Data-driven molecular discovery leverages artificial intelligence/machine learning (AI/ML) and generative modeling to filter and design novel molecules. Discovering novel molecules requires accurate out-of-distribution (OOD) predictions, but ML models struggle to generalize OOD. Currently, no systematic benchmarks exist for molecular OOD prediction tasks. We present BOOM, $\textbf{b}$enchmarks for $\textbf{o}$ut-$\textbf{o}f$-$\textbf{d}$istribution $\textbf{m}$olecular property predictions: a chemically-informed benchmark for OOD performance on common molecular property prediction tasks. We evaluate over 150 model-task combinations to benchmark deep learning models on OOD performance. Overall, we find that no existing model achieves strong generalization across all tasks: even the top-performing model exhibited an average OOD error 3$\times$ higher than in-distribution. Current chemical foundation models do not show strong OOD extrapolation, while models with high inductive bias can perform well on OOD tasks with simple, specific properties. We perform extensive ablation experiments, highlighting how data generation, pre-training, hyperparameter optimization, model architecture, and molecular representation impact OOD performance. Developing models with strong OOD generalization is a new frontier challenge in chemical ML. This open-source benchmark is available at https://github.com/FLASK-LLNL/BOOM

Are you fed up with "lies, d---ned lies, and statistics" made up from data in our media? For claims based on structured data, we present a system to automatically assess the quality of claims (beyond their correctness) and counter misleading claims that cherry-pick data to advance their conclusions. The key insight is to model such claims as parameterized queries and consider how parameter perturbations affect their results. We demonstrate our system on claims drawn from U.S. congressional voting records, sports statistics, and publication records of database researchers.

JFLAP is an educational software tool for experimenting with several types of automata, grammars and proofs related to the area of theoretical computer science. For example, one can create an NFA, convert it to a DFA, minimize the number of states in the DFA, convert it to a regular expression or convert it to a regular grammar. Once created, one can simulate test strings. JFLAP includes a large number of topics and proofs to experiment with from the regular languages, context-free languages, recursively enumerable languages and l-systems. This poster describes several new features in a redesign of JFLAP that includes a more flexible grammar representation, explicit formal definitions, basic set theory tools, and new algorithms.

For four years we have been integrating computing into a variety of middle school disciplines via the Alice programing language. Early on we focused on creating curriculum materials for teachers to use in teaching the basic programming and animation concepts. This paper describes our efforts over the past two years in creating model projects for students to build in all disciplines, and our most recent focus on science and mathematics projects. For science we have introduced experiments in Alice and the tools needed for them. In mathematics we have created projects to increase their understanding of programming and to use the projects to increase their understanding of mathematics. We also discuss our efforts in workshops to teach K-12 teachers Alice and an analysis of the teachers' lesson plans and worlds developed in the most recent workshop.

Data collection using Autonomous Underwater Vehicles (AUVs) is increasing in importance within the oceanographic research community. Contrary to traditional moored or static platforms, mobile sensors require intelligent planning strategies to maneuver through the ocean. However, the ability to navigate to high-value locations and collect data with specific scientific merit is worth the planning efforts. In this study, we examine the use of ocean model predictions to determine the locations to be visited by an AUV, and aid in planning the trajectory that the vehicle executes during the sampling mission. The objectives are: a) to provide near-real time, in situ measurements to a large-scale ocean model to increase the skill of future predictions, and b) to utilize ocean model predictions as a component in an end-to-end autonomous prediction and tasking system for aquatic, mobile sensor networks. We present an algorithm designed to generate paths for AUVs to track a dynamically evolving ocean feature utilizing ocean model predictions. This builds on previous work in this area by incorporating the predicted current velocities into the path planning to assist in solving the 3-D motion planning problem of steering an AUV between two selected locations. We present simulation results for tracking a fresh water plume by use of our algorithm. Additionally, we present experimental results from field trials that test the skill of the model used as well as the incorporation of the model predictions into an AUV trajectory planner. These results indicate a modest, but measurable, improvement in surfacing error when the model predictions are incorporated into the planner.

This work addresses mission planning for autonomous underwater gliders based on predictions of an uncertain, time-varying current field. Glider submersibles are highly sensitive to prevailing currents so mission planners must account for ocean tides and eddies. Previous work in variable-current path planning assumes that current predictions are perfect, but in practice these forecasts may be inaccurate. Here we evaluate plan fragility using empirical tests on historical ocean forecasts for which followup data is available. We present methods for glider path planning and control in a time-varying current field. A case study scenario in the Southern California Bight uses current predictions drawn from the Regional Ocean Monitoring System (ROMS).

JPL's Remote Interactive Visualization and Analysis System (RIVA) is described in detail. The RIVA system integrates workstation graphics, massively parallel computing technology, and gigabit communication networks to provide a flexible interactive environment for scientific data perusal, analysis, and visualization, RIVA's kernel is a highly scalable parallel perspective renderer tailored especially for the demands of large datasets beyond the sensible reach of workstations. Early experience with using RIVA to interactively explore and process multivariate, multiresolution datasets is reported; several examples using data from a variety of remote sensing instruments are discussed in detail and the results shown. Particular attention is placed on describing the algorithmic details of RIVA's parallel renderer kernel, with emphasis on the key aspects of achieving the algorithm's overall scalability. The paper summarizes the performance achieved for machine sizes up to more than 500 nodes and for initial input image/terrain bases in the 2 Gbyte range.