Search

Many real-world applications of tabular data involve using historic events to predict properties of new ones, for example whether a credit card transaction is fraudulent or what rating a customer will assign a product on a retail platform. Existing approaches to event prediction include costly, brittle, and application-dependent techniques such as time-aware positional embeddings, learned row and field encodings, and oversampling methods for addressing class imbalance. Moreover, these approaches often assume specific use-cases, for example that we know the labels of all historic events or that we only predict a pre-specified label and not the data's features themselves. In this work, we propose a simple but flexible baseline using standard autoregressive LLM-style transformers with elementary positional embeddings and a causal language modeling objective. Our baseline outperforms existing approaches across popular datasets and can be employed for various use-cases. We demonstrate that the same model can predict labels, impute missing values, or model event sequences., Comment: 10 pages, 6 pages of references+appendix

Lubricated contacts in soft materials are important in various engineering systems and natural settings. Three major lubrication regimes are boundary (BL), mixed (ML), and elasto-hydrodynamic (EHL) lubrication, where the contact region is dry, partially wetted, or fully wetted, respectively. The transition between these regimes is insufficiently understood, especially for soft contacts, which impedes desired control of lubricated sliding friction. Here, we report on the role of solid wettability and adhesion on these transitions. Wettability of glycerol on polydimethylsiloxane (PDMS) surface, and adhesion between a glass indenter and PDMS, were varied by exposure of the PDMS to an ultraviolet light-ozone (UV-Ozone) cleaner. By combining friction tests and visualization, we demonstrate that the transition from ML to BL regime is dominated by the wettability of the lubricant; increasing wettability of glycerol makes removal of liquid from the contact region more difficult. Transition from EHL to ML is related to a series of events with increasing normal load, which are thinning of the lubricant layer, sudden jump to contact between the glass indenter and solid substrate across a gap of tens to a few hundreds of nanometers, and attendant elastic instabilities such as wrinkling and stick-slip. These results provide a deeper understanding of transitions in lubricated frictional behavior of soft materials which govern the maximum and minimum friction achievable.

The miniaturisation of transistors is approaching its limits due to challenges in heat management and information transfer speed. To overcome these obstacles, emerging technologies such as spintronics are being developed, which leverage the electron's spin in addition to its charge. Local phenomena at interfaces or structural defects will greatly influence the efficiency of spin-based devices, making the ability to study and control spin wave propagation at the nano- and atomic scales a key challenge. The development of high-spatial-resolution tools to probe spin waves, also called magnons, at relevant lengthscales is thus essential to understand how their properties are affected by such local features. Here, we show the first experimental detection of bulk magnons at the nanoscale using scanning transmission electron microscopy. By employing high-resolution electron energy loss spectroscopy with hybrid-pixel direct electron detectors optimized for low acceleration voltages, we successfully overcome the challenges posed by weak signals and identify magnon excitations in a thin NiO nanocrystal. Advanced inelastic electron scattering simulations corroborate our findings. These results open new avenues for detecting magnons, exploring their dispersions and their modifications arising from nanoscale structural or chemical defects. This marks an important milestone in magnonics and presents exciting opportunities for the future development of spintronic devices.

U.S. presidential elections are decided by the Electoral College, established in 1789, and designed to mitigate potential risks arising from the collusion of large groups of citizens. A statewide winner-take-all popular voting system for electors is implemented in all but two states, which has led to instances where narrow victories in key states were decisive in several recent elections. Small groups of voters can significantly impact the election, for example, through voter turnout. However, another dynamic can also influence this: a surprisingly small number of dedicated voters moving short distances across state lines. The extent to which the election's outcome is sensitive to small and well-coordinated movements of people has not been investigated in detail. Using a combination of forecasting, simulation, and optimization, we show that a candidate's probability of winning can be increased by 1% through the strategic relocation of approximately 10,000 people no farther than 100 miles from their current county of residence, less than 0.006% of the eligible voting population. Moreover, an 8% probability increase can be realized by a mere 50,000 voters relocating across state lines, or 0.03% of the voting population. Given the remarkably small number of people involved and the fact that establishing electoral residence in many states takes about a month, this coordinated relocation of voters is not nearly as challenging as previously thought. As it stands, U.S. presidential elections may be vulnerable to the exploitation of the aforementioned loophole. Therefore, we anticipate our findings will have direct consequences on policymaking and campaign strategy, as well as motivate new operations research methods within the political sciences., Comment: 16 pages, 2 figures, under review

Beyond classical Bragg diffraction, we report on new sub-Bragg phenomena to achieve enhanced sound wave tunability in the low-frequency range in a multilayered acoustic metamaterial system. Remarkably, we reveal and study the formation of genuinely sub-wavelength Bragg-like band-splitting induced bandgaps, generalizing the band-folding induced bandgaps in the literature. Additionally, we propose a methodology to widen sub-wavelength local resonance bandgaps by simultaneously hosting two local resonances within the same bandgap. These sub-Bragg bandgaps are realized in an axisymmetric vibroacoustic phononic metamaterial consisting of repetitive multilayered unit cells, each composed of two layers of membrane-cavity resonators. The resulting coupled sound-structure interaction system is analytically solved in closed form. Furthermore, the studied multilayered vibroacoustic metamaterial exhibits sub-wavelength acoustical transparency, akin to electromagnetically induced transparency, and an acoustic analogue of low-frequency plasma oscillations, resulting in a plasma bandgap where wave propagation is entirely prohibited below the plasma frequency. These sub-wavelength phenomena, achieved predictively and well below the ubiquitous first-order Bragg diffraction range, provide broadband attenuation and superior wave manipulation capabilities for low-frequency sound. This highlights the potential of utilizing sound-structure interaction across diverse physical applications. Moreover, our findings can be extrapolated to wave propagation in broader classes of physical systems where sub-wavelength phenomena are crucial, including phononic and photonic crystals of more general configurations.

Magnetoelasticity plays a crucial role in numerous magnetic phenomena, including magnetocalorics, magnon excitation via acoustic waves, and ultrafast demagnetization/Einstein-de Haas effect. Despite a long-standing discussion on anisotropy-mediated magnetoelastic interactions of relativistic origin, including \textit{ab-initio} calculations, the exchange-mediated magnetoelastic parameters within an atomistic framework have only recently begun to be investigated. As a result, many of their behaviors and values for real materials remain poorly understood. Therefore, by using a proposed simple modification of the embedded cluster approach that reduces the computational complexity, we critically analyze the properties of exchange-mediated spin-lattice coupling parameters for elemental $3d$ ferromagnets (bcc Fe, fcc Ni, and fcc Co), comparing methods used for their extraction and relating their realistic values to symmetry considerations and orbitally-decomposed contributions. Additionally, we investigate the effects of noncollinearity (spin temperature) and applied pressure on these parameters. For Fe, we find that single-site rotations, associated with spin temperatures around $\sim100$ K, induce significant modifications, particularly in Dzyaloshinskii-Moriya-type couplings; in contrast, such interactions in Co and Ni remain almost configuration independent. Moreover, we demonstrate a notable change in the exchange-mediated magnetoelastic constants for Fe under isostatic compression. Finally, the conversion between atomistic, quantum-mechanically derived parameters and the phenomenological magnetoelastic theory is discussed, which can be an useful tool towards larger and more realistic dynamics simulations involving coupled subsystems.

If $\mathcal{C}$ is a category of algebras closed under finite direct products, and $M_\mathcal{C}$ the commutative monoid of isomorphism classes of members of $\mathcal{C},$ with operation induced by direct product, A.Tarski defines a nonidentity element $p$ of $M_\mathcal{C}$ to be prime if, whenever it divides a product of two elements in that monoid, it divides one of them, and calls an object of $\mathcal{C}$ prime if its isomorphism class has this property. McKenzie, McNulty and Taylor ask whether the category of nonempty semigroups has any prime objects. We show in section 2 that it does not. However, for the category of monoids, and some other subcategories of semigroups, we obtain examples of prime objects in sections 3-5 In section 6, two related questions open so far as I know, are recalled. In section 7, which can be read independently of the rest of this note, we recall two related conditions that are called primeness by semigroup theorists, and obtain results and examples on the relationships among those two conditions and Tarski's, in categories of groups. Section 8 notes an interesting characterization of one of those conditions when applied to finite algebras in an arbitrary variety. Various questions are raised., Comment: 16 pages. Copy at http://math.berkeley.edu/~gbergman/papers may be updated more frequently than arXiv copy

Let $f$ be a holomorphic self-map of the unit disc. We show that if $\log (1-\lvert f(z) \rvert)$ is integrable on a sub-arc of the unit circle, $I$, then the set of Carath\'eodory angular derivatives of $f$ on $I$ is a countable union of Beurling-Carleson sets of finite entropy. Conversely, given a countable union of Beurling-Carleson sets, $E$, we construct a holomorphic self-map of the unit disc, such that its set of Carath\'eodory angular derivatives is equal to $E$ and $\log(1-\lvert f(z) \rvert)$ is integrable on the unit circle. Our main technical tools are the Aleksandrov disintegration Theorem and a characterization of countable unions of Beurling-Carleson sets due to Makarov and Nikolski.

Memory has become the primary cost driver in cloud data centers. Yet, a significant portion of memory allocated to VMs in public clouds remains unused. To optimize this resource, "cold" memory can be reclaimed from VMs and stored on slower storage or compressed, enabling memory overcommit. Current overcommit systems rely on general-purpose OS swap mechanisms, which are not optimized for virtualized workloads, leading to missed memory-saving opportunities and ineffective use of optimizations like prefetchers. This paper introduces a userspace memory management framework designed for VMs. It enables custom policies that have full control over the virtual machines' memory using a simple userspace API, supports huge page-based swapping to satisfy VM performance requirements, is easy to deploy by leveraging Linux/KVM, and supports zero-copy I/O virtualization with shared VM memory. Our evaluation demonstrates that an overcommit system based on our framework outperforms the state-of-the-art solutions on both micro-benchmarks and commonly used cloud workloads. Specifically our implementation outperforms the Linux Kernel baseline implementation by up to 25% while saving a similar amount of memory. We also demonstrate the benefits of custom policies by implementing workload-specific reclaimers and prefetchers that save $10\%$ additional memory, improve performance in a limited memory scenario by 30% over the Linux baseline, and recover faster from hard limit releases., Comment: 13 pages, 13 figures

We present Jamba-1.5, new instruction-tuned large language models based on our Jamba architecture. Jamba is a hybrid Transformer-Mamba mixture of experts architecture, providing high throughput and low memory usage across context lengths, while retaining the same or better quality as Transformer models. We release two model sizes: Jamba-1.5-Large, with 94B active parameters, and Jamba-1.5-Mini, with 12B active parameters. Both models are fine-tuned for a variety of conversational and instruction-following capabilties, and have an effective context length of 256K tokens, the largest amongst open-weight models. To support cost-effective inference, we introduce ExpertsInt8, a novel quantization technique that allows fitting Jamba-1.5-Large on a machine with 8 80GB GPUs when processing 256K-token contexts without loss of quality. When evaluated on a battery of academic and chatbot benchmarks, Jamba-1.5 models achieve excellent results while providing high throughput and outperforming other open-weight models on long-context benchmarks. The model weights for both sizes are publicly available under the Jamba Open Model License and we release ExpertsInt8 as open source., Comment: Webpage: https://www.ai21.com/jamba

SPIDER is a balloon-borne instrument designed to map the cosmic microwave background at degree-angular scales in the presence of Galactic foregrounds. SPIDER has mapped a large sky area in the Southern Hemisphere using more than 2000 transition-edge sensors (TESs) during two NASA Long Duration Balloon flights above the Antarctic continent. During its first flight in January 2015, SPIDER observed in the 95 GHz and 150 GHz frequency bands, setting constraints on the B-mode signature of primordial gravitational waves. Its second flight in the 2022-23 season added new receivers at 280 GHz, each using an array of TESs coupled to the sky through feedhorns formed from stacks of silicon wafers. These receivers are optimized to produce deep maps of polarized Galactic dust emission over a large sky area, providing a unique data set with lasting value to the field. In this work, we describe the instrument's performance during SPIDER's second flight., Comment: Submitted to SPIE Astronomical Telescopes + Instrumentation 2024, JATIS

This study investigates a vibroacoustic phononic metamaterial system composed of repeated monolayered membrane-air cavity unit-cells to assess its efficacy in controlling sound waves. Assuming low-frequency axisymmetric modes, the coupled membrane-cavity vibroacoustic system for a representative unit-cell is solved entirely analytically. Unlike previous research that relied on an infinite series of eigenfunctions, our analysis offers a single-term exact solution for the membrane's displacement field, fully accounting for coupling with the acoustic cavities. Utilizing the transfer matrix method and the Bloch-Floquet theorem, we offer a comprehensive analytical characterization of the band structure, including closed-form analytical expressions for determining the bounding frequencies of the bandgaps and the dispersion branches. Interaction between Bragg and local resonance bandgaps is examined by adjusting Bragg bandgap positions, with detailed mathematical descriptions provided for their overlapping and transition. Additionally, a "plasma bandgap" analogous to metallic plasma oscillations is identified, with a derived analytical expression for its frequency. First two passbands remain robust against cavity depth variations, limiting wave manipulation capabilities. Analysis of the finite phononic system involves constructing the global transfer matrix to study natural frequencies and scattering coefficients. Interaction between Bragg and local resonance bandgaps in finite systems results in ultra-narrow passbands, creating transparency windows analogous to electromagnetically induced transparency by quantum interference. This theoretical framework enables precise characterization and engineering of bandgaps in the monolayered vibroacoustic phononic metamaterial, highlighting its potential for controlling low-frequency sound wave propagation across multiple frequencies.

Using data from the first flight of SPIDER and from Planck HFI, we probe the properties of polarized emission from interstellar dust in the SPIDER observing region. Component separation algorithms operating in both the spatial and harmonic domains are applied to probe their consistency and to quantify modeling errors associated with their assumptions. Analyses spanning the full SPIDER region demonstrate that i) the spectral energy distribution of diffuse Galactic dust emission is broadly consistent with a modified-blackbody (MBB) model with a spectral index of $\beta_\mathrm{d}=1.45\pm0.05$ $(1.47\pm0.06)$ for $E$ ($B$)-mode polarization, slightly lower than that reported by Planck for the full sky; ii) its angular power spectrum is broadly consistent with a power law; and iii) there is no significant detection of line-of-sight decorrelation of the astrophysical polarization. The size of the SPIDER region further allows for a statistically meaningful analysis of the variation in foreground properties within it. Assuming a fixed dust temperature $T_\mathrm{d}=19.6$ K, an analysis of two independent sub-regions of that field results in inferred values of $\beta_\mathrm{d}=1.52\pm0.06$ and $\beta_\mathrm{d}=1.09\pm0.09$, which are inconsistent at the $3.9\,\sigma$ level. Furthermore, a joint analysis of SPIDER and Planck 217 and 353 GHz data within a subset of the SPIDER region is inconsistent with a simple MBB at more than $3\,\sigma$, assuming a common morphology of polarized dust emission over the full range of frequencies. These modeling uncertainties have a small--but non-negligible--impact on limits on the cosmological tensor-to-scalar ratio derived from the \spider dataset. The fidelity of the component separation approaches of future CMB polarization experiments may thus have a significant impact on their constraining power., Comment: 21 pages, 15 figures

We investigate relaxations for a class of discrete bilevel programs where the interaction constraints linking the leader and the follower are linear. Our approach reformulates the upper-level optimality constraints by projecting the leader's decisions onto vectors that map to distinct follower solution values, each referred to as a state. Based on such a state representation, we develop a network-flow linear program via a decision diagram that captures the convex hull of the follower's value function graph, leading to a new single-level reformulation of the bilevel problem. We also present a reduction procedure that exploits symmetry to identify the reformulation of minimal size. For large networks, we introduce parameterized relaxations that aggregate states by considering tractable hyperrectangles based on lower and upper bounds associated with the interaction constraints, and can be integrated into existing mixed-integer bilevel linear programming (MIBLP) solvers. Numerical experiments suggest that the new relaxations, whether used within a simple cutting-plane procedure or integrated into state-of-the-art MIBLP solvers, significantly reduce runtimes or solve additional benchmark instances. Our findings also highlight the correlation between the quality of relaxations and the properties of the interaction matrix, underscoring the potential of our approach in enhancing solution methods for structured bilevel optimization instances.

This paper explores multi-entry strategies for betting pools related to single-elimination tournaments. In such betting pools, participants select winners of games, and their respective score is a weighted sum of the number of correct selections. Most betting pools have a top-heavy payoff structure, so the paper focuses on strategies that maximize the expected score of the best-performing entry. There is no known closed-formula expression for the estimation of this metric, so the paper investigates the challenges associated with the estimation and the optimization of multi-entry solutions. We present an exact dynamic programming approach for calculating the maximum expected score of any given fixed solution, which is exponential in the number of entries. We explore the structural properties of the problem to develop several solution techniques. In particular, by extracting insights from the solutions produced by one of our algorithms, we design a simple yet effective problem-specific heuristic that was the best-performing technique in our experiments, which were based on real-world data extracted from recent March Madness tournaments. In particular, our results show that the best 100-entry solution identified by our heuristic had a 2.2% likelihood of winning a $1 million prize in a real-world betting pool.

High-dimensional action spaces remain a challenge for dynamic algorithm configuration (DAC). Interdependencies and varying importance between action dimensions are further known key characteristics of DAC problems. We argue that these Coupled Action Dimensions with Importance Differences (CANDID) represent aspects of the DAC problem that are not yet fully explored. To address this gap, we introduce a new white-box benchmark within the DACBench suite that simulates the properties of CANDID. Further, we propose sequential policies as an effective strategy for managing these properties. Such policies factorize the action space and mitigate exponential growth by learning a policy per action dimension. At the same time, these policies accommodate the interdependence of action dimensions by fostering implicit coordination. We show this in an experimental study of value-based policies on our new benchmark. This study demonstrates that sequential policies significantly outperform independent learning of factorized policies in CANDID action spaces. In addition, they overcome the scalability limitations associated with learning a single policy across all action dimensions. The code used for our experiments is available under https://github.com/PhilippBordne/candidDAC., Comment: 5 pages main paper, 11 pages references and appendix, 9 figures, to be published in: Proceedings of the Third International Conference on Automated Machine Learning (AutoML 2024), Workshop Track

We propose a holographic description of the operators implementing $U(1)$ global symmetries that are dual to superstring gauge fields in terms of non-BPS D-branes. We check the consistency of our proposal in a number of examples., Comment: 14 pages, 3 figures. References added

We report an estimation of the injected mass composition of ultra-high energy cosmic rays (UHECRs) at energies higher than 10 EeV. The composition is inferred from an energy-dependent sky distribution of UHECR events observed by the Telescope Array surface detector by comparing it to the Large Scale Structure of the local Universe. In the case of negligible extra-galactic magnetic fields the results are consistent with a relatively heavy injected composition at E ~ 10 EeV that becomes lighter up to E ~ 100 EeV, while the composition at E > 100 EeV is very heavy. The latter is true even in the presence of highest experimentally allowed extra-galactic magnetic fields, while the composition at lower energies can be light if a strong EGMF is present. The effect of the uncertainty in the galactic magnetic field on these results is subdominant., Comment: 8 pages, 3 figures, accepted for publication in PRL

We use a new method to estimate the injected mass composition of ultrahigh cosmic rays (UHECRs) at energies higher than 10 EeV. The method is based on comparison of the energy-dependent distribution of cosmic ray arrival directions as measured by the Telescope Array experiment (TA) with that calculated in a given putative model of UHECR under the assumption that sources trace the large-scale structure (LSS) of the Universe. As we report in the companion letter, the TA data show large deflections with respect to the LSS which can be explained, assuming small extra-galactic magnetic fields (EGMF), by an intermediate composition changing to a heavy one (iron) in the highest energy bin. Here we show that these results are robust to uncertainties in UHECR injection spectra, the energy scale of the experiment and galactic magnetic fields (GMF). The assumption of weak EGMF, however, strongly affects this interpretation at all but the highest energies E > 100 EeV, where the remarkable isotropy of the data implies a heavy injected composition even in the case of strong EGMF. This result also holds if UHECR sources are as rare as $2 \times 10^{-5}$ Mpc$^{-3}$, that is the conservative lower limit for the source number density., Comment: 18 pages, 11 figures, accepted for publication in PRD

Toroidal classification and geometric duality in quantum spin systems is presented. Through our classification and duality, we reveal that various bipartite quantum features in magnon-systems can manifest equivalently in both bipartite ferromagnetic and antiferromagnetic materials, based upon the availability of relevant Hamiltonian parameters. Additionally, the results highlight the antiferromagnetic regime as an ultra-fast dual counterpart to the ferromagnetic regime, both exhibiting identical capabilities for quantum spintronics and technological applications. Concrete illustrations are provided, demonstrating how splitting and squeezing types of two-mode magnon quantum correlations can be realized across ferro- and antiferromagnetic regimes., Comment: 6 pages, 2 figures

Algorithmic fairness has emerged as a critical concern in artificial intelligence (AI) research. However, the development of fair AI systems is not an objective process. Fairness is an inherently subjective concept, shaped by the values, experiences, and identities of those involved in research and development. To better understand the norms and values embedded in current fairness research, we conduct a meta-analysis of algorithmic fairness papers from two leading conferences on AI fairness and ethics, AIES and FAccT, covering a final sample of 139 papers over the period from 2018 to 2022. Our investigation reveals two concerning trends: first, a US-centric perspective dominates throughout fairness research; and second, fairness studies exhibit a widespread reliance on binary codifications of human identity (e.g., "Black/White", "male/female"). These findings highlight how current research often overlooks the complexities of identity and lived experiences, ultimately failing to represent diverse global contexts when defining algorithmic bias and fairness. We discuss the limitations of these research design choices and offer recommendations for fostering more inclusive and representative approaches to fairness in AI systems, urging a paradigm shift that embraces nuanced, global understandings of human identity and values.

This research introduces STAR, a sociotechnical framework that improves on current best practices for red teaming safety of large language models. STAR makes two key contributions: it enhances steerability by generating parameterised instructions for human red teamers, leading to improved coverage of the risk surface. Parameterised instructions also provide more detailed insights into model failures at no increased cost. Second, STAR improves signal quality by matching demographics to assess harms for specific groups, resulting in more sensitive annotations. STAR further employs a novel step of arbitration to leverage diverse viewpoints and improve label reliability, treating disagreement not as noise but as a valuable contribution to signal quality., Comment: 8 pages, 5 figures, 5 pages appendix. * denotes equal contribution

We report on an observation of the difference between northern and southern skies of the ultrahigh energy cosmic ray energy spectrum with a significance of ${\sim}8\sigma$. We use measurements from the two largest experiments$\unicode{x2014}$the Telescope Array observing the northern hemisphere and the Pierre Auger Observatory viewing the southern hemisphere. Since the comparison of two measurements from different observatories introduces the issue of possible systematic differences between detectors and analyses, we validate the methodology of the comparison by examining the region of the sky where the apertures of the two observatories overlap. Although the spectra differ in this region, we find that there is only a $1.8\sigma$ difference between the spectrum measurements when anisotropic regions are removed and a fiducial cut in the aperture is applied., Comment: 8 pages, 6 figures

The consistency model (CM) has recently made significant progress in accelerating the generation of diffusion models. However, its application to high-resolution, text-conditioned image generation in the latent space (a.k.a., LCM) remains unsatisfactory. In this paper, we identify three key flaws in the current design of LCM. We investigate the reasons behind these limitations and propose the Phased Consistency Model (PCM), which generalizes the design space and addresses all identified limitations. Our evaluations demonstrate that PCM significantly outperforms LCM across 1--16 step generation settings. While PCM is specifically designed for multi-step refinement, it achieves even superior or comparable 1-step generation results to previously state-of-the-art specifically designed 1-step methods. Furthermore, we show that PCM's methodology is versatile and applicable to video generation, enabling us to train the state-of-the-art few-step text-to-video generator. More details are available at https://g-u-n.github.io/projects/pcm/.

State-of-the-art automated machine learning systems for tabular data often employ cross-validation; ensuring that measured performances generalize to unseen data, or that subsequent ensembling does not overfit. However, using k-fold cross-validation instead of holdout validation drastically increases the computational cost of validating a single configuration. While ensuring better generalization and, by extension, better performance, the additional cost is often prohibitive for effective model selection within a time budget. We aim to make model selection with cross-validation more effective. Therefore, we study early stopping the process of cross-validation during model selection. We investigate the impact of early stopping on random search for two algorithms, MLP and random forest, across 36 classification datasets. We further analyze the impact of the number of folds by considering 3-, 5-, and 10-folds. In addition, we investigate the impact of early stopping with Bayesian optimization instead of random search and also repeated cross-validation. Our exploratory study shows that even a simple-to-understand and easy-to-implement method consistently allows model selection to converge faster; in ~94% of all datasets, on average by ~214%. Moreover, stopping cross-validation enables model selection to explore the search space more exhaustively by considering +167% configurations on average within one hour, while also obtaining better overall performance., Comment: Accepted at Third International Conference on Automated Machine Learning (AutoML 2024); for code, see https://github.com/automl/DontWasteYourTime-early-stopping

The impact of the local chemical environment on the Gilbert damping in the binary alloy Fe$_{100-x}$Co$_{x}$ is investigated, using computations based on density functional theory. By varying the alloy composition x as well as Fe/Co atom positions we reveal that the effective damping of the alloy is highly sensitive to the nearest neighbor environment, especially to the amount of Co and the average distance between Co-Co atoms at nearest neighbor sites. Both lead to a significant local increase (up to an order of magnitude) of the effective Gilbert damping, originating mainly from variations of the density of states at the Fermi energy. In a global perspective (i.e., making a configuration average for a real material), those differences in damping are masked by statistical averages. When low-temperature explicit atomistic dynamics simulations are performed, the impact of short-range disorder on local dynamics is observed to also alter the overall relaxation rate. Our results illustrate the possibility of local chemical engineering of the Gilbert damping, which may stimulate the study of new ways to tune and control materials aiming for spintronics applications.

With the increasing computational costs associated with deep learning, automated hyperparameter optimization methods, strongly relying on black-box Bayesian optimization (BO), face limitations. Freeze-thaw BO offers a promising grey-box alternative, strategically allocating scarce resources incrementally to different configurations. However, the frequent surrogate model updates inherent to this approach pose challenges for existing methods, requiring retraining or fine-tuning their neural network surrogates online, introducing overhead, instability, and hyper-hyperparameters. In this work, we propose FT-PFN, a novel surrogate for Freeze-thaw style BO. FT-PFN is a prior-data fitted network (PFN) that leverages the transformers' in-context learning ability to efficiently and reliably do Bayesian learning curve extrapolation in a single forward pass. Our empirical analysis across three benchmark suites shows that the predictions made by FT-PFN are more accurate and 10-100 times faster than those of the deep Gaussian process and deep ensemble surrogates used in previous work. Furthermore, we show that, when combined with our novel acquisition mechanism (MFPI-random), the resulting in-context freeze-thaw BO method (ifBO), yields new state-of-the-art performance in the same three families of deep learning HPO benchmarks considered in prior work., Comment: Published at the 41st International Conference on Machine Learning (ICML), Vienna, Austria

This paper focuses on the opportunities and the ethical and societal risks posed by advanced AI assistants. We define advanced AI assistants as artificial agents with natural language interfaces, whose function is to plan and execute sequences of actions on behalf of a user, across one or more domains, in line with the user's expectations. The paper starts by considering the technology itself, providing an overview of AI assistants, their technical foundations and potential range of applications. It then explores questions around AI value alignment, well-being, safety and malicious uses. Extending the circle of inquiry further, we next consider the relationship between advanced AI assistants and individual users in more detail, exploring topics such as manipulation and persuasion, anthropomorphism, appropriate relationships, trust and privacy. With this analysis in place, we consider the deployment of advanced assistants at a societal scale, focusing on cooperation, equity and access, misinformation, economic impact, the environment and how best to evaluate advanced AI assistants. Finally, we conclude by providing a range of recommendations for researchers, developers, policymakers and public stakeholders.

YMn$_6$Sn$_6$ consists of two types of Mn-based kagome planes stacked along $c$-axis having a complex magnetic interactions. We report a spin reconstruction in YMn$_6$Sn$_6$ from ferromagnet (FM) into a combination of two incommensurate spin spirals (SSs) originating from two different type of Mn kagome planes driven by frustrated magnetic exchanges along the $c$-axis with inclusion of Hubbard U. The pitch angle and wave vector of the incommensurate SSs are $\sim$89.3$^{\circ}$ and $\sim$ (0 0 0.248) respectively which are in excellent agreement with experiment. We employed an effective model Hamiltonian constructed out of exchange interactions to capture experimentally observed non-equivalent nature of the two incommensurate SSs which also explain FM-SS crossover due to antiferromagntic spin exchange with correlation. We further report the existence of topological magnon with spin-orbit coupling in incommensurate SS phase of YMn$_6$Sn$_6$ by calculating the topological invariants and Berry curvature profile. The location of Dirac magnon in energy landscape at 73 meV matches with another experimental report. We demonstrate the accuracy of our results by highlighting experimental features in YMn$_6$Sn$_6$.

To develop new devices based on synthetic ferrimagnetic (S-FiM) heterostructures, understanding the material's physical properties is pivotal. Here, the induced magnetic moment (IMM), magnetic exchange-coupling, and spin textures were investigated at room-temperature in Pt/Co/Gd multilayers using a multiscale approach. The magnitude and direction of the IMM were interpreted experimentally and theoretically in the framework of both X-ray magnetic circular dichroism (XMCD) and density functional theory (DFT). The results demonstrate that the IMM transferred by Co across the Gd paramagnetic (PM) thickness leads to a flipped spin state (FSS) within the Gd layers, in which their magnetic moments couple antiparallel/parallel with the ferromagnetic (FM) Co near/far from the Co/Gd interface, respectively. For the Pt, in both Pt/Co and Gd/Pt interfaces the IMM follows the same direction as the Co magnetic moment, with negligible IMM in the Gd/Pt interface. Additionally, zero-field spin spirals were imaged using scanning transmission X-ray microscopy (STXM), while micromagnetic simulations employed to unfold the interactions stabilizing the FiM configurations, where the existence of a sizable Dzyaloshinskii-Moriya interaction is demonstrated to be crucial for the formation of those spin textures. Our outcomes may add fundamental physical and technological aspects for using FiM films in antiferromagnetic spintronic devices.

We present Jamba, a new base large language model based on a novel hybrid Transformer-Mamba mixture-of-experts (MoE) architecture. Specifically, Jamba interleaves blocks of Transformer and Mamba layers, enjoying the benefits of both model families. MoE is added in some of these layers to increase model capacity while keeping active parameter usage manageable. This flexible architecture allows resource- and objective-specific configurations. In the particular configuration we have implemented, we end up with a powerful model that fits in a single 80GB GPU. Built at large scale, Jamba provides high throughput and small memory footprint compared to vanilla Transformers, and at the same time state-of-the-art performance on standard language model benchmarks and long-context evaluations. Remarkably, the model presents strong results for up to 256K tokens context length. We study various architectural decisions, such as how to combine Transformer and Mamba layers, and how to mix experts, and show that some of them are crucial in large scale modeling. We also describe several interesting properties of these architectures which the training and evaluation of Jamba have revealed, and plan to release checkpoints from various ablation runs, to encourage further exploration of this novel architecture. We make the weights of our implementation of Jamba publicly available under a permissive license., Comment: Webpage: https://www.ai21.com/jamba

Femtosecond laser pulses can be used to induce ultrafast changes of the magnetization in magnetic materials. Several microscopic mechanisms have been proposed to explain the observations, including the transport of ultrashort spin-polarized hot-electrons (SPHE). Such ultrafast spin currents find growing interest because of the recent challenges in ultrafast spintronics however they are only poorly characterized. One of the key challenges is to characterize the spin-polarized ultrafast currents and the microscopic mechanisms behind SPHE induced manipulation of the magnetization, especially in the case of technologically relevant ferrimagnetic alloys. Here, we have used a combined approach using time- and element-resolved X-ray magnetic circular dichroism and theoretical calculations based on atomistic spin-dynamics simulations to address the ultrafast transfer of the angular momentum from spin-polarized currents into ferrimagnetic Fe74Gd26 films and the concomitant reduction of sub-lattice magnetization. Our study shows that using a Co/Pt multilayer as a polarizer in a spin-valve structure, the SPHE drives the demagnetization of the two sub-lattices of the Fe74Gd26 film. This behaviour is explained based on two physical mechanisms, i.e., spin transfer torque and thermal fluctuations induced by the SPHE. We provide a quantitative description of the heat transfer of the ultrashort SPHE pulse to the Fe74Gd26 films, as well as the degree of spin-polarization of the SPHE current density responsible for the observed magnetization dynamics. Our work finally characterizes the spin-polarization of the SPHEs revealing unexpected opposite spin polarization to the Co magnetization, explaining our experimental results., Comment: 22 pages, 19 figures

The Landau-Lifshitz-Gilbert (LLG) and Landau-Lifshitz (LL) equations play an essential role for describing the dynamics of magnetization in solids. While a quantum analog of the LL dynamics has been proposed in [Phys. Rev. Lett. 110, 147201 (2013)], the corresponding quantum version of LLG remains unknown. Here, we propose such a quantum LLG equation that inherently conserves purity of the quantum state. We examine the quantum LLG dynamics of a dimer consisting of two interacting spin-1/2 particles. Our analysis reveals that, in the case of ferromagnetic coupling, the evolution of initially uncorrelated spins mirrors the classical LLG dynamics. However, in the antiferromagnetic scenario, we observe pronounced deviations from classical behavior, underscoring the unique dynamics of becoming a spinless state, which is non-locally correlated. Moreover, when considering spins that are initially correlated, our study uncovers an unusual form of transient quantum correlation dynamics, which differ significantly from what is typically seen in open quantum systems.

Polymer microgels exhibit intriguing macroscopic flow properties arising from their unique microscopic structure. Microgel colloids comprise a crosslinked polymer network with a radially decaying density profile, resulting in a dense core surrounded by a fuzzy corona. Notably, microgels synthesized from poly(N-isopropylacrylamide) (PNIPAM) are thermoresponsive, capable of adjusting their size and density profile based on temperature. Above the lower critical solution temperature ($T_\text{LCST}\sim 33$ $^\circ$C), the microgel's polymer network collapses, leading to the expulsion of water through a reversible process. Conversely, below $33$ $^\circ$C, the microgel's network swells, becoming highly compressible and allowing overpacking to effective volume fractions exceeding one. Under conditions of dense packing, microgels undergo deformation in distinct stages: corona compression and faceting, interpenetration, and finally, isotropic compression. Each stage exhibits a characteristic signature in the yield stress and elastic modulus of the dense microgel suspensions. Here, we introduce a model for the linear elastic shear modulus through the minimization of a quasi-equilibrium free energy, encompassing all relevant energetic contributions. We validate our model by comparing its predictions to experimental results from oscillatory shear rheology tests on microgel suspensions at different densities and temperatures. Our findings demonstrate that combining macroscopic rheological measurements with the model allows for temperature-dependent characterization of polymer interaction parameters., Comment: 23 pages, 4 figures; We have corrected the polymer concentrations; they should be given in wt % instead of g/ml

In multicriteria decision-making, a user seeks a set of non-dominated solutions to a (constrained) multiobjective optimization problem, the so-called Pareto frontier. In this work, we seek to bring a state-of-the-art method for exact multiobjective integer linear programming into the heuristic realm. We focus on binary decision diagrams (BDDs) which first construct a graph that represents all feasible solutions to the problem and then traverse the graph to extract the Pareto frontier. Because the Pareto frontier may be exponentially large, enumerating it over the BDD can be time-consuming. We explore how restricted BDDs, which have already been shown to be effective as heuristics for single-objective problems, can be adapted to multiobjective optimization through the use of machine learning (ML). MORBDD, our ML-based BDD sparsifier, first trains a binary classifier to eliminate BDD nodes that are unlikely to contribute to Pareto solutions, then post-processes the sparse BDD to ensure its connectivity via optimization. Experimental results on multiobjective knapsack problems show that MORBDD is highly effective at producing very small restricted BDDs with excellent approximation quality, outperforming width-limited restricted BDDs and the well-known evolutionary algorithm NSGA-II.

While deep learning has celebrated many successes, its results often hinge on the meticulous selection of hyperparameters (HPs). However, the time-consuming nature of deep learning training makes HP optimization (HPO) a costly endeavor, slowing down the development of efficient HPO tools. While zero-cost benchmarks, which provide performance and runtime without actual training, offer a solution for non-parallel setups, they fall short in parallel setups as each worker must communicate its queried runtime to return its evaluation in the exact order. This work addresses this challenge by introducing a user-friendly Python package that facilitates efficient parallel HPO with zero-cost benchmarks. Our approach calculates the exact return order based on the information stored in file system, eliminating the need for long waiting times and enabling much faster HPO evaluations. We first verify the correctness of our approach through extensive testing and the experiments with 6 popular HPO libraries show its applicability to diverse libraries and its ability to achieve over 1000x speedup compared to a traditional approach. Our package can be installed via pip install mfhpo-simulator., Comment: Accepted to AutoML Conference 2024 ABCD Track

We report on the stabilization of ferromagnetic skyrmions in zero external magnetic fields, in exchange-biased systems composed of ferromagnetic-antiferromagnetic (FM-AFM) bilayers. By performing atomistic spin dynamics simulations, we study cases of compensated, uncompensated, and partly uncompensated FM-AFM interfaces, and investigate the impact of important parameters such as temperature, inter-plane exchange interaction, Dzyaloshinskii-Moria interaction, and magnetic anisotropy on the skyrmions appearance and stability. The model with an uncompensated FM-AFM interface leads to the stabilization of individual skyrmions and skyrmion lattices in the FM layer, caused by the effective field from the AFM instead of an external magnetic field. Similarly, in the case of a fully compensated FM-AFM interface, we show that FM skyrmions can be stabilized. We also demonstrate that accounting for interface roughness leads to stabilization of skyrmions both in compensated and uncompensated interfaces. Moreover, in bilayers with a rough interface, skyrmions in the FM layer are observed for a wide range of exchange interaction values through the FM-AFM interface, and the chirality of the skyrmions depends critically on the exchange interaction., Comment: 8 pages, 8 figures