electroencephalogram (EEG) to track subject cognition. The Unit Task level in this game involves coordinating with a partner to kill a fortress. At this highest level of the Cognitive Band, there is a good match between subject behavior and the model. The EEG signals were also strong enough to track when Unit Tasks succeeded or failed. The intermediate Operation level in this task involves legs of flight to achieve a kill. The EEG signals associated with these operations are much weaker than the signals associated with the Unit Tasks. Still, it was possible to reconstruct subject play with much better than chance success. There were significant differences in the leg behavior of subjects and models. Model behavior did not provide a good basis for interpreting a subject's behavior at this level. At the lowest Deliberate-Act level, we observed overlapping key actions, which the model did not display. Such overlapping key actions also frustrated efforts to identify EEG signals of motor actions. We conclude that the Unit-task level is the appropriate level both for understanding open-ended tasks and for using EEG to track the performance of open-ended tasks.

NASA’s Evolutionary Xenon Thruster (NEXT) is a candidate for future deep space missions that offers high efficiency and specific impulse over a large power throttling range. One of the key life-limiting components is the ion accelerator system, which is subject to sputter erosion by low energy discharge plasma ions incident on the upstream screen grid and higher energy charge exchange ions that impact the downstream accelerator grid. The grid erosion codes CEX2D and CEX3D were validated with data from tests of NEXT as well as the NSTAR ion thruster and then used to assess the time to failure in space due to screen grid erosion and electron backstreaming caused by accelerator grid aperture erosion. Screen grid erosion was found to be important only at the lowest throttle levels, and was conservatively estimated to lead to failure after processing over 900 kg of xenon. The first failure mode at high power levels was found to be electron backstreaming due to accelerator grid hole wall erosion, which would occur after processing over 700 kg of propellant.

The data presented indicate that excitatory and inhibitory amino acids play important transmitter roles in both peripheral and central vestibular structures. Other transmitters and neuropeptides appear to have more subtle neuromodulatory roles. Acetylcholine has been proposed to act as a neuromodulator in the cerebellar cortex rather than as a transmitter. While several other neuropeptides are localized within the vestibular complex or have been shown physiologically or behaviorally to affect vestibular function, more investigation is required to clarify their role as neuromodulators in the vestibular system. It will be especially important to distinquish the possible effects of adrenal steroids from sex steroids, and to focus on possible gender differences with respect to the effect of steroids on vestibular function. It is now clear that excitatory and inhibitory amino acids and their receptors play a major role in vestibular function, while acetylcholine, monoamines, and neuropeptides play somewhat secondary, neuromodulatory roles.

Clouds, especially low clouds, are crucial for regulating Earth's energy balance and mediating the response of the climate system to changes in greenhouse gas concentrations. Despite their importance for climate, they remain relatively poorly understood and are inaccurately represented in climate models. A principal reason is that the high computational expense of simulating them with large-eddy simulations (LES) has inhibited broad and systematic numerical experimentation and the generation of large datasets for training parametrization schemes for climate models. Here we demonstrate LES of low clouds on Tensor Processing Units (TPUs), application-specific integrated circuits that were originally developed for machine learning applications. We show that TPUs in conjunction with tailored software implementations can be used to simulate computationally challenging stratocumulus clouds in conditions observed during the Dynamics and Chemistry of Marine Stratocumulus (DYCOMS) field study. The TPU-based LES code successfully reproduces clouds during DYCOMS and opens up the large computational resources available on TPUs to cloud simulations. The code enables unprecedented weak and strong scaling of LES, making it possible, for example, to simulate stratocumulus with $10\times$ speedup over real-time evolution in domains with a $34.7~\mathrm{km} \times 53.8~\mathrm{km}$ horizontal cross section. The results open up new avenues for computational experiments and for substantially enlarging the sample of LES available to train parameterizations of low clouds.

We studied collaborative skill acquisition in a dynamic setting with the game Co-op Space Fortress. While gaining expertise, the majority of subjects became increasingly consistent in the role they adopted without being able to communicate. Moreover, they acted in anticipation of the future task state. We constructed a collaborative skill acquisition model in the cognitive architecture ACT-R that reproduced subject skill acquisition trajectory. It modeled role adoption through reinforcement learning and predictive processes through motion extrapolation and learned relevant control parameters using both a reinforcement learning procedure and a new to ACT-R supervised learning procedure. This is the first integrated cognitive model of collaborative skill acquisition and, as such, gives us valuable insights into the multiple cognitive processes that are involved in learning to collaborate.

In this paper, we study the stability and instability of plane wave solutions to semilinear systems of wave equations satisfying the null condition. We identify a condition which allows us to prove the global nonlinear asymptotic stability of the plane wave. The proof of global stability requires us to analyze the geometry of the interaction between the background plane wave and the perturbation. When this condition is not met, we are able to prove linear instability assuming an additional genericity condition. The linear instability is shown using a geometric optics ansatz.

We introduce a two-stage probabilistic framework for statistical downscaling using unpaired data. Statistical downscaling seeks a probabilistic map to transform low-resolution data from a biased coarse-grained numerical scheme to high-resolution data that is consistent with a high-fidelity scheme. Our framework tackles the problem by composing two transformations: (i) a debiasing step via an optimal transport map, and (ii) an upsampling step achieved by a probabilistic diffusion model with a posteriori conditional sampling. This approach characterizes a conditional distribution without needing paired data, and faithfully recovers relevant physical statistics from biased samples. We demonstrate the utility of the proposed approach on one- and two-dimensional fluid flow problems, which are representative of the core difficulties present in numerical simulations of weather and climate. Our method produces realistic high-resolution outputs from low-resolution inputs, by upsampling resolutions of 8x and 16x. Moreover, our procedure correctly matches the statistics of physical quantities, even when the low-frequency content of the inputs and outputs do not match, a crucial but difficult-to-satisfy assumption needed by current state-of-the-art alternatives. Code for this work is available at: https://github.com/google-research/swirl-dynamics/tree/main/swirl_dynamics/projects/probabilistic_diffusion.

Memory should make more available things that are more likely to be needed. Across multiple environmental domains, it has been shown that such a system would match qualitatively the memory effects involving repetition, delay, and spacing (Schooler & Anderson, 2017). To obtain data of sufficient size to study how detailed patterns of past appearance predict probability of being needed again, we examined the patterns with which words appear in large two data sets: tweets from popular sources and comments on popular subreddits. The two data sets show remarkably similar statistics, which are also consistent with earlier, smaller studies of environmental statistics. None of a candidate set of mathematical models of memory do well at predicting the observed patterns in these environments. A new model of human memory based on the environmental model proposed by Anderson and Milson (1989) did better at predicting the environmental data and a wide range of behavioral studies that measure memory availability by probability of recall and speed of retrieval. A critical variable in this model was range, the span of time over which an item occurs, which was discovered in mining the environmental data. These results suggest that theories of memory can be guided by mining of the statistical structure of the environment. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

NASA's Europa Clipper mission aims to conduct detailed reconnaissance of Jupiter's icy moon Europa and to investigate whether the moon could harbor conditions suitable for life. To perform these tasks Clipper will carry a suite of state-of-the-art scientific instruments, many of which are susceptible to the effects of molecular contamination and to interactions with the natural Jovian radiation environment. Recent ground testing conducted by the JPL Contamination Control group in the JPL Dynamitron particle accelerator (high-energy radiation source) has demonstrated that many common spacecraft materials exhibit significantly increased rates of molecular outgassing under exposure to high-energy radiation characteristic of the Europan environment [1]. This includes materials to be used in Clipper's thermal blankets and solar arrays, and subsequent free-molecular transport analyses showed that the increases in expected outgassing attributable to radiation would lead to exceedances of molecular deposition requirements for several of Clipper's instruments during the planned mission. The JPL Contamination Control group and the Europa Clipper project investigated testing and analysis refines in parallel with project mitigations strategies to protect Clipper's instruments. Refinements included improving the outgassing testing configuration and performing higher fidelity free-molecular flow analyses of instrument interiors. The project mitigations investigated include developing and testing alternate thermal blanket materials with a lower outgassing response under radiation and developing contamination shields that block line-of-sight between outgassing source surfaces and sensitive instrument surfaces. These mitigations were considered and coordinated with the impacted instrument teams such that any updates to the instrument requirements, available operation mitigations, or science robustness could be considered holistically. A combination of these strategies applied uniquely to each instrument proved most effective at mitigating predicted increases in molecular deposition caused by radiation induced outgassing. This approach demonstrates a novel method of identifying, assessing, and mitigating radiation induced outgassing that will be relevant to future space exploration missions with exposure to high-radiation environments.

Abstract : This paper describes the ACT theory of the learning of procedures. ACT is a computer simulation program that uses a propositional network to represent knowledge of general facts and a set of productions (condition action rules) to represent knowledge of procedures. There are currently four different mechanisms by which ACT can make additions and modifications to its set of productions as required for procedural learning: designation, strengthening, generatlization, and discrimination. Designation refers to the ability of productions to call for the creation of new productions. Strengthening a production may have important consequences for performance, since a production's strength determines the amount of system resources that will be allocated to its processing. Finally, generalization and discrimination refer to complementary processes that produce better performance by either extending or restricting the range of situations in which a production will apply. Each of these four mechanisms is discussed in detail and related to the available psychological data on procedural learning. The small-scale simulations of learning provided as examples are drawn from the domains of language processing and computer programming, since our ultimate goal is for ACT to learn the complex procedures required in such domains. (Author)

We examined the detailed behavioral characteristics of transfer of skill and the ability of the adaptive control of thought rational (ACT-R) architecture to account for this with its new Controller module. We employed a simple action video game called Auto Orbit and investigated the control tuning of timing skills across speed perturbations of the environment. In Auto Orbit , players needed to learn to alternate turn and shot actions to blow and burst balloons under time constraints imposed by balloon resets and deflations. Cognitive and motor skill transfer was assessed both in terms of game performance and in terms of the details of their motor actions. We found that skill transfer across speeds necessitated the recalibration of action timing skills. In addition, we found that acquiring skill in Auto Orbit involved a progressive decrease in variability of behavior. Finally, we found that players with higher skill levels tended to be less variable in terms of action chunking and action timing. These findings further shed light on the complex cognitive and motor mechanisms of skill transfer across speeds in complex task environments.

The LISP Intelligent Tutoring System (LISPITS) is an instructional program that helps students learn to program in the computer language LISP. Specifically, the tutor helps students with homework exercises. In each exercise the student is given a written description of a short program to write and, as the student types the program at the terminal, the tutor monitors the student's performance and provides assistance when errors are made. The tutor currently covers the first twelve chapters of an introductory LISP text (Anderson, Corbett & Reiser, 1987) and includes approximately 240 exercises.

Use of continuous glucose monitoring (CGM) improves clinical outcomes in type 1 diabetes, and significant benefits been demonstrated in patients with type 2 diabetes, including improved glycemic control, better treatment adherence, and an increased understanding of their treatment regimens. Currently, there are two types of CGM systems: real-time CGM (rtCGM) and flash CGM (FCGM). Retrospective analysis of CGM data allows patients and their clinicians to identify glycemic patterns that support and facilitate informed therapy decisions. With the increasing prevalence of diabetes, primary care physicians will be compelled to take on more responsibility for managing patients with diabetes. This article focuses on practical approaches and decision-making strategies for utilizing FCGM in primary care settings.

Insect-borne flaviviruses produce a 300-500-base long noncoding RNA, termed subgenomic flavivirus RNA (sfRNA), by stalling the cellular 5'-3'-exoribonuclease 1 (XRN1) via structures located in their 3' UTRs. In this study, we demonstrate that sfRNA production by Zika virus represses XRN1 analogous to what we have previously shown for other flaviviruses. Using protein-RNA reconstitution and a stringent RNA pulldown assay with human choriocarcinoma (JAR) cells, we demonstrate that the sfRNAs from both dengue type 2 and Zika viruses interact with a common set of 21 RNA-binding proteins that contribute to the regulation of post-transcriptional processes in the cell, including splicing, RNA stability, and translation. We found that four of these sfRNA-interacting host proteins, DEAD-box helicase 6 (DDX6) and enhancer of mRNA decapping 3 (EDC3) (two RNA decay factors), phosphorylated adaptor for RNA export (a regulator of the biogenesis of the splicing machinery), and apolipoprotein B mRNA-editing enzyme catalytic subunit 3C (APOBEC3C, a nucleic acid-editing deaminase), inherently restrict Zika virus infection. Furthermore, we demonstrate that the regulations of cellular mRNA decay and RNA splicing are compromised by Zika virus infection as well as by sfRNA alone. Collectively, these results reveal the large extent to which Zika virus-derived sfRNAs interact with cellular RNA-binding proteins and highlight the potential for widespread dysregulation of post-transcriptional control that likely limits the effective response of these cells to viral infection.

We study the problem of learning similarity functions over very large corpora using neural network embedding models. These models are typically trained using SGD with sampling of random observed and unobserved pairs, with a number of samples that grows quadratically with the corpus size, making it expensive to scale to very large corpora. We propose new efficient methods to train these models without having to sample unobserved pairs. Inspired by matrix factorization, our approach relies on adding a global quadratic penalty to all pairs of examples and expressing this term as the matrix-inner-product of two generalized Gramians. We show that the gradient of this term can be efficiently computed by maintaining estimates of the Gramians, and develop variance reduction schemes to improve the quality of the estimates. We conduct large-scale experiments that show a significant improvement in training time and generalization quality compared to traditional sampling methods.