Search

This paper presents a novel approach utilizing a scalable neural decoder application-specific integrated circuit (ASIC) based on metal oxide memristors in a 180nm CMOS technology. The ASIC architecture employs in-memory computing with memristor crossbars for efficient vector-matrix multiplications (VMM). The ASIC decoder architecture includes an input layer implemented with a VMM and an analog sigmoid activation function, a recurrent layer with analog memory, and an output layer with a VMM and a threshold activation function. Cryogenic characterization of the ASIC is conducted, demonstrating its performance at both room temperature and cryogenic temperatures down to 1.2K. Results indicate stable activation function shapes and pulse responses at cryogenic temperatures. Moreover, power consumption measurements reveal consistent behavior at room and cryogenic temperatures. Overall, this study lays the foundation for developing efficient and scalable neural decoders for quantum error correction in cryogenic environments.

Recent advances in prompting techniques and multi-agent systems for Large Language Models (LLMs) have produced increasingly complex approaches. However, we lack a framework for characterizing and comparing prompting techniques or understanding their relationship to multi-agent LLM systems. This position paper introduces and explains the concepts of linear contexts (a single, continuous sequence of interactions) and non-linear contexts (branching or multi-path) in LLM systems. These concepts enable the development of an agent-centric projection of prompting techniques, a framework that can reveal deep connections between prompting strategies and multi-agent systems. We propose three conjectures based on this framework: (1) results from non-linear prompting techniques can predict outcomes in equivalent multi-agent systems, (2) multi-agent system architectures can be replicated through single-LLM prompting techniques that simulate equivalent interaction patterns, and (3) these equivalences suggest novel approaches for generating synthetic training data. We argue that this perspective enables systematic cross-pollination of research findings between prompting and multi-agent domains, while providing new directions for improving both the design and training of future LLM systems., Comment: 8 pages, 5 figures. Accepted at ICAART 2025. Derived from an early draft at 2312.17601. arXiv admin note: substantial text overlap with arXiv:2312.17601

Recent advancements in prompting techniques for Large Language Models (LLMs) have improved their reasoning, planning, and action abilities. This paper examines these prompting techniques through the lens of model predictive control (MPC). We show that LLMs act as implicit planning cost function minimizers when planning prompts are used. Under our framework we demonstrate that LLM planning performance can be improved further by incorporating real planning cost functions and evaluators.

Jammed (mechanically rigid) polydisperse circular-disk packings in two dimensions (2D) are popular models for structural glass formers. Maximally random jammed (MRJ) states, which are the most disordered packings subject to strict jamming, have been shown to be hyperuniform. The characterization of the hyperuniformity of MRJ circular-disk packings has covered only a very small part of the possible parameter space for the disk-size distributions. Hyperuniform heterogeneous media are those that anomalously suppress large-scale volume-fraction fluctuations compared to those in typical disordered systems, i.e., their spectral densities $\tilde{\chi}_{_V}(\mathbf{k})$ tend to 0 as the wavenumber $k\equiv|\mathbf{k}|$ tends to 0 and are described by the power-law $\tilde{\chi}_{_V}(\mathbf{k})\sim k^{\alpha}$ as $k\rightarrow0$ where $\alpha$ is the hyperuniformity scaling exponent. In this work, we generate and characterize the structure of strictly jammed binary circular-disk packings with disk-size ratio $\beta$ and a molar ratio of 1:1. By characterizing the rattler fraction, the fraction of isostatic configurations in an ensemble with fixed $\beta$, and the $n$-fold orientational order metrics of ensembles of packings with a wide range of $\beta$, we show that size ratios $1.2\lesssim \beta\lesssim 2.0$ produce MRJ-like states, which we show are the most disordered packings according to several criteria. Using the large-length-scale scaling of the volume fraction variance, we extract $\alpha$ from these packings, and find the function $\alpha(\beta)$ is maximized at $\beta$ = 1.4 (with $\alpha = 0.450\pm0.002$) within the range $1.2\leq\beta\leq2.0$, and decreases rapidly outside of this range. The results from this work can inform the experimental design of disordered hyperuniform thin-film materials with tunable degrees of orientational and translational disorder. (abridged), Comment: 13 pages, 8 figures

The rapid adoption of generative artificial intelligence (GenAI) in research presents both opportunities and ethical challenges that should be carefully navigated. Although GenAI tools can enhance research efficiency through automation of tasks such as literature review and data analysis, their use raises concerns about aspects such as data accuracy, privacy, bias, and research integrity. This paper develops the ETHICAL framework, which is a practical guide for responsible GenAI use in research. Employing a constructivist case study examining multiple GenAI tools in real research contexts, the framework consists of seven key principles: Examine policies and guidelines, Think about social impacts, Harness understanding of the technology, Indicate use, Critically engage with outputs, Access secure versions, and Look at user agreements. Applying these principles will enable researchers to uphold research integrity while leveraging GenAI benefits. The framework addresses a critical gap between awareness of ethical issues and practical action steps, providing researchers with concrete guidance for ethical GenAI integration. This work has implications for research practice, institutional policy development, and the broader academic community while adapting to an AI-enhanced research landscape. The ETHICAL framework can serve as a foundation for developing AI literacy in academic settings and promoting responsible innovation in research methodologies., Comment: 28 pages, 1 figure

In this work, we investigate the problem of the boundedness of the Gross' solutions of the planar Skorokhod embedding problem, where we show that the solution is bounded under some mild conditions on the underlying probability distribution.

Solar energetic particles (SEPs) associated with shocks driven by fast coronal mass ejections (CMEs) or shocks developed by corotating interaction regions (CIRs) often extend to high energies, and are thus key elements of space weather. The PUNCH mission, set to be launched in 2025, is equipped with photometric that enables 3D tracking of solar wind structures in the interplanetary space through polarized light. Tracking techniques are used to estimate speeds and speed gradients of solar structures, including speed jumps at fast shocks. We report on a strong and a robust relation between the shock speed jump magnitude at CME and CIR shocks and the peak fluxes of associated energetic particles from the analysis of 59 CME-driven shocks and 74 CIRs observed by Wind/STEP between 1997-2023. We demonstrate that this relation, along with PUNCH anticipated observations of solar structures can be used to forecast shock-associated particle events close to the Sun; thus, advancing and providing a crucial input to forecasting of SEP fluxes in the heliosphere.

The synthesis of heavy elements in supernovae is affected by low-energy (n,p) and (p,n) reactions on unstable nuclei, yet experimental data on such reaction rates are scarce. The SECAR (SEparator for CApture Reactions) recoil separator at FRIB (Facility for Rare Isotope Beams) was originally designed to measure astrophysical reactions that change the mass of a nucleus significantly. We used a novel approach that integrates machine learning with ion-optical simulations to find an ion-optical solution for the separator that enables the measurement of (p,n) reactions, despite the reaction leaving the mass of the nucleus nearly unchanged. A new measurement of the $^{58}$Fe(p,n)$^{58}$Co reaction in inverse kinematics with a 3.66$\pm$0.12 MeV/nucleon $^{58}$Fe beam (corresponding to 3.69$\pm$0.12 MeV proton energy in normal kinematics) yielded a cross-section of 20.3$\pm$6.3 mb and served as a benchmark for the new technique demonstrating its effectiveness in achieving the required performance criteria. This novel approach marks a significant advancement in experimental nuclear astrophysics, as it paves the way for studying astrophysically important (p,n) reactions on unstable nuclei produced at FRIB.

We examined the origin of 3He abundance enhancement in 23 high-energy (25-50 MeV) solar proton events that coincide with 3He-rich periods detected by ACE ULEIS in 1997-2021. In seven events, 3He enhancement was due to 3He leftover from preceding events or independent 3He events occurring during proton events. One event is the most likely impulsive (3He-rich), and another is unclear. Reaccelerated remnant flare material was the most probable cause of 3He enhancements in the remaining 14 proton events. Imaging observations showed coronal jets in the parent active regions in six of these 14 events. Remarkably, the highest 3He/4He occurred in events with jets, implying their contribution to 3He enhancement., Comment: 6 pages, 4 figures, 2 tables, accepted for publication in IAU Symposium 388 Proceedings

Contextual bandits have become an increasingly popular solution for personalized recommender systems. Despite their growing use, the interpretability of these systems remains a significant challenge, particularly for the often non-expert operators tasked with ensuring their optimal performance. In this paper, we address this challenge by designing a new interface to explain to domain experts the underlying behaviour of a bandit. Central is a metric we term "value gain", a measure derived from off-policy evaluation to quantify the real-world impact of sub-components within a bandit. We conduct a qualitative user study to evaluate the effectiveness of our interface. Our findings suggest that by carefully balancing technical rigour with accessible presentation, it is possible to empower non-experts to manage complex machine learning systems. We conclude by outlining guiding principles that other researchers should consider when building similar such interfaces in future., Comment: 10 pages, 1 figure, Accepted at the IntRS 24 workshop, co-located with ACM RecSys 24

The complementary information found in different modalities of patient data can aid in more accurate modelling of a patient's disease state and a better understanding of the underlying biological processes of a disease. However, the analysis of multi-modal, multi-omics data presents many challenges, including high dimensionality and varying size, statistical distribution, scale and signal strength between modalities. In this work we compare the performance of a variety of ensemble machine learning algorithms that are capable of late integration of multi-class data from different modalities. The ensemble methods and their variations tested were i) a voting ensemble, with hard and soft vote, ii) a meta learner, iii) a multi-modal Adaboost model using a hard vote, a soft vote and a meta learner to integrate the modalities on each boosting round, the PB-MVBoost model and a novel application of a mixture of experts model. These were compared to simple concatenation as a baseline. We examine these methods using data from an in-house study on hepatocellular carcinoma (HCC), along with four validation datasets on studies from breast cancer and irritable bowel disease (IBD). Using the area under the receiver operating curve as a measure of performance we develop models that achieve a performance value of up to 0.85 and find that two boosted methods, PB-MVBoost and Adaboost with a soft vote were the overall best performing models. We also examine the stability of features selected, and the size of the clinical signature determined. Finally, we provide recommendations for the integration of multi-modal multi-class data., Comment: 21 pages, 5 figures

Multi-armed Bandits (MABs) are increasingly employed in online platforms and e-commerce to optimize decision making for personalized user experiences. In this work, we focus on the Contextual Bandit problem with linear rewards, under conditions of sparsity and batched data. We address the challenge of fairness by excluding irrelevant features from decision-making processes using a novel algorithm, Online Batched Sequential Inclusion (OBSI), which sequentially includes features as confidence in their impact on the reward increases. Our experiments on synthetic data show the superior performance of OBSI compared to other algorithms in terms of regret, relevance of features used, and compute., Comment: 4 pages, 4 figures, Accepted at the CONSEQUENCES 24 workshop, co-located with ACM RecSys 24

This paper presents a curriculum, "AI Literacy for All," to promote an interdisciplinary understanding of AI, its socio-technical implications, and its practical applications for all levels of education. With the rapid evolution of artificial intelligence (AI), there is a need for AI literacy that goes beyond the traditional AI education curriculum. AI literacy has been conceptualized in various ways, including public literacy, competency building for designers, conceptual understanding of AI concepts, and domain-specific upskilling. Most of these conceptualizations were established before the public release of Generative AI (Gen-AI) tools like ChatGPT. AI education has focused on the principles and applications of AI through a technical lens that emphasizes the mastery of AI principles, the mathematical foundations underlying these technologies, and the programming and mathematical skills necessary to implement AI solutions. In AI Literacy for All, we emphasize a balanced curriculum that includes technical and non-technical learning outcomes to enable a conceptual understanding and critical evaluation of AI technologies in an interdisciplinary socio-technical context. The paper presents four pillars of AI literacy: understanding the scope and technical dimensions of AI, learning how to interact with Gen-AI in an informed and responsible way, the socio-technical issues of ethical and responsible AI, and the social and future implications of AI. While it is important to include all learning outcomes for AI education in a Computer Science major, the learning outcomes can be adjusted for other learning contexts, including, non-CS majors, high school summer camps, the adult workforce, and the public. This paper advocates for a shift in AI literacy education to offer a more interdisciplinary socio-technical approach as a pathway to broaden participation in AI., Comment: Published at 2024 IEEE Frontiers in Education Conference

Formulating order metrics that sensitively quantify the degree of order/disorder in many-particle systems in $d$-dimensional Euclidean space $\mathbb{R}^d$ across length scales is an outstanding challenge in physics, chemistry, and materials science. Since an infinite set of $n$-particle correlation functions is required to fully characterize a system, one must settle for a reduced set of structural information, in practice. We initiate a program to use the local number variance $\sigma_N^2(R)$ associated with a spherical sampling window of radius $R$ (which encodes pair correlations) and an integral measure derived from it $\Sigma_N(R_i,R_j)$ that depends on two specified radial distances $R_i$ and $R_j$. Across the first three space dimensions ($d = 1,2,3$), we find these metrics can sensitively describe and categorize the degree of order/disorder of 41 different models of antihyperuniform, nonhyperuniform, disordered hyperuniform, and ordered hyperuniform many-particle systems at a specified length scale $R$. Using our local variance metrics, we demonstrate the importance of assessing order/disorder with respect to a specific value of $R$. These local order metrics could also aid in the inverse design of structures with prescribed length-scale-specific degrees of order/disorder that yield desired physical properties. In future work, it would be fruitful to explore the use of higher-order moments of the number of points within a spherical window of radius $R$ [S. Torquato {\it et al}., Phys. Rev. X, \textbf{11}, 021028 (2021)] to devise even more sensitive order metrics., Comment: 19 pages, 8 figures

Process-based hydrologic models are invaluable tools for understanding the terrestrial water cycle and addressing modern water resources problems. However, many hydrologic models are computationally expensive and, depending on the resolution and scale, simulations can take on the order of hours to days to complete. While techniques such as uncertainty quantification and optimization have become valuable tools for supporting management decisions, these analyses typically require hundreds of model simulations, which are too computationally expensive to perform with a process-based hydrologic model. To address this gap, we propose a hybrid modeling workflow in which a process-based model is used to generate an initial set of simulations and a machine learning (ML) surrogate model is then trained to perform the remaining simulations required for downstream analysis. As a case study, we apply this workflow to simulations of variably saturated groundwater flow at a prospective managed aquifer recharge (MAR) site. We compare the accuracy and computational efficiency of several ML architectures, including deep convolutional networks, recurrent neural networks, vision transformers, and networks with Fourier transforms. Our results demonstrate that ML surrogate models can achieve under 10% mean absolute percentage error and yield order-of-magnitude runtime savings over processed-based models. We also offer practical recommendations for training hydrologic surrogate models, including implementing data normalization to improve accuracy, using a normalized loss function to improve training stability and downsampling input features to decrease memory requirements., Comment: 32 pages, 14 figures, 11 tables

Nearest neighbor machine translation is a successful approach for fast domain adaption, which interpolates the pre-trained transformers with domain-specific token-level k-nearest-neighbor (kNN) retrieval without retraining. Despite kNN MT's success, searching large reference corpus and fixed interpolation between the kNN and pre-trained model led to computational complexity and translation quality challenges. Among other papers, Dai et al. proposed methods to obtain a small number of reference samples dynamically for which they introduced a distance-aware interpolation method using an equation that includes free parameters. This paper proposes a simply trainable nearest neighbor machine translation and carry out inference experiments on GPU. Similar to Dai et al., we first adaptively construct a small datastore for each input sentence. Second, we train a single-layer network for the interpolation coefficient between the knnMT and pre-trained result to automatically interpolate in different domains. Experimental results on different domains show that our proposed method either improves or sometimes maintain the translation quality of methods in Dai et al. while being automatic. In addition, our GPU inference results demonstrate that knnMT can be integrated into GPUs with a drop of only 5% in terms of speed., Comment: 6 pages