Search

Observations of the Milky Way and external galaxies support the idea that large-scale magnetic fields are concentrated in galactic disks, with halo magnetic fields at least an order of magnitude weaker. However, very little is known about the transition between the two. We present the discovery of linearly polarized radio emission at the interface between interacting shells of gas within a well-known grouping of high-velocity clouds (HVCs), the Anti-Center Shell. Faraday rotation of diffuse emission and of background extragalactic compact sources demonstrates an enhancement of the field at the interface. This is the clearest observed example of an HVC altering the large-scale magnetic field at the disk-halo interface and is the first image of magnetic field effects in an HVC. These results demonstrate the possibility of future three-dimensional reconstruction of the Galactic magnetic field and showcase the versatility of the Synthesis Telescope at the Dominion Radio Astrophysical Observatory as one of the few existing telescopes which can exploit this new method of probing Galactic magnetism., Comment: Accepted for publication in the Astrophysical Journal Nov. 13, 2024

In this study, we explore the use of Co-55 as a radioisotope for multiplexed PET (mPET) by utilizing its emission of a prompt gamma-ray in cascade with a positron during decay. We leverage the prompt-gamma signal to generate triple coincidences for a Co-55-labeled tracer, allowing us to distinguish it from a tracer labeled with a pure positron emitter, such as F-18. By employing triple versus double coincidence detection and signal processing methodology, we successfully separate the Co-55 signal from that of F-18. Phantom studies were conducted to establish the correlation between Co-55 double and triple coincidence counts and Co-55 activity. Additionally, we demonstrate the potential for quantifying hot spots within a warm background produced by both Co-55 and F-18 signals in a single PET scan. Finally, we showcase the ability to simultaneously image two tracers in vivo in a single PET session with mouse models of cancer., Comment: 5 pages, 4 figures, submitted IEEE ISBI 2025; v2: clarification on unstable Co-55-labeled antibody

Governments are increasingly employing funding for open source software (OSS) development as a policy lever to support the security of software supply chains, digital sovereignty, economic growth, and national competitiveness in science and innovation, among others. However, the impacts of public funding on OSS development remain poorly understood, with a lack of consensus on how to meaningfully measure them. This gap hampers assessments of the return on public investment and impedes the optimisation of public-interest funding strategies. We address this gap with a toolkit of methodological considerations that may inform such measurements, drawing on prior work on OSS valuations and community health metrics by the Community Health Analytics Open Source Software (CHAOSS) project as well as our first-hand learnings as practitioners tasked with evaluating funding programmes by the Next Generation Internet initiative and the Sovereign Tech Agency. We discuss salient considerations, including the importance of accounting for funding objectives, project life stage and social structure, and regional and organisational cost factors. Next, we present a taxonomy of potential social, economic, and technological impacts that can be both positive and negative, direct and indirect, internal (i.e. within a project) and external (i.e. among a project's ecosystem of dependents and users), and manifest over various time horizons. Furthermore, we discuss the merits and limitations of qualitative, quantitative, and mixed-methods approaches, as well as options for and hazards of estimating multiplier effects. With this toolkit, we contribute to the multi-stakeholder conversation about the value and impacts of funding on OSS developers and society at large., Comment: 23 pages, 3 tables

From CMB polarization data alone we reconstruct the CMB lensing power spectrum, comparable in overall constraining power to previous temperature-based reconstructions, and an unlensed E-mode power spectrum. The observations, taken in 2019 and 2020 with the South Pole Telescope (SPT) and the SPT-3G camera, cover 1500 deg$^2$ at 95, 150, and 220 GHz with arcminute resolution and roughly 4.9$\mu$K-arcmin coadded noise in polarization. The power spectrum estimates, together with systematic parameter estimates and a joint covariance matrix, follow from a Bayesian analysis using the Marginal Unbiased Score Expansion (MUSE) method. The E-mode spectrum at $\ell>2000$ and lensing spectrum at $L>350$ are the most precise to date. Assuming the $\Lambda$CDM model, and using only these SPT data and priors on $\tau$ and absolute calibration from Planck, we find $H_0=66.81\pm0.81$ km/s/Mpc, comparable in precision to the Planck determination and in 5.4$\sigma$ tension with the most precise $H_0$ inference derived via the distance ladder. We also find $S_8=0.850\pm0.017$, providing further independent evidence of a slight tension with low-redshift structure probes. The $\Lambda$CDM model provides a good simultaneous fit to the combined Planck, ACT, and SPT data, and thus passes a powerful test. Combining these CMB datasets with BAO observations, we find that the effective number of neutrino species, spatial curvature, and primordial helium fraction are consistent with standard model values, and that the 95% confidence upper limit on the neutrino mass sum is 0.075 eV. The SPT data are consistent with the somewhat weak preference for excess lensing power seen in Planck and ACT data relative to predictions of the $\Lambda$CDM model. We also detect at greater than 3$\sigma$ the influence of non-linear evolution in the CMB lensing power spectrum and discuss it in the context of the $S_8$ tension.(abridged), Comment: 28 pages, 21 figures + appendices

Many popular open source projects are owned and driven by vendors, and in today's difficult economic climate, those vendors are under increasing pressure from investors to deliver a strong return on their investments. One response to this pressure has been the relicensing of popular open source projects to more restrictive licenses in the hopes of generating more revenue, disrupting the idea of open source as a digital commons. In some cases, relicensing has resulted in a hard fork of the original project. These relicensing events and resulting forks can be disruptive to the organizations and individuals using these open source projects. This research compares and contrasts organizational affiliation data from three case studies based on license changes that resulted in forks: Elasticsearch / OpenSearch, Redis / Valkey, and Terraform / OpenTofu. The research indicates that the forks resulting from these relicensing events have more organizational diversity than the original projects, especially when the forks are created under a neutral foundation, like the Linux Foundation, rather than by a single company., Comment: Presented at OpenForum Academy (OFA) Symposium in Boston, Massachusetts, November 13 & 14, 2024

Deduplication is a major focus for assembling and curating training datasets for large language models (LLM) -- detecting and eliminating additional instances of the same content -- in large collections of technical documents. Unrestrained, duplicates in the training dataset increase training costs and lead to undesirable properties such as memorization in trained models or cheating on evaluation. Contemporary approaches to document-level deduplication are often extremely expensive in both runtime and memory. We propose LSHBloom, an extension to MinhashLSH, which replaces the expensive LSHIndex with lightweight Bloom filters. LSHBloom demonstrates the same deduplication performance as MinhashLSH with only a marginal increase in false positives (as low as 1e-5 in our experiments); demonstrates competitive runtime (270\% faster than MinhashLSH on peS2o); and, crucially, uses just 0.6\% of the disk space required by MinhashLSH to deduplicate peS2o. We demonstrate that this space advantage scales with increased dataset size -- at the extreme scale of several billion documents, LSHBloom promises a 250\% speedup and a 54$\times$ space advantage over traditional MinHashLSH scaling deduplication of text datasets to many billions of documents.

We provide bounds on the tail probabilities for simple procedures that generate random samples _without replacement_, when the probabilities of being selected need not be equal.

The detection of satellite thermal emission at millimeter wavelengths is presented using data from the 3rd-Generation receiver on the South Pole Telescope (SPT-3G). This represents the first reported detection of thermal emission from artificial satellites at millimeter wavelengths. Satellite thermal emission is shown to be detectable at high signal-to-noise on timescales as short as a few tens of milliseconds. An algorithm for downloading orbital information and tracking known satellites given observer constraints and time-ordered observatory pointing is described. Consequences for cosmological surveys and short-duration transient searches are discussed, revealing that the integrated thermal emission from all large satellites does not contribute significantly to the SPT-3G survey intensity map. Measured satellite positions are found to be discrepant from their two-line element (TLE) derived ephemerides up to several arcminutes which may present a difficulty in cross-checking or masking satellites from short-duration transient searches.

The growing number of seismic and elastic attributes poses a challenge, making the full benefit from each attribute in characterizing geological formation very difficult, if not impossible. Various approaches are routinely employed to select the best attributes for specific purposes. Machine learning (ML) algorithms have demonstrated good capabilities in combining appropriate attributes to address reservoir characterization problems. This study aims to use and combine seismic and elastic attributes to detect hydrocarbon-saturated reservoirs, source rock, and seal rocks in the Poseidon field, Offshore Australia. A large number of attributes are extracted from seismic data and from impedance data. Artificial Neural Networks (ANN) are implemented to combine the extracted attributes and convert them into Resistivity volume, and Gamma Ray volume from which Shale probability volume, Sand volume probability volume, Effective Porosity volume, and Gas Chimney Probability Cubes. The cubes are deployed for a detailed analysis of the petroleum system in the area. The produced Shale volume and Resistivity cube helped delineate the seal rock and source rock in the area. Next, the reservoir intervals were identified using Porosity, Shale, and Resistivity volumes. A pre-trained Convolutional Neural Network (CNN) is trained using another carefully selected attribute set to detect subtle faults that hydrocarbons might migrated through from source rock to trap. The integration of all the extracted cubes contributed to find new prospects in the area and assess their geological probability of success. The proposed approach stands out for its multi-physical attribute integration, Machine Learning and Human expertise incorporation, possible applicability to other fields., Comment: 20 pages, 10 figures

We describe a way to decompose the chromatic symmetric function as a positive sum of smaller pieces. We show that these pieces are $e$-positive for cycles. Then we prove that attaching a cycle to a graph preserves the $e$-positivity of these pieces. From this, we prove an $e$-positive formula for graphs of cycles connected at adjacent vertices. We extend these results to graphs formed by connecting a sequence of cycles and cliques., Comment: 28 pages, 12 figures

Carbohydrates, essential biological building blocks, exhibit functional mechanisms tied to their intricate stereochemistry. Subtle stereochemical differences, such as those between the anomers maltose and cellobiose, lead to distinct properties due to their differing glycosidic bonds; the former is digestible by humans, while the latter is not. This underscores the importance of precise structural determination of individual carbohydrate molecules for deeper functional insights. However, their structural complexity and conformational flexibility, combined with the high spatial resolution needed, have hindered direct imaging of carbohydrate stereochemistry. Here, we employ non-contact atomic force microscopy integrated with a data-efficient, multi-fidelity structure search approach accelerated by machine learning integration to determine the precise 3D atomic coordinates of two carbohydrate anomers. We observe that glycosidic bond stereochemistry regulates on-surface chiral selection in carbohydrate self-assemblies. The reconstructed models, validated against experimental data, provide reliable atomic-scale structural evidence, uncovering the origin of on-surface chirality from carbohydrate anomerism. Our study confirms that nc-AFM is a reliable technique for real-space discrimination of carbohydrate stereochemistry at the single-molecule level, providing a pathway for bottom-up investigations into the structure-property relationships of carbohydrates in biological research and materials science.

Training large-scale models under given resources requires careful design of parallelism strategies. In particular, the efficiency notion of critical batch size (CBS), concerning the compromise between time and compute, marks the threshold beyond which greater data parallelism leads to diminishing returns. To operationalize it, we propose a measure of CBS and pre-train a series of auto-regressive language models, ranging from 85 million to 1.2 billion parameters, on the C4 dataset. Through extensive hyper-parameter sweeps and careful control of factors such as batch size, momentum, and learning rate along with its scheduling, we systematically investigate the impact of scale on CBS. Then we fit scaling laws with respect to model and data sizes to decouple their effects. Overall, our results demonstrate that CBS scales primarily with data size rather than model size, a finding we justify theoretically through the analysis of infinite-width limits of neural networks and infinite-dimensional least squares regression. Of independent interest, we highlight the importance of common hyper-parameter choices and strategies for studying large-scale pre-training beyond fixed training durations.

Real-world applications of reinforcement learning often involve environments where agents operate on complex, high-dimensional observations, but the underlying (''latent'') dynamics are comparatively simple. However, outside of restrictive settings such as small latent spaces, the fundamental statistical requirements and algorithmic principles for reinforcement learning under latent dynamics are poorly understood. This paper addresses the question of reinforcement learning under $\textit{general}$ latent dynamics from a statistical and algorithmic perspective. On the statistical side, our main negative result shows that most well-studied settings for reinforcement learning with function approximation become intractable when composed with rich observations; we complement this with a positive result, identifying latent pushforward coverability as a general condition that enables statistical tractability. Algorithmically, we develop provably efficient observable-to-latent reductions -- that is, reductions that transform an arbitrary algorithm for the latent MDP into an algorithm that can operate on rich observations -- in two settings: one where the agent has access to hindsight observations of the latent dynamics [LADZ23], and one where the agent can estimate self-predictive latent models [SAGHCB20]. Together, our results serve as a first step toward a unified statistical and algorithmic theory for reinforcement learning under latent dynamics.

We present spectroscopic properties of 22 Lyman-alpha emitters(LAEs) at z=5.5-6.6 with Lyman-alpha(Lya) luminosity log($L_{Lya}$[$ergs^{-1}$])=42.4-43.5, obtained using VLT/MUSE as part of the Middle Ages Galaxy Properties with Integral Field Spectroscopy(MAGPI) survey. Additionally, we incorporate broad-band photometric data from the Subaru Hyper Suprime-Cam(HSC) for 17 LAEs in our sample. The HSC-y band magnitudes show that our LAEs are UV-bright, with rest-frame absolute UV magnitudes -19.7 < $M_{UV}$ < -23.3. We find that the Lya line width increases with luminosity, and this trend becomes more prominent at z > 6 where Lya lines become significantly broadened (> 260 $kms^{-1}$) at luminosities log($L_{Lya}$[$ergs^{-1}$]) > 43. This broadening is consistent with previous studies, suggesting that these sources are located inside larger ionized bubbles. We observe a slightly elevated ionizing photon production efficiency estimated for LAEs at z > 6, indicating that younger galaxies could be producing more ionizing photons per UV luminosity. A tentative anti-correlation between ionizing photon production efficiency and Lya rest-frame equivalent width is noticed, which could indicate a time delay between production and escape of ionizing photon primarily due to supernovae activity. Furthermore, we find a positive correlation between bubble radius and Lya line width, which again suggests that large ionized bubbles are created around these LAEs, allowing them to self-shield from the scattering effects of the intergalactic medium (IGM). We also detect two closely separated LAEs at z=6.046 (projected spatial separation is 15.92 kpc). The size of their respective bubbles suggests that they likely sit inside a common large ionized region. Such a closely-separated LAE pair increases the size of ionized bubble, potentially allowing a boosted transmission of Lya through neutral IGM. (Abridged), Comment: 14 pages, 6 figures, 2 tables, Accepted for publication in PASA

The Workflows Community Summit gathered 111 participants from 18 countries to discuss emerging trends and challenges in scientific workflows, focusing on six key areas: time-sensitive workflows, AI-HPC convergence, multi-facility workflows, heterogeneous HPC environments, user experience, and FAIR computational workflows. The integration of AI and exascale computing has revolutionized scientific workflows, enabling higher-fidelity models and complex, time-sensitive processes, while introducing challenges in managing heterogeneous environments and multi-facility data dependencies. The rise of large language models is driving computational demands to zettaflop scales, necessitating modular, adaptable systems and cloud-service models to optimize resource utilization and ensure reproducibility. Multi-facility workflows present challenges in data movement, curation, and overcoming institutional silos, while diverse hardware architectures require integrating workflow considerations into early system design and developing standardized resource management tools. The summit emphasized improving user experience in workflow systems and ensuring FAIR workflows to enhance collaboration and accelerate scientific discovery. Key recommendations include developing standardized metrics for time-sensitive workflows, creating frameworks for cloud-HPC integration, implementing distributed-by-design workflow modeling, establishing multi-facility authentication protocols, and accelerating AI integration in HPC workflow management. The summit also called for comprehensive workflow benchmarks, workflow-specific UX principles, and a FAIR workflow maturity model, highlighting the need for continued collaboration in addressing the complex challenges posed by the convergence of AI, HPC, and multi-facility research environments.

Understanding neural networks is crucial to creating reliable and trustworthy deep learning models. Most contemporary research in interpretability analyzes just one model at a time via causal intervention or activation analysis. Yet despite successes, these methods leave significant gaps in our understanding of the training behaviors of neural networks, how their inner representations emerge, and how we can predictably associate model components with task-specific behaviors. Seeking new insights from work in related fields, here we survey literature in the field of model merging, a field that aims to combine the abilities of various neural networks by merging their parameters and identifying task-specific model components in the process. We analyze the model merging literature through the lens of loss landscape geometry, an approach that enables us to connect observations from empirical studies on interpretability, security, model merging, and loss landscape analysis to phenomena that govern neural network training and the emergence of their inner representations. To systematize knowledge in this area, we present a novel taxonomy of model merging techniques organized by their core algorithmic principles. Additionally, we distill repeated empirical observations from the literature in these fields into characterizations of four major aspects of loss landscape geometry: mode convexity, determinism, directedness, and connectivity. We argue that by improving our understanding of the principles underlying model merging and loss landscape geometry, this work contributes to the goal of ensuring secure and trustworthy machine learning in practice.

Applications are increasingly written as dynamic workflows underpinned by an execution framework that manages asynchronous computations across distributed hardware. However, execution frameworks typically offer one-size-fits-all solutions for data flow management, which can restrict performance and scalability. ProxyStore, a middleware layer that optimizes data flow via an advanced pass-by-reference paradigm, has shown to be an effective mechanism for addressing these limitations. Here, we investigate integrating ProxyStore with Dask Distributed, one of the most popular libraries for distributed computing in Python, with the goal of supporting scalable and portable scientific workflows. Dask provides an easy-to-use and flexible framework, but is less optimized for scaling certain data-intensive workflows. We investigate these limitations and detail the technical contributions necessary to develop a robust solution for distributed applications and demonstrate improved performance on synthetic benchmarks and real applications., Comment: To be presented as a demo at the SC24 Workshop on High Performance Python for Science at Scale (HPPSS)

This paper explores the mechanical behaviour of the composite materials used in modern lithium-ion battery electrodes. These contain relatively high modulus active particle inclusions within a two-component matrix of liquid electrolyte which penetrates the pore space within a viscoelastic polymer binder. Deformations are driven by a combination of (i) swelling/contraction of the electrode particles in response to lithium insertion/extraction, (ii) swelling of the binder as it absorbs electrolyte, (iii) external loading and (iv) flow of the electrolyte within the pores. We derive the macroscale response of the composite using systematic multiple scales homomgenisation by exploiting the disparity in lengthscales associated with the size of an electrode particle and the electrode as a whole. The resulting effective model accurately replicates the behaviour of the original model (as is demonstrated by a series of relevant case studies) but, crucially, is markedly {simpler and hence} cheaper to solve. This is significant practical value because it facilitates low-cost, realistic computations of the mechanical states of battery electrodes, thereby allowing model-assisted development of battery designs that are better able to withstand the mechanical abuse encountered in practice and ultimately paving the way for longer-lasting batteries., Comment: 46 pages, 19 figures

Neural ODEs are a recently developed model class that combine the strong model priors of differential equations with the high-capacity function approximation of neural networks. One advantage of Neural ODEs is the potential for memory-efficient training via the continuous adjoint method. However, memory-efficient training comes at the cost of approximate gradients. Therefore, in practice, gradients are often obtained by simply backpropagating through the internal operations of the forward ODE solve - incurring high memory cost. Interestingly, it is possible to construct algebraically reversible ODE solvers that allow for both exact gradients and the memory-efficiency of the continuous adjoint method. Unfortunately, current reversible solvers are low-order and suffer from poor numerical stability. The use of these methods in practice is therefore limited. In this work, we present a class of algebraically reversible solvers that are both high-order and numerically stable. Moreover, any explicit numerical scheme can be made reversible by our method. This construction naturally extends to numerical schemes for Neural CDEs and SDEs., Comment: Preprint

In order to formally verify robotic controllers, we must tackle the inherent uncertainty of sensing and actuation in a physical environment. We can model uncertainty using stochastic hybrid systems, which combine discrete jumps with continuous, stochastic behaviour. The verification of these systems is intractable for state-exploration based approaches, so we instead propose a deductive verification approach. As a first step towards a deductive verification tool, we present a mechanisation of Brownian motion within Isabelle/HOL. For this, we mechanise stochastic kernels and Markov semigroups, which allow us to specify a range of processes with stationary, independent increments. Further, we prove the Kolmogorov-Chentsov theorem, which allows us to construct H\"older continuous modifications of processes that satisfy certain bounds on their expectation. This paves the way for modelling and verifying stochastic hybrid systems in Isabelle/HOL.

In this study, we propose a novel approach to uncover subgroup-specific and subgroup-common latent factors addressing the challenges posed by the heterogeneity of neurological and mental disorders, which hinder disease understanding, treatment development, and outcome prediction. The proposed approach, sparse Group Factor Analysis (GFA) with regularised horseshoe priors, was implemented with probabilistic programming and can uncover associations (or latent factors) among multiple data modalities differentially expressed in sample subgroups. Synthetic data experiments showed the robustness of our sparse GFA by correctly inferring latent factors and model parameters. When applied to the Genetic Frontotemporal Dementia Initiative (GENFI) dataset, which comprises patients with frontotemporal dementia (FTD) with genetically defined subgroups, the sparse GFA identified latent disease factors differentially expressed across the subgroups, distinguishing between "subgroup-specific" latent factors within homogeneous groups and "subgroup common" latent factors shared across subgroups. The latent disease factors captured associations between brain structure and non-imaging variables (i.e., questionnaires assessing behaviour and disease severity) across the different genetic subgroups, offering insights into disease profiles. Importantly, two latent factors were more pronounced in the two more homogeneous FTD patient subgroups (progranulin (GRN) and microtubule-associated protein tau (MAPT) mutation), showcasing the method's ability to reveal subgroup-specific characteristics. These findings underscore the potential of sparse GFA for integrating multiple data modalities and identifying interpretable latent disease factors that can improve the characterization and stratification of patients with neurological and mental health disorders., Comment: 38 pages, 14 figures

Charge-conserving dynamics of non-interacting fermions monitored by local charge measurements have been shown to possess in 1D space only an area-law-entangled phase, with no measurement-induced phase transition (MIPT). This phenomenon was elegantly explained by an effective replica field theory given by a non-linear sigma model featuring large replica symmetries. In this work, we focus on the charge-conserving monitored dynamics of interacting fermions in 1D. Using a unifying replica Keldysh field theory, we show that more dynamical phases and transitions emerge owing to the reduction of replica symmetry when interactions are introduced. The reduced symmetry combines a discrete replica permutation symmetry and the charge-conservation within each replica. The former and its spontaneous breaking govern the MIPT between area-law and volume-law entanglement scalings, which can be described by a separatrix in the weak-coupling renormalization group flow. The replica-resolved charge conservation dictates the Kosterlitz-Thouless type ``charge-sharpening" transition between the two kinds of dynamics where the global charge information is either hidden or reconstructible from the local measurements. All the relevant phases and transitions can be understood within the same unifying field theory. Additionally, this field theory provides insight into why the charge-sharpening transition should only happen in the presence of volume-law entanglement scaling. Moreover, we show that our field theory, despite being constructed using the Keldysh formalism, is equivalent to a statistical mechanics model that describes the (replicated) time-evolution of the system's density matrix mapped to a doubled Hilbert space under the Choi-Jamiolkowski isomorphism. This equivalence is the result of the trivialization of the fermion distribution function by measurement-induced heating effects., Comment: 30 pages, 4 figures

In this paper, we develop a unified framework for lower bound methods in statistical estimation and interactive decision making. Classical lower bound techniques -- such as Fano's inequality, Le Cam's method, and Assouad's lemma -- have been central to the study of minimax risk in statistical estimation, yet they are insufficient for the analysis of methods that collect data in an interactive manner. The recent minimax lower bounds for interactive decision making via the Decision-Estimation Coefficient (DEC) appear to be genuinely different from the classical methods. We propose a unified view of these distinct methodologies through a general algorithmic lower bound method. We further introduce a novel complexity measure, decision dimension, which facilitates the derivation of new lower bounds for interactive decision making. In particular, decision dimension provides a characterization of bandit learnability for any structured bandit model class. Further, we characterize the sample complexity of learning convex model class up to a polynomial gap with the decision dimension, addressing the remaining gap between upper and lower bounds in Foster et al. (2021, 2023).

Language models (LMs) can "memorize" information, i.e., encode training data in their weights in such a way that inference-time queries can lead to verbatim regurgitation of that data. This ability to extract training data can be problematic, for example, when data are private or sensitive. In this work, we investigate methods to mitigate memorization: three regularizer-based, three finetuning-based, and eleven machine unlearning-based methods, with five of the latter being new methods that we introduce. We also introduce TinyMem, a suite of small, computationally-efficient LMs for the rapid development and evaluation of memorization-mitigation methods. We demonstrate that the mitigation methods that we develop using TinyMem can successfully be applied to production-grade LMs, and we determine via experiment that: regularizer-based mitigation methods are slow and ineffective at curbing memorization; fine-tuning-based methods are effective at curbing memorization, but overly expensive, especially for retaining higher accuracies; and unlearning-based methods are faster and more effective, allowing for the precise localization and removal of memorized information from LM weights prior to inference. We show, in particular, that our proposed unlearning method BalancedSubnet outperforms other mitigation methods at removing memorized information while preserving performance on target tasks.

We calculate higher-order soft-gluon corrections for the associated production of a top-antitop quark pair and a $Z$ boson ($t{\bar t}Z$ production) in proton-proton collisions. We find that these corrections are numerically dominant and large. We present approximate NNLO (aNNLO) and approximate N$^3$LO (aN$^3$LO) cross sections that add, respectively, second-order and third-order soft-gluon corrections to the exact NLO QCD result. We also add electroweak (EW) corrections through NLO. We estimate uncertainties that arise from scale dependence, which are smaller at higher orders, and from parton distribution functions. We compare our aN$^3$LO QCD + NLO EW theoretical results to measurements of $t{\bar t}Z$ cross sections from the LHC and find good agreement. We also calculate single-particle-inclusive differential distributions in top-quark transverse momentum and rapidity., Comment: 10 pages, 3 figures. arXiv admin note: text overlap with arXiv:2312.00861

Scanning Probe Microscopy (SPM) techniques have shown great potential in fabricating nanoscale structures endowed with exotic quantum properties achieved through various manipulations of atoms and molecules. However, the selection of proper manipulation parameters requires extensive domain knowledge, which is not necessarily transferable to new systems. Therefore, efficient and autonomous SPM techniques are needed to reduce the reliance on user supervision and learn optimal strategies for new systems, in particular for the challenge of controlling chemical reactions. In this paper, we developed a software infrastructure named AutoOSS (Autonomous On-Surface Synthesis) to automate bromine removal from Zn(II)-5,15- bis(4-bromo-2,6-dimethylphenyl)porphyrin (ZnBr2Me4DPP) on Au(111), using neural network models to interpret STM outputs and deep reinforcement learning models to optimize manipulation parameters. This is further supported by Bayesian Optimization Structure Search (BOSS) and Density Functional Theory (DFT) computations to explore 3D structures and reaction mechanisms based on STM images.

Protein-ligand binding is the process by which a small molecule (drug or inhibitor) attaches to a target protein. The binding affinity, which refers to the strength of this interaction, is central to many important problems in bioinformatics such as drug design. An extensive amount of work has been devoted to predicting binding affinity over the past decades due to its significance. In this paper, we review all significant recent works, focusing on the methods, features, and benchmark datasets. We have observed a rising trend in the use of traditional machine learning and deep learning models for predicting binding affinity, accompanied by an increasing amount of data on proteins and small drug-like molecules. While prediction results are constantly improving, we also identify several open questions and potential directions that remain unexplored in the field. This paper could serve as an excellent starting point for machine learning researchers who wish to engage in the study of binding affinity, or for anyone with general interests in machine learning, drug discovery, and bioinformatics.

The scaling of superconducting nanowire-based devices to larger arrays is often limited by the cabling required to interface with each device. Cryogenic integrated circuits constructed from nanowire cryotrons, or nanocryotrons, can address this limitation by performing signal processing on chip. In this study, we characterize key performance metrics of the nanocryotron to elucidate its potential as a logical element in cryogenic integrated circuits and develop an electro-thermal model to connect material parameters with device performance. We find that the performance of the nanocryotron depends significantly on the device geometry, and trade-offs are associated with optimizing the gain, jitter, and energy dissipation. We demonstrate that nanocryotrons fabricated on niobium nitride can achieve a grey zone less than 210 nA wide for a 5 ns long input pulse corresponding to a maximum achievable gain of 48 dB, an energy dissipation of less than 20 aJ per operation, and a jitter of less than 60 ps., Comment: 5 pages, 5 figures

This paper addresses the capacitated periodic review inventory control problem, focusing on a retailer managing multiple products with limited shared resources, such as storage or inbound labor at a facility. Specifically, this paper is motivated by the questions of (1) what does it mean to backtest a capacity control mechanism, (2) can we devise and backtest a capacity control mechanism that is compatible with recent advances in deep reinforcement learning for inventory management? First, because we only have a single historic sample path of Amazon's capacity limits, we propose a method that samples from a distribution of possible constraint paths covering a space of real-world scenarios. This novel approach allows for more robust and realistic testing of inventory management strategies. Second, we extend the exo-IDP (Exogenous Decision Process) formulation of Madeka et al. 2022 to capacitated periodic review inventory control problems and show that certain capacitated control problems are no harder than supervised learning. Third, we introduce a `neural coordinator', designed to produce forecasts of capacity prices, guiding the system to adhere to target constraints in place of a traditional model predictive controller. Finally, we apply a modified DirectBackprop algorithm for learning a deep RL buying policy and a training the neural coordinator. Our methodology is evaluated through large-scale backtests, demonstrating RL buying policies with a neural coordinator outperforms classic baselines both in terms of cumulative discounted reward and capacity adherence (we see improvements of up to 50% in some cases).

Federated Learning (FL) is a decentralized machine learning paradigm where models are trained on distributed devices and are aggregated at a central server. Existing FL frameworks assume simple two-tier network topologies where end devices are directly connected to the aggregation server. While this is a practical mental model, it does not exploit the inherent topology of real-world distributed systems like the Internet-of-Things. We present Flight, a novel FL framework that supports complex hierarchical multi-tier topologies, asynchronous aggregation, and decouples the control plane from the data plane. We compare the performance of Flight against Flower, a state-of-the-art FL framework. Our results show that Flight scales beyond Flower, supporting up to 2048 simultaneous devices, and reduces FL makespan across several models. Finally, we show that Flight's hierarchical FL model can reduce communication overheads by more than 60%.

We report measurements of ten custom-made high-homogeneity LYSO crystals. The investigation is motivated by the need for a compact, high-resolution, and fast electromagnetic calorimeter for a new rare pion decay experiment. Each $2.5\times 2.5 \times 18$ cm$^3$ crystal was first characterized for general light yield properties and then its longitudinal response uniformity and energy resolution were measured using low-energy gamma sources. The ten crystals were assembled as an array and subjected to a 30 - 100 MeV positron beam with excellent momentum definition. The energy and timing resolutions were measured as a function of energy, and the spatial resolution was determined at 70 MeV. An additional measurement using monoenergetic 17.6 MeV gammas produced through a p-Li resonance was later made after the photosensors used in positron testing were improved. As an example of the results, the energy resolution at 70 MeV of 1.80 $\pm$ 0.05% is more than two times better than reported results using previous generation LYSO crystals.

Dreams offer a unique window into the cognitive and affective dynamics of the sleeping and the waking mind. Recent quantitative linguistic approaches have shown promise in obtaining corpus-level measures of dream sentiment and topic occurrence. However, it is currently unclear how the affective content of individual dreams relates to their semantic content and structure. Here, we combine word embedding, topic modeling, and network analysis to investigate this relationship. By applying Discourse Atom Topic Modeling (DATM) to the DreamBank corpus of >18K dream reports, we represent the latent themes arising within dreams as a sparse dictionary of topics and identify the affective associations of those topics. We show that variation in dream affect (valence and arousal) is associated with changes in topical content. By representing each dream report as a network of topics, we demonstrate that the affective content of dreams is also coupled to semantic structure. Specifically, positively valenced dreams exhibit more coherent, structured, and linear narratives, whilst negatively valenced dreams have more narrative loops and dominant topics. Additionally, topic networks of high arousal dreams are structurally dominated by few high arousal topics and incoherent topical connections, whereas low arousal dreams contain more loops. These findings suggest that affective processes are associated with both the content and structure of dreams. Our approach showcases the potential of integrating natural language processing and network analysis with psychology to elucidate the interplay of affect, cognition and narrative in dreams. This methodology has broad applications for the study of narrated experience and psychiatric symptomatology.

We prove that a tree with a vertex of degree at least five must be missing a connected partition of some type and therefore its chromatic symmetric function cannot be $e$-positive. We prove that this also holds for a tree with a vertex of degree four as long as it is not adjacent to any leaf. This brings us very close to the conjecture by Dahlberg, She, and van Willigenburg of non-$e$-positivity for all trees with a vertex of degree at least four. We also prove that spiders with four legs cannot have an $e$-positive chromatic symmetric function., Comment: 14 pages

The signature is a collection of iterated integrals describing the "shape" of a path. It appears naturally in the Taylor expansions of controlled differential equations and, as a consequence, is arguably the central object within rough path theory. In this paper, we will consider the signature of Brownian motion with time, and present both new and recently developed approximations for some of its integrals. Since these integrals (or equivalent L\'{e}vy areas) are nonlinear functions of the Brownian path, they are not Gaussian and known to be challenging to simulate. To conclude the paper, we will present some applications of these approximations to the high order numerical simulation of stochastic differential equations (SDEs)., Comment: 69 pages, 12 figures

We conduct a systematic search for astrophysical transients using data from the Atacama Cosmology Telescope (ACT). The data were taken from 2017 to 2022 in three frequency bands spanning 77 GHz to 277 GHz. In this paper we present a pipeline for transient detection using single observation maps where each pixel of a map contains one observation with an integration time of approximately four minutes. We find 34 transient events at 27 unique locations. All but two of the transients are associated with Galactic stars and exhibit a wide range of properties. We also detect an event coincident with the classical nova, YZ Ret and one event consistent with a flaring active galactic nucleus. We notably do not detect any reverse shock emission from gamma ray bursts, a non-detection which is in tension with current models., Comment: 23 pages, 10 figures, 10 tables. First and second author share equal contributions. Article and accompanying data submitted to ApJ. Data tables will be made available upon publication

We demonstrate an exact equivalence between classical population dynamics and Lindbladian evolution admitting a dark state and obeying a set of certain local symmetries. We then introduce {\em quantum population dynamics} as models in which this local symmetry condition is relaxed. This allows for non-classical processes in which animals behave like Schr\"odinger's cat and enter superpositions of live and dead states, thus resulting in coherent superpositions of different population numbers. We develop a field theory treatment of quantum population models as a synthesis of Keldysh and third quantization techniques and draw comparisons to the stochastic Doi-Peliti field theory description of classical population models. We apply this formalism to study a prototypical ``Schr\"odigner cat'' population model on a $d$-dimensional lattice, which exhibits a phase transition between a dark extinct phase and an active phase that supports a stable quantum population. Using a perturbative renormalization group approach, we find a critical scaling of the Schr\"odinger cat population distinct from that observed in both classical population dynamics and usual quantum phase transitions.

The advent of back-illuminated complementary metal-oxide-semiconductor (CMOS) sensors and their well-known advantages over charge-coupled devices (CCDs) make them an attractive technology for future X-ray missions. However, numerous challenges remain, including improving their depletion depth and identifying effective methods to calculate per-pixel gain conversion. We have tested a commercial Sony IMX290LLR CMOS sensor under X-ray light using an $^{55}$Fe radioactive source and collected X-ray photons for $\sim$15 consecutive days under stable conditions at regulated temperatures of 21{\deg}C and 26{\deg}C. At each temperature, the data set contained enough X-ray photons to produce one spectrum per pixel consisting only of single-pixel events. We determined the gain dispersion of its 2.1 million pixels using the peak fitting and the Energy Calibration by Correlation (ECC) methods. We measured a gain dispersion of 0.4\% at both temperatures and demonstrated the advantage of the ECC method in the case of spectra with low statistics. The energy resolution at 5.9 keV after the per-pixel gain correction is improved by $\gtrsim$10 eV for single-pixel and all event spectra, with single-pixel event energy resolution reaching $123.6\pm 0.2$ eV, close to the Fano limit of silicon sensors at room temperature. Finally, our long data acquisition demonstrated the excellent stability of the detector over more than 30 days under a flux of $10^4$ photons per second., Comment: 15 pages, 7 figures, accepted for publication in JATIS

We describe the new design and current performance of the Mid-InfraRed Spectrometer and Imager (MIRSI) on the NASA Infrared Telescope Facility (IRTF). The system has been converted from a liquid nitrogen/liquid helium cryogen system to one that uses a closed-cycle cooler, which allows it to be kept on the telescope at operating temperature and available for observing on short notice, requiring less effort by the telescope operators and day crew to maintain operating temperature. Several other enhancements have been completed, including new detector readout electronics, an IRTF-style standard instrument user interface, new stepper motor driver electronics, and an optical camera that views the same field as the mid-IR instrument using a cold dichroic mirror, allowing for guiding and/or simultaneous optical imaging. The instrument performance is presented, both with an engineering-grade array used from 2021-2023, and a science-grade array installed in the fall of 2023. Some sample astronomical results are also shown. The upgraded MIRSI is a facility instrument at the IRTF available to all users., Comment: 15 pages, 17 figures, accepted to PASP

Variational ab-initio methods in quantum chemistry stand out among other methods in providing direct access to the wave function. This allows in principle straightforward extraction of any other observable of interest, besides the energy, but in practice this extraction is often technically difficult and computationally impractical. Here, we consider the electron density as a central observable in quantum chemistry and introduce a novel method to obtain accurate densities from real-space many-electron wave functions by representing the density with a neural network that captures known asymptotic properties and is trained from the wave function by score matching and noise-contrastive estimation. We use variational quantum Monte Carlo with deep-learning ans\"atze (deep QMC) to obtain highly accurate wave functions free of basis set errors, and from them, using our novel method, correspondingly accurate electron densities, which we demonstrate by calculating dipole moments, nuclear forces, contact densities, and other density-based properties., Comment: 12 pages, 9 figures in the main text