Search

Uncertainty quantification enables users to assess the reliability of responses generated by large language models (LLMs). We present a novel Question Rephrasing technique to evaluate the input uncertainty of LLMs, which refers to the uncertainty arising from equivalent variations of the inputs provided to LLMs. This technique is integrated with sampling methods that measure the output uncertainty of LLMs, thereby offering a more comprehensive uncertainty assessment. We validated our approach on property prediction and reaction prediction for molecular chemistry tasks.

We present a comprehensive evaluation of proprietary and open-weights large language models using the first astronomy-specific benchmarking dataset. This dataset comprises 4,425 multiple-choice questions curated from the Annual Review of Astronomy and Astrophysics, covering a broad range of astrophysical topics. Our analysis examines model performance across various astronomical subfields and assesses response calibration, crucial for potential deployment in research environments. Claude-3.5-Sonnet outperforms competitors by up to 4.6 percentage points, achieving 85.0% accuracy. For proprietary models, we observed a universal reduction in cost every 3-to-12 months to achieve similar score in this particular astronomy benchmark. Open-source models have rapidly improved, with LLaMA-3-70b (80.6%) and Qwen-2-72b (77.7%) now competing with some of the best proprietary models. We identify performance variations across topics, with non-English-focused models generally struggling more in exoplanet-related fields, stellar astrophysics, and instrumentation related questions. These challenges likely stem from less abundant training data, limited historical context, and rapid recent developments in these areas. This pattern is observed across both open-weights and proprietary models, with regional dependencies evident, highlighting the impact of training data diversity on model performance in specialized scientific domains. Top-performing models demonstrate well-calibrated confidence, with correlations above 0.9 between confidence and correctness, though they tend to be slightly underconfident. The development for fast, low-cost inference of open-weights models presents new opportunities for affordable deployment in astronomy. The rapid progress observed suggests that LLM-driven research in astronomy may become feasible in the near future., Comment: 45 pages, 12 figures, 7 tables. Submitted to ApJ. Comments welcome. AstroMLab homepage: https://astromlab.org/

Numerical models of the ocean and ice sheets are crucial for understanding and simulating the impact of greenhouse gases on the global climate. Oceanic processes affect phenomena such as hurricanes, extreme precipitation, and droughts. Ocean models rely on subgrid-scale parameterizations that require calibration and often significantly affect model skill. When model sensitivities to parameters can be computed by using approaches such as automatic differentiation, they can be used for such calibration toward reducing the misfit between model output and data. Because the SOMA model code is challenging to differentiate, we have created neural network-based surrogates for estimating the sensitivity of the ocean model to model parameters. We first generated perturbed parameter ensemble data for an idealized ocean model and trained three surrogate neural network models. The neural surrogates accurately predicted the one-step forward ocean dynamics, of which we then computed the parametric sensitivity.

Information theoretic quantities play a central role in machine learning. The recent surge in the complexity of data and models has increased the demand for accurate estimation of these quantities. However, as the dimension grows the estimation presents significant challenges, with existing methods struggling already in relatively low dimensions. To address this issue, in this work, we introduce $\texttt{REMEDI}$ for efficient and accurate estimation of differential entropy, a fundamental information theoretic quantity. The approach combines the minimization of the cross-entropy for simple, adaptive base models and the estimation of their deviation, in terms of the relative entropy, from the data density. Our approach demonstrates improvement across a broad spectrum of estimation tasks, encompassing entropy estimation on both synthetic and natural data. Further, we extend important theoretical consistency results to a more generalized setting required by our approach. We illustrate how the framework can be naturally extended to information theoretic supervised learning models, with a specific focus on the Information Bottleneck approach. It is demonstrated that the method delivers better accuracy compared to the existing methods in Information Bottleneck. In addition, we explore a natural connection between $\texttt{REMEDI}$ and generative modeling using rejection sampling and Langevin dynamics., Comment: To appear in ICML 2024. 30 pages, 18 figures

Weather forecasting is a fundamental problem for anticipating and mitigating the impacts of climate change. Recently, data-driven approaches for weather forecasting based on deep learning have shown great promise, achieving accuracies that are competitive with operational systems. However, those methods often employ complex, customized architectures without sufficient ablation analysis, making it difficult to understand what truly contributes to their success. Here we introduce Stormer, a simple transformer model that achieves state-of-the-art performance on weather forecasting with minimal changes to the standard transformer backbone. We identify the key components of Stormer through careful empirical analyses, including weather-specific embedding, randomized dynamics forecast, and pressure-weighted loss. At the core of Stormer is a randomized forecasting objective that trains the model to forecast the weather dynamics over varying time intervals. During inference, this allows us to produce multiple forecasts for a target lead time and combine them to obtain better forecast accuracy. On WeatherBench 2, Stormer performs competitively at short to medium-range forecasts and outperforms current methods beyond 7 days, while requiring orders-of-magnitude less training data and compute. Additionally, we demonstrate Stormer's favorable scaling properties, showing consistent improvements in forecast accuracy with increases in model size and training tokens. Code and checkpoints are available at https://github.com/tung-nd/stormer., Comment: Neural Information Processing Systems (NeurIPS 2024)

Modeling is crucial to understanding the effect of greenhouse gases, warming, and ice sheet melting on the ocean. At the same time, ocean processes affect phenomena such as hurricanes and droughts. Parameters in the models that cannot be physically measured have a significant effect on the model output. For an idealized ocean model, we generated perturbed parameter ensemble data and trained surrogate neural network models. The neural surrogates accurately predicted the one-step forward dynamics, of which we then computed the parametric sensitivity.

Machine learning-based performance models are increasingly being used to build critical job scheduling and application optimization decisions. Traditionally, these models assume that data distribution does not change as more samples are collected over time. However, owing to the complexity and heterogeneity of production HPC systems, they are susceptible to hardware degradation, replacement, and/or software patches, which can lead to drift in the data distribution that can adversely affect the performance models. To this end, we develop continually learning performance models that account for the distribution drift, alleviate catastrophic forgetting, and improve generalizability. Our best model was able to retain accuracy, regardless of having to learn the new distribution of data inflicted by system changes, while demonstrating a 2x improvement in the prediction accuracy of the whole data sequence in comparison to the naive approach., Comment: Presented at Workshop on Machine Learning for Systems at 36th Conference on Neural Information Processing Systems (NeurIPS 2022)

The ability to learn continuously from an incoming data stream without catastrophic forgetting is critical to designing intelligent systems. Many approaches to continual learning rely on stochastic gradient descent and its variants that employ global error updates, and hence need to adopt strategies such as memory buffers or replay to circumvent its stability, greed, and short-term memory limitations. To address this limitation, we have developed a biologically inspired lightweight neural network architecture that incorporates synaptic plasticity mechanisms and neuromodulation and hence learns through local error signals to enable online continual learning without stochastic gradient descent. Our approach leads to superior online continual learning performance on Split-MNIST, Split-CIFAR-10, and Split-CIFAR-100 datasets compared to other memory-constrained learning approaches and matches that of the state-of-the-art memory-intensive replay-based approaches. We further demonstrate the effectiveness of our approach by integrating key design concepts into other backpropagation-based continual learning algorithms, significantly improving their accuracy. Our results provide compelling evidence for the importance of incorporating biological principles into machine learning models and offer insights into how we can leverage them to design more efficient and robust systems for online continual learning.

In this work we have extended AutoML inspired approaches to the exploration and optimization of neuromorphic architectures. Through the integration of a parallel asynchronous model-based search approach with a simulation framework to simulate spiking architectures, we are able to efficiently explore the configuration space of neuromorphic architectures and identify the subset of conditions leading to the highest performance in a targeted application. We have demonstrated this approach on an exemplar case of real time, on-chip learning application. Our results indicate that we can effectively use optimization approaches to optimize complex architectures, therefore providing a viable pathway towards application-driven codesign.

Artificial intelligence methods show great promise in increasing the quality and speed of work with large astronomical datasets, but the high complexity of these methods leads to the extraction of dataset-specific, non-robust features. Therefore, such methods do not generalize well across multiple datasets. We present a universal domain adaptation method, \textit{DeepAstroUDA}, as an approach to overcome this challenge. This algorithm performs semi-supervised domain adaptation and can be applied to datasets with different data distributions and class overlaps. Non-overlapping classes can be present in any of the two datasets (the labeled source domain, or the unlabeled target domain), and the method can even be used in the presence of unknown classes. We apply our method to three examples of galaxy morphology classification tasks of different complexities ($3$-class and $10$-class problems), with anomaly detection: 1) datasets created after different numbers of observing years from a single survey (LSST mock data of $1$ and $10$ years of observations); 2) data from different surveys (SDSS and DECaLS); and 3) data from observing fields with different depths within one survey (wide field and Stripe 82 deep field of SDSS). For the first time, we demonstrate the successful use of domain adaptation between very discrepant observational datasets. \textit{DeepAstroUDA} is capable of bridging the gap between two astronomical surveys, increasing classification accuracy in both domains (up to $40\%$ on the unlabeled data), and making model performance consistent across datasets. Furthermore, our method also performs well as an anomaly detection algorithm and successfully clusters unknown class samples even in the unlabeled target dataset., Comment: Accepted in Machine Learning Science and Technology (MLST); 24 pages, 14 figures

Despite the advancement of machine learning techniques in recent years, state-of-the-art systems lack robustness to "real world" events, where the input distributions and tasks encountered by the deployed systems will not be limited to the original training context, and systems will instead need to adapt to novel distributions and tasks while deployed. This critical gap may be addressed through the development of "Lifelong Learning" systems that are capable of 1) Continuous Learning, 2) Transfer and Adaptation, and 3) Scalability. Unfortunately, efforts to improve these capabilities are typically treated as distinct areas of research that are assessed independently, without regard to the impact of each separate capability on other aspects of the system. We instead propose a holistic approach, using a suite of metrics and an evaluation framework to assess Lifelong Learning in a principled way that is agnostic to specific domains or system techniques. Through five case studies, we show that this suite of metrics can inform the development of varied and complex Lifelong Learning systems. We highlight how the proposed suite of metrics quantifies performance trade-offs present during Lifelong Learning system development - both the widely discussed Stability-Plasticity dilemma and the newly proposed relationship between Sample Efficient and Robust Learning. Further, we make recommendations for the formulation and use of metrics to guide the continuing development of Lifelong Learning systems and assess their progress in the future., Comment: To appear in Neural Networks

We have developed a model for online continual or lifelong reinforcement learning (RL) inspired on the insect brain. Our model leverages the offline training of a feature extraction and a common general policy layer to enable the convergence of RL algorithms in online settings. Sharing a common policy layer across tasks leads to positive backward transfer, where the agent continuously improved in older tasks sharing the same underlying general policy. Biologically inspired restrictions to the agent's network are key for the convergence of RL algorithms. This provides a pathway towards efficient online RL in resource-constrained scenarios.

In the era of big astronomical surveys, our ability to leverage artificial intelligence algorithms simultaneously for multiple datasets will open new avenues for scientific discovery. Unfortunately, simply training a deep neural network on images from one data domain often leads to very poor performance on any other dataset. Here we develop a Universal Domain Adaptation method DeepAstroUDA, capable of performing semi-supervised domain alignment that can be applied to datasets with different types of class overlap. Extra classes can be present in any of the two datasets, and the method can even be used in the presence of unknown classes. For the first time, we demonstrate the successful use of domain adaptation on two very different observational datasets (from SDSS and DECaLS). We show that our method is capable of bridging the gap between two astronomical surveys, and also performs well for anomaly detection and clustering of unknown data in the unlabeled dataset. We apply our model to two examples of galaxy morphology classification tasks with anomaly detection: 1) classifying spiral and elliptical galaxies with detection of merging galaxies (three classes including one unknown anomaly class); 2) a more granular problem where the classes describe more detailed morphological properties of galaxies, with the detection of gravitational lenses (ten classes including one unknown anomaly class)., Comment: 3 figures, 1 table; accepted to Machine Learning and the Physical Sciences - Workshop at the 36th conference on Neural Information Processing Systems (NeurIPS)

Robust machine learning models with accurately calibrated uncertainties are crucial for safety-critical applications. Probabilistic machine learning and especially the Bayesian formalism provide a systematic framework to incorporate robustness through the distributional estimates and reason about uncertainty. Recent works have shown that approximate inference approaches that take the weight space uncertainty of neural networks to generate ensemble prediction are the state-of-the-art. However, architecture choices have mostly been ad hoc, which essentially ignores the epistemic uncertainty from the architecture space. To this end, we propose a Unified probabilistic architecture and weight ensembling Neural Architecture Search (UraeNAS) that leverages advances in probabilistic neural architecture search and approximate Bayesian inference to generate ensembles form the joint distribution of neural network architectures and weights. The proposed approach showed a significant improvement both with in-distribution (0.86% in accuracy, 42% in ECE) CIFAR-10 and out-of-distribution (2.43% in accuracy, 30% in ECE) CIFAR-10-C compared to the baseline deterministic approach.

Distributed data storage services tailored to specific applications have grown popular in the high-performance computing (HPC) community as a way to address I/O and storage challenges. These services offer a variety of specific interfaces, semantics, and data representations. They also expose many tuning parameters, making it difficult for their users to find the best configuration for a given workload and platform. To address this issue, we develop a novel variational-autoencoder-guided asynchronous Bayesian optimization method to tune HPC storage service parameters. Our approach uses transfer learning to leverage prior tuning results and use a dynamically updated surrogate model to explore the large parameter search space in a systematic way. We implement our approach within the DeepHyper open-source framework, and apply it to the autotuning of a high-energy physics workflow on Argonne's Theta supercomputer. We show that our transfer-learning approach enables a more than $40\times$ search speedup over random search, compared with a $2.5\times$ to $10\times$ speedup when not using transfer learning. Additionally, we show that our approach is on par with state-of-the-art autotuning frameworks in speed and outperforms them in resource utilization and parallelization capabilities., Comment: Accepted at IEEE Cluster 2022

Deep ensembles have emerged as a powerful technique for improving predictive performance and enhancing model robustness across various applications by leveraging model diversity. However, traditional deep ensemble methods are often computationally expensive and rely on deterministic models, which may limit their flexibility. Additionally, while sparse subnetworks of dense models have shown promise in matching the performance of their dense counterparts and even enhancing robustness, existing methods for inducing sparsity typically incur training costs comparable to those of training a single dense model, as they either gradually prune the network during training or apply thresholding post-training. In light of these challenges, we propose an approach for sequential ensembling of dynamic Bayesian neural subnetworks that consistently maintains reduced model complexity throughout the training process while generating diverse ensembles in a single forward pass. Our approach involves an initial exploration phase to identify high-performing regions within the parameter space, followed by multiple exploitation phases that take advantage of the compactness of the sparse model. These exploitation phases quickly converge to different minima in the energy landscape, corresponding to high-performing subnetworks that together form a diverse and robust ensemble. We empirically demonstrate that our proposed approach outperforms traditional dense and sparse deterministic and Bayesian ensemble models in terms of prediction accuracy, uncertainty estimation, out-of-distribution detection, and adversarial robustness.

I/O efficiency is crucial to productivity in scientific computing, but the increasing complexity of the system and the applications makes it difficult for practitioners to understand and optimize I/O behavior at scale. Data-driven machine learning-based I/O throughput models offer a solution: they can be used to identify bottlenecks, automate I/O tuning, or optimize job scheduling with minimal human intervention. Unfortunately, current state-of-the-art I/O models are not robust enough for production use and underperform after being deployed. We analyze multiple years of application, scheduler, and storage system logs on two leadership-class HPC platforms to understand why I/O models underperform in practice. We propose a taxonomy consisting of five categories of I/O modeling errors: poor application and system modeling, inadequate dataset coverage, I/O contention, and I/O noise. We develop litmus tests to quantify each category, allowing researchers to narrow down failure modes, enhance I/O throughput models, and improve future generations of HPC logging and analysis tools.

The information bottleneck framework provides a systematic approach to learning representations that compress nuisance information in the input and extract semantically meaningful information about predictions. However, the choice of a prior distribution that fixes the dimensionality across all the data can restrict the flexibility of this approach for learning robust representations. We present a novel sparsity-inducing spike-slab categorical prior that uses sparsity as a mechanism to provide the flexibility that allows each data point to learn its own dimension distribution. In addition, it provides a mechanism for learning a joint distribution of the latent variable and the sparsity and hence can account for the complete uncertainty in the latent space. Through a series of experiments using in-distribution and out-of-distribution learning scenarios on the MNIST, CIFAR-10, and ImageNet data, we show that the proposed approach improves accuracy and robustness compared to traditional fixed-dimensional priors, as well as other sparsity induction mechanisms for latent variable models proposed in the literature.

Gaussian Process Regression (GPR) is a Bayesian method for inferring profiles based on input data. The technique is increasing in popularity in the fusion community due to its many advantages over traditional fitting techniques including intrinsic uncertainty quantification and robustness to over-fitting. This work investigates the use of a new method, the change-point method, for handling the varying length scales found in different tokamak regimes. The use of the Student's t-distribution for the Bayesian likelihood probability is also investigated and shown to be advantageous in providing good fits in profiles with many outliers. To compare different methods, synthetic data generated from analytic profiles is used to create a database enabling a quantitative statistical comparison of which methods perform the best. Using a full Bayesian approach with the change-point method, Mat\'ern kernel for the prior probability, and Student's t-distribution for the likelihood is shown to give the best results., Comment: submitted to PPCF

With increased adoption of supervised deep learning methods for processing and analysis of cosmological survey data, the assessment of data perturbation effects (that can naturally occur in the data processing and analysis pipelines) and the development of methods that increase model robustness are increasingly important. In the context of morphological classification of galaxies, we study the effects of perturbations in imaging data. In particular, we examine the consequences of using neural networks when training on baseline data and testing on perturbed data. We consider perturbations associated with two primary sources: 1) increased observational noise as represented by higher levels of Poisson noise and 2) data processing noise incurred by steps such as image compression or telescope errors as represented by one-pixel adversarial attacks. We also test the efficacy of domain adaptation techniques in mitigating the perturbation-driven errors. We use classification accuracy, latent space visualizations, and latent space distance to assess model robustness. Without domain adaptation, we find that processing pixel-level errors easily flip the classification into an incorrect class and that higher observational noise makes the model trained on low-noise data unable to classify galaxy morphologies. On the other hand, we show that training with domain adaptation improves model robustness and mitigates the effects of these perturbations, improving the classification accuracy by 23% on data with higher observational noise. Domain adaptation also increases by a factor of ~2.3 the latent space distance between the baseline and the incorrectly classified one-pixel perturbed image, making the model more robust to inadvertent perturbations., Comment: 20 pages, 6 figures, 5 tables; accepted in MLST

The potential of deep learning based image-to-image translations has recently drawn a lot of attention; one intriguing possibility is that of generating cosmological predictions with a drastic reduction in computational cost. Such an effort requires optimization of neural networks with loss functions beyond low-order statistics like pixel-wise mean square error, and validation of results beyond simple visual comparisons and summary statistics. In order to study learning-based cosmological mappings, we choose a tractable analytical prescription - the Zel'dovich approximation - modeled using U-Net, a convolutional image translation framework. A comprehensive list of metrics is proposed, including higher-order correlation functions, conservation laws, topological indicators, dynamical robustness, and statistical independence of density fields. We find that the U-Net approach does well with some metrics but has difficulties with others. In addition to validating AI approaches using rigorous physical benchmarks, this study motivates advancements in domain-specific optimization schemes for scientific machine learning., Comment: Accepted in Neural Information Processing Systems (NeurIPS) 2021 AI for Science Workshop

Deep learning models are being increasingly adopted in wide array of scientific domains, especially to handle high-dimensionality and volume of the scientific data. However, these models tend to be brittle due to their complexity and overparametrization, especially to the inadvertent adversarial perturbations that can appear due to common image processing such as compression or blurring that are often seen with real scientific data. It is crucial to understand this brittleness and develop models robust to these adversarial perturbations. To this end, we study the effect of observational noise from the exposure time, as well as the worst case scenario of a one-pixel attack as a proxy for compression or telescope errors on performance of ResNet18 trained to distinguish between galaxies of different morphologies in LSST mock data. We also explore how domain adaptation techniques can help improve model robustness in case of this type of naturally occurring attacks and help scientists build more trustworthy and stable models., Comment: Accepted in: Fourth Workshop on Machine Learning and the Physical Sciences (35th Conference on Neural Information Processing Systems; NeurIPS2021); final version

In this community review report, we discuss applications and techniques for fast machine learning (ML) in science -- the concept of integrating power ML methods into the real-time experimental data processing loop to accelerate scientific discovery. The material for the report builds on two workshops held by the Fast ML for Science community and covers three main areas: applications for fast ML across a number of scientific domains; techniques for training and implementing performant and resource-efficient ML algorithms; and computing architectures, platforms, and technologies for deploying these algorithms. We also present overlapping challenges across the multiple scientific domains where common solutions can be found. This community report is intended to give plenty of examples and inspiration for scientific discovery through integrated and accelerated ML solutions. This is followed by a high-level overview and organization of technical advances, including an abundance of pointers to source material, which can enable these breakthroughs., Comment: 66 pages, 13 figures, 5 tables

In astronomy, neural networks are often trained on simulation data with the prospect of being used on telescope observations. Unfortunately, training a model on simulation data and then applying it to instrument data leads to a substantial and potentially even detrimental decrease in model accuracy on the new target dataset. Simulated and instrument data represent different data domains, and for an algorithm to work in both, domain-invariant learning is necessary. Here we employ domain adaptation techniques$-$ Maximum Mean Discrepancy (MMD) as an additional transfer loss and Domain Adversarial Neural Networks (DANNs)$-$ and demonstrate their viability to extract domain-invariant features within the astronomical context of classifying merging and non-merging galaxies. Additionally, we explore the use of Fisher loss and entropy minimization to enforce better in-domain class discriminability. We show that the addition of each domain adaptation technique improves the performance of a classifier when compared to conventional deep learning algorithms. We demonstrate this on two examples: between two Illustris-1 simulated datasets of distant merging galaxies, and between Illustris-1 simulated data of nearby merging galaxies and observed data from the Sloan Digital Sky Survey. The use of domain adaptation techniques in our experiments leads to an increase of target domain classification accuracy of up to ${\sim}20\%$. With further development, these techniques will allow astronomers to successfully implement neural network models trained on simulation data to efficiently detect and study astrophysical objects in current and future large-scale astronomical surveys., Comment: Submitted to MNRAS; 21 pages, 9 figures, 9 tables