Search

We present universal relations between entanglement entropy, which quantifies the quantum correlation between subsystems, and the elastic cross section, which is the primary observable for high energy particle scattering, by employing a careful formulation of wave packets for the incoming particles. For 2-to-2 elastic scattering with no initial entanglement and subdividing the system along particle labels, we show that both the R\'enyi and Tsallis entropies in the final states are directly proportional to the elastic cross section in unit of the transverse size for the initial wave packets, which is then interpreted as the elastic scattering probability. The relations do not depend on the underlying dynamics of the quantum field theory and are valid to all orders in coupling strengths. Furthermore, computing quantum correlations between momentum and non-kinematic data leads to entanglement entropies expressed as various semi-inclusive elastic cross sections. Our result gives rise to a novel ``area law'' for entanglement entropy in a two-body system., Comment: 32 pages + appendices, 10 figures

In this work, we study the cosmological effects of a tower of warm dark matter states on the cosmic microwave background (CMB) and on large-scale structure (LSS). For concreteness, we consider the $N$naturalness model, which is a proposed mechanism to solve the Higgs hierarchy problem. In this framework, the sector of particles of the Standard Model is copied $N$ times where the Higgs mass-squared value is the only parameter that changes between sectors. The other sectors are similar to our own, except their particles are proportionally heavier and cooler compared to the Standard Model sector. Since each sector is extremely weakly coupled to other sectors, direct observations of the new particles are not expected. The addition of new photon-like species and new neutrino-like species, however, can be detected through the CMB and in LSS data. These additional neutrinos form a tower of states with increasing mass and decreasing temperature compared to the SM neutrinos. This tower causes a more gradual suppression of the matter power spectrum across different comoving wavenumbers than a single warm dark matter state would. We quantitatively explore these effects in the $N$naturalness model and compute the parameter space allowed by the Planck 2018, weak lensing, and Lyman-$\alpha$ datasets. Depending on the underlying parameters, Planck 2018 and weak lensing data can require $N$naturalness to be tuned at the $10\%$ level for Dirac neutrinos and at the $5\%$ level for Majorana neutrinos. The additional neutrino states are crucial in constraining the model, particularly through their suppression of the power spectrum at small scales. The inclusion of many warm dark matter species is computationally very challenging and we make detailed assessments of our approximations and comment on potential future improvements., Comment: 36 pages, 15 captioned figures, 4 tables

Quantum algorithms for linear systems produce the solution state $A^{-1}|b\rangle$ by querying two oracles: $O_A$ that block encodes the coefficient matrix and $O_b$ that prepares the initial state. We present a quantum linear system algorithm making $\mathbf{\Theta}\left(1/\sqrt{p}\right)$ queries to $O_b$, which is optimal in the success probability, and $\mathbf{O}\left(\kappa\log\left(1/p\right)\left(\log\log\left(1/p\right)+\log\left({1}/{\epsilon}\right)\right)\right)$ queries to $O_A$, nearly optimal in all parameters including the condition number and accuracy. Notably, our complexity scaling of initial state preparation holds even when $p$ is not known $\textit{a priori}$. This contrasts with recent results achieving $\mathbf{O}\left(\kappa\log\left({1}/{\epsilon}\right)\right)$ complexity to both oracles, which, while optimal in $O_A$, is highly suboptimal in $O_b$ as $\kappa$ can be arbitrarily larger than $1/\sqrt{p}$. In various applications such as solving differential equations, preparing ground states of operators with real spectra, and estimating and transforming eigenvalues of non-normal matrices, we can further improve the dependence on $p$ using a block preconditioning scheme to nearly match or outperform best previous results based on other methods, which also furnishes an extremely simple quantum linear system algorithm with an optimal query complexity to $O_A$. Underlying our results is a new Variable Time Amplitude Amplification algorithm with Tunable thresholds (Tunable VTAA), which fully characterizes generic nested amplitude amplifications, improves the $\ell_1$-norm input cost scaling of Ambainis to an $\ell_{\frac{2}{3}}$-quasinorm scaling, and admits a deterministic amplification schedule for the quantum linear system problem., Comment: 86 pages, 3 figures

The AEOS project introduces a series of high-resolution cosmological simulations that model star-by-star chemical enrichment and galaxy formation in the early Universe, achieving 1 pc resolution. These simulations capture the complexities of galaxy evolution within the first ~300 Myr by modeling individual stars and their feedback processes. By incorporating chemical yields from individual stars, AEOS generates galaxies with diverse stellar chemical abundances, linking them to hierarchical galaxy formation and early nucleosynthetic events. These simulations underscore the importance of chemical abundance patterns in ancient stars as vital probes of early nucleosynthesis, star formation histories, and galaxy formation. We examine the metallicity floors of various elements resulting from Pop III enrichment, providing best-fit values for eight different metals (e.g., [O/H] = -4.0) to guide simulations without Pop III models. Additionally, we identify galaxies that begin star formation with Pop II after external enrichment and investigate the frequency of CEMP stars at varying metallicities. The AEOS simulations offer detailed insights into the relationship between star formation, feedback, and chemical enrichment. Future work will extend these simulations to later epochs to interpret the diverse stellar populations of the Milky Way and its satellites., Comment: 16 pages, 10 figures, submitted to ApJ

The interpretation of groups of particle spins at colliders as quantum states has opened up the possibility of using colliders for quantum information. While most efforts have focused on utilizing the decays of the particles to infer their spins to reconstruct the quantum density matrix, we show that the production kinematics of the particles provides sufficient information about the spins to establish quantum tomography without using the decays. We perform a comparative study, highlighting the advantages and disadvantages of using this "kinematic approach" relative to the usual "decay approach." Since the kinematic approach leverages the simplicity of scattering kinematics, this approach promises to achieve the optimal statistical results for quantum tomography at colliders.

We study the evolution of populations of binary stars within massive cluster-forming regions. We simulate the formation of young massive star clusters within giant molecular clouds with masses ranging from 2 x 10$^{4}$ to 3.2 x 10$^{5}$ M$_{\odot}$. We use Torch, which couples stellar dynamics, magnetohydrodynamics, star and binary formation, stellar evolution, and stellar feedback through the AMUSE framework. We find that the binary fraction decreases during cluster formation at all molecular cloud masses. The binaries' orbital properties also change, with stronger and quicker changes in denser, more massive clouds. Most of the changes we see can be attributed to the disruption of binaries wider than 100 au, although the close binary fraction also decreases in the densest cluster-forming region. The binary fraction for O stars remains above 90%, but exchanges and dynamical hardening are ubiquitous, indicating that O stars undergo frequent few-body interactions early during the cluster formation process. Changes to the populations of binaries are a by-product of hierarchical cluster assembly: most changes to the binary population take place when the star formation rate is high and there are frequent mergers between sub-clusters in the cluster-forming region. A universal primordial binary distribution based on observed inner companions in the Galactic field is consistent with the binary populations of young clusters with resolved stellar populations, and the scatter between clusters of similar masses could be explained by differences in their formation history., Comment: 15 pages, 8 figures, 3 tables. Submitted to ApJ

Data valuation is a class of techniques for quantitatively assessing the value of data for applications like pricing in data marketplaces. Existing data valuation methods define a value for a discrete dataset. However, in many use cases, users are interested in not only the value of the dataset, but that of the distribution from which the dataset was sampled. For example, consider a buyer trying to evaluate whether to purchase data from different vendors. The buyer may observe (and compare) only a small preview sample from each vendor, to decide which vendor's data distribution is most useful to the buyer and purchase. The core question is how should we compare the values of data distributions from their samples? Under a Huber characterization of the data heterogeneity across vendors, we propose a maximum mean discrepancy (MMD)-based valuation method which enables theoretically principled and actionable policies for comparing data distributions from samples. We empirically demonstrate that our method is sample-efficient and effective in identifying valuable data distributions against several existing baselines, on multiple real-world datasets (e.g., network intrusion detection, credit card fraud detection) and downstream applications (classification, regression)., Comment: Accepted to NeurIPS 2024 as a poster. Main paper with appendix (38 pages in total). Code will be released soon at https://github.com/XinyiYS/Data_Distribution_Valuation

Designing control policies for large, distributed systems is challenging, especially in the context of critical, temporal logic based specifications (e.g., safety) that must be met with high probability. Compositional methods for such problems are needed for scalability, yet relying on worst-case assumptions for decomposition tends to be overly conservative. In this work, we use the framework of Constrained Markov Decision Processes (CMDPs) to provide an assume-guarantee based decomposition for synthesizing decentralized control policies, subject to logical constraints in a multi-agent setting. The returned policies are guaranteed to satisfy the constraints with high probability and provide a lower bound on the achieved objective reward. We empirically find the returned policies to achieve near-optimal rewards while enjoying an order of magnitude reduction in problem size and execution time., Comment: 6 pages, 1 figure, accepted for publication at the 63rd IEEE Conf. on Decision and Control (2024)

Power system operators must ensure that dispatch decisions remain feasible in case of grid outages or contingencies to prevent cascading failures and ensure reliable operation. However, checking the feasibility of all $N - k$ contingencies -- every possible simultaneous failure of $k$ grid components -- is computationally intractable for even small $k$, requiring system operators to resort to heuristic screening methods. Because of the increase in uncertainty and changes in system behaviors, heuristic lists might not include all relevant contingencies, generating false negatives in which unsafe scenarios are misclassified as safe. In this work, we propose to use input-convex neural networks (ICNNs) for contingency screening. We show that ICNN reliability can be determined by solving a convex optimization problem, and by scaling model weights using this problem as a differentiable optimization layer during training, we can learn an ICNN classifier that is both data-driven and has provably guaranteed reliability. Namely, our method can ensure a zero false negative rate. We empirically validate this methodology in a case study on the IEEE 39-bus test network, observing that it yields substantial (10-20x) speedups while having excellent classification accuracy., Comment: 11 pages, 4 figures

The stark contrast in the design philosophy of an event camera makes it particularly ideal for operating under high-speed, high dynamic range and low-light conditions, where standard cameras underperform. Nonetheless, event cameras still suffer from some amount of motion blur, especially under these challenging conditions, in contrary to what most think. This is attributed to the limited bandwidth of the event sensor pixel, which is mostly proportional to the light intensity. Thus, to ensure that event cameras can truly excel in such conditions where it has an edge over standard cameras, it is crucial to account for event motion blur in downstream applications, especially reconstruction. However, none of the recent works on reconstructing Neural Radiance Fields (NeRFs) from events, nor event simulators, have considered the full effects of event motion blur. To this end, we propose, Deblur e-NeRF, a novel method to directly and effectively reconstruct blur-minimal NeRFs from motion-blurred events generated under high-speed motion or low-light conditions. The core component of this work is a physically-accurate pixel bandwidth model proposed to account for event motion blur under arbitrary speed and lighting conditions. We also introduce a novel threshold-normalized total variation loss to improve the regularization of large textureless patches. Experiments on real and novel realistically simulated sequences verify our effectiveness. Our code, event simulator and synthetic event dataset will be open-sourced., Comment: Accepted to ECCV 2024. Project website is accessible at https://wengflow.github.io/deblur-e-nerf

In this conference summary I first provide some historical notes on my own collaboration with You-Hua Chu and on the discovery of X-rays from superbubbles by Margarita Rosado S. I then consider the central subject of the conference, stellar feedback, and how interactions between stellar winds and the interstellar medium can limit or enhance the effects of feedback compared to models including only supernova explosions. Finally, I review results in other areas covered by the conference, including planet and star formation, nova and supernova remnants, different topics in stellar evolution and interaction with the interstellar medium, star clusters, observational surveys, and observational and numerical techniques., Comment: 10 pages, for publication in RevMexAA (Ser Conf) proceedings of the conference

Pre-trained deep learning embeddings have consistently shown superior performance over handcrafted acoustic features in speech emotion recognition (SER). However, unlike acoustic features with clear physical meaning, these embeddings lack clear interpretability. Explaining these embeddings is crucial for building trust in healthcare and security applications and advancing the scientific understanding of the acoustic information that is encoded in them. This paper proposes a modified probing approach to explain deep learning embeddings in the SER space. We predict interpretable acoustic features (e.g., f0, loudness) from (i) the complete set of embeddings and (ii) a subset of the embedding dimensions identified as most important for predicting each emotion. If the subset of the most important dimensions better predicts a given emotion than all dimensions and also predicts specific acoustic features more accurately, we infer those acoustic features are important for the embedding model for the given task. We conducted experiments using the WavLM embeddings and eGeMAPS acoustic features as audio representations, applying our method to the RAVDESS and SAVEE emotional speech datasets. Based on this evaluation, we demonstrate that Energy, Frequency, Spectral, and Temporal categories of acoustic features provide diminishing information to SER in that order, demonstrating the utility of the probing classifier method to relate embeddings to interpretable acoustic features.

Quantum computation for chemical problems will require the construction of guiding states with sufficient overlap with a target state. Since easily available and initializable mean-field states are characterized by an overlap that is reduced for multi-configurational electronic structures and even vanishes with growing system size, we here investigate the severity of state preparation for reaction chemistry. We emphasize weaknesses in current traditional approaches (even for weakly correlated molecules) and highlight the advantage of quantum phase estimation algorithms. An important result is the introduction of a new classification scheme for electronic structures based on orbital entanglement information. We identify two categories of multi-configurational molecules. Whereas class-1 molecules are dominated by very few determinants and often found in reaction chemistry, class-2 molecules do not allow one to single out a reasonably sized number of important determinants. The latter are particularly hard for traditional approaches and an ultimate target for quantum computation. Some open-shell iron-sulfur clusters belong to class 2. We discuss the role of the molecular orbital basis set and show that true class-2 molecules remain in this class independent of the choice of the orbital basis, with the iron-molybdenum cofactor of nitrogenase being a prototypical example. We stress that class-2 molecules can be build in a systematic fashion from open-shell centers or unsaturated carbon atoms. Our key result is that it will always be possible to initialize a guiding state for chemical reaction chemistry in the ground state based on initial low-cost approximate electronic structure information, which is facilitated by the finite size of the atomistic structures to be considered., Comment: 58 pages, 10 figures, 3 tables

The Brewster effect, dating back to the pioneering work of Sir David Brewster in 1815, offers a crucial route to achieve 100% energy conversion between the incident and transmitted propagating waves at an optical interface and is of fundamental importance to many practical applications, such as polarization filtering, beam steering, and optical broadband angular selectivity. However, whether the Brewster effect of surface waves can be implemented without the involvement of negative-permittivity or negative-permeability materials remains elusive. This is due to the formidable challenge to fully suppress both the parasitic scattering into propagating waves and the reflection into surface waves under the incidence of surface waves. Here, we reveal a feasible route to achieve scattering-free plasmonic Brewster effect via isotropic metasurfaces, along with the usage of positive-permittivity and positive-permeability metamaterials with both anisotropic and magnetic responses. In essence, the anisotropic response of metamaterials is judiciously designed to fully suppress the parasitic scattering into propagating waves, while the magnetic response of metamaterials facilitates the full suppression of the reflection into surface waves supported by metasurfaces. Moreover, we find that this plasmonic Brewster effect via metasurfaces can be further engineered to occur for arbitrary incident angles, giving rise to the exotic phenomenon of all-angle scattering-free plasmonic Brewster effect.

Research on microwave signal measurement techniques is risen, driven by the expanding urgent demands of wireless communication, global positioning systems, remote sensing and 6G networks. In stark contrast with traditional electronic-based realization, the implementations of microwave signal measurement systems based on integrated compact photonic chip have exhibited distinct advantages in high operation bandwidth, light weight, and strong immunity to electromagnetic interference. However, although numerous integrated microwave photonic signal measurement systems have been reported, measurement bandwidth of the majority of them is still below 30 GHz due to the bandwidth limitation of electro-optical modulators (EOMs). Furthermore, previous studies often are more focused on the measurement of one single parameter (typically the frequency) of microwave signals, which has hindered their practical application in complex situations. Here, an integrated photonic microwave multi-parameter measurement system composed of microwave frequency measurement module and microwave phase amplitude measurement module based on thin-film lithium niobate (TFLN) platform is reported. Utilizing this system, not only the ultra-high bandwidth (up to 60GHz) of microwave frequency, phase and amplitude measurement with low root-mean-squares errors (450MHz, 3.43{\deg} and 1.64% of the measurement for frequency, phase and amplitude, respectively), but also the time-domain reconstruction of sinusoidal microwave signals is achieved. This demonstration further broadens the application of integrated TFLN photonic devices in microwave signal measurement technology to address the bandwidth bottleneck of the ever-growing microwave networks in the future information society., Comment: 23 pages,3 figures

We demonstrate that two-dimensional Kramers-Weyl fermions can be engineered in spin-orbit coupled twisted bilayers, where the chiral structure of these moir\'e systems breaks all mirror symmetries, confining Kramers-Weyl fermions to high-symmetry points in the Brillouin zone under time reversal symmetry. Our theoretical analysis reveals a symmetry-enforced Weyl-like interlayer moir\'e coupling that universally ensures an ideal radial spin-texture at arbitrary twist angles, under $C_{nz}$ symmetry with n>2. First principles density functional calculation confirm the realization of these fermions in twisted $\alpha$-In$_2$Se$_3$ bilayers, where flat bands and out-of-plane ferroelectric polarization in each layer guarantee two-dimensional Kramers-Weyl physics with perfectly ideal radial spin textures.

Large Language Models (LLMs) have become indispensable in numerous real-world applications. Unfortunately, fine-tuning these models at scale, especially in federated settings where data privacy and communication efficiency are critical, presents significant challenges. Existing methods often resort to parameter-efficient fine-tuning (PEFT) to mitigate communication overhead, but this typically comes at the cost of model accuracy. To address these limitations, we propose federated full-parameter tuning at scale for LLMs (Ferret), the first first-order method with shared randomness to enable scalable full-parameter tuning of LLMs across decentralized data sources while maintaining competitive model accuracy. Ferret accomplishes this through three aspects: (1) it employs widely applied first-order methods for efficient local updates; (2) it projects these updates into a low-dimensional space to considerably reduce communication overhead; and (3) it reconstructs local updates from this low-dimensional space with shared randomness to facilitate effective full-parameter global aggregation, ensuring fast convergence and competitive final performance. Our rigorous theoretical analyses and insights along with extensive experiments, show that Ferret significantly enhances the scalability of existing federated full-parameter tuning approaches by achieving high computational efficiency, reduced communication overhead, and fast convergence, all while maintaining competitive model accuracy. Our implementation is available at https://github.com/allen4747/Ferret.

We demonstrate the first end-to-end integration of high-performance computing (HPC), reliable quantum computing, and AI in a case study on catalytic reactions producing chiral molecules. We present a hybrid computation workflow to determine the strongly correlated reaction configurations and estimate, for one such configuration, its active site's ground state energy. We combine 1) the use of HPC tools like AutoRXN and AutoCAS to systematically identify the strongly correlated chemistry within a large chemical space with 2) the use of logical qubits in the quantum computing stage to prepare the quantum ground state of the strongly correlated active site, demonstrating the advantage of logical qubits compared to physical qubits, and 3) the use of optimized quantum measurements of the logical qubits with so-called classical shadows to accurately predict various properties of the ground state including energies. The combination of HPC, reliable quantum computing, and AI in this demonstration serves as a proof of principle of how future hybrid chemistry applications will require integration of large-scale quantum computers with classical computing to be able to provide a measurable quantum advantage., Comment: 15 pages, 9 figures

Mass segregation is seen in many star clusters, but whether massive stars form in the center of a cluster or migrate there dynamically is still debated. N-body simulations have shown that early dynamical mass segregation is possible when sub-clusters merge to form a dense core with a small crossing time. However, the effect of gas dynamics on both the formation and dynamics of the stars could inhibit the formation of the dense core. We aim to study the dynamical mass segregation of star cluster models that include gas dynamics and self-consistently form stars from the dense substructure in the gas. Our models use the Torch framework, which is based on AMUSE and includes stellar and magnetized gas dynamics, as well as stellar evolution and feedback from radiation, stellar winds, and supernovae. Our models consist of three star clusters forming from initial turbulent spherical clouds of mass $10^{4,5,6}\rm~M_\odot$ and radius $11.7\rm~pc$ that have final stellar masses of $3.6\times10^3\rm~M_\odot$, $6.5\times10^4\rm~M_\odot$, and $8.9\times10^5\rm~M_\odot$, respectively. There is no primordial mass segregation in the model by construction. All three clusters become dynamically mass segregated at early times via collapse confirming that this mechanism occurs within sub-clusters forming directly out of the dense substructure in the gas. The dynamics of the embedded gas and stellar feedback do not inhibit the collapse of the cluster. We find that each model cluster becomes mass segregated within $2~$Myr of the onset of star formation, reaching the levels observed in young clusters in the Milky Way. However, we note that the exact values are highly time-variable during these early phases of evolution. Massive stars that segregate to the center during core collapse are likely to be dynamically ejected, a process that can decrease the overall level of mass segregation again., Comment: Submitted to A&A. 8 pages, 4 figures

This paper considers a novel online fair division problem involving multiple agents in which a learner observes an indivisible item that has to be irrevocably allocated to one of the agents while satisfying a fairness and efficiency constraint. Existing algorithms assume a small number of items with a sufficiently large number of copies, which ensures a good utility estimation for all item-agent pairs. However, such an assumption may not hold in many real-life applications, e.g., an online platform that has a large number of users (items) who only use the platform's service providers (agents) a few times (a few copies of items), which makes it difficult to estimate the utility for all item-agent pairs. To overcome this challenge, we model the online fair division problem using contextual bandits, assuming the utility is an unknown function of the item-agent features. We then propose algorithms for online fair division with sub-linear regret guarantees. Our experimental results also verify the different performance aspects of the proposed algorithms., Comment: We study an online fair division problem that has a large number of items with only a few copies of each item and propose contextual bandits-based algorithms with sub-linear regret guarantees

The modeling of dynamical systems is a pervasive concern for not only describing but also predicting and controlling natural phenomena and engineered systems. Current data-driven approaches often assume prior knowledge of the relevant state variables or result in overparameterized state spaces. Boyuan Chen and his co-authors proposed a neural network model that estimates the degrees of freedom and attempts to discover the state variables of a dynamical system. Despite its innovative approach, this baseline model lacks a connection to the physical principles governing the systems it analyzes, leading to unreliable state variables. This research proposes a method that leverages the physical characteristics of second-order Hamiltonian systems to constrain the baseline model. The proposed model outperforms the baseline model in identifying a minimal set of non-redundant and interpretable state variables.

Existing sample-based methods, like influence functions and representer points, measure the importance of a training point by approximating the effect of its removal from training. As such, they are skewed towards outliers and points that are very close to the decision boundaries. The explanations provided by these methods are often static and not specific enough for different test points. In this paper, we propose a method to generate an explanation in the form of support spectrums which are based on two main ideas: the support sets and a global-to-local importance measure. The support set is the set of training points, in the predicted class, that ``lie in between'' the test point and training points in the other classes. They indicate how well the test point can be distinguished from the points not in the predicted class. The global-to-local importance measure is obtained by decoupling existing methods into the global and local components which are then used to select the points in the support set. Using this method, we are able to generate explanations that are tailored to specific test points. In the experiments, we show the effectiveness of the method in image classification and text generation tasks.

This research study uses hierarchical logistic regression to identify predictors of first-class STEM degree outcomes at a research-intensive Russell Group university in the UK between 2012 and 2022. By building a multivariate binary logistic model with random intercepts for different STEM degree subjects, we find that prior academic attainment, ethnicity, gender, socioeconomic status, age, and course duration are statistically significant predictors of achieving a first-class degree. By determining the odds ratios and average marginal effects of socio-demographic predictors, we find evidence for the existence of age, ethnicity, gender, and socioeconomic awarding gaps. The largest awarding gap exists between Black and White students, with Black students having 0.45 (95\% CI: 0.30-0.68) times the odds, and a 14\% lower probability, of achieving a first-class degree compared to White students, holding all other variables constant. Students who graduate from 4-year degrees are found to have, on average, a 27\% higher probability of achieving a first-class degree than students on 3-year degrees. Despite raw data suggesting that male students outperform female students, the multivariate hierarchical analysis revealed higher odds for female students after controlling for other factors and accounting for nested data structures. Analysis using year-specific average marginal effects indicates that awarding gaps have not significantly changed between 2012 and 2022. This research study provides a robust analytical framework for use by other departments and institutions aiming to identify and address awarding gaps., Comment: 20 pages

WalledEval is a comprehensive AI safety testing toolkit designed to evaluate large language models (LLMs). It accommodates a diverse range of models, including both open-weight and API-based ones, and features over 35 safety benchmarks covering areas such as multilingual safety, exaggerated safety, and prompt injections. The framework supports both LLM and judge benchmarking and incorporates custom mutators to test safety against various text-style mutations, such as future tense and paraphrasing. Additionally, WalledEval introduces WalledGuard, a new, small, and performant content moderation tool, and two datasets: SGXSTest and HIXSTest, which serve as benchmarks for assessing the exaggerated safety of LLMs and judges in cultural contexts. We make WalledEval publicly available at https://github.com/walledai/walledeval., Comment: Under review

In this work, we derive a generalized constitutive relation describing the current response to external electromagnetic fields in electrically biased quantum materials. While our semiclassical Boltzmann approach reveals the existence of electro-optic effects induced by the Berry curvature dipole of Bloch electrons, we also find a wealth of alternative electro-optic effects originating from the interplay between Berry curvature and magnetic moment. In particular, our symmetry analysis indicates the existence of a magnetoelectric electro-optic effect, derived from the simultaneous presence of Berry curvature and magnetic moment, that requires either time-reversal or inversion symmetry breaking. The revealed conductivity coefficients are explicitly written and we derive the tensor shape describing such alternative electromagnetic responses for chiral materials pertaining to space groups 152 and 198.

The Blatter radical has been suggested as a building block in future molecular spintronic devices due to its radical character and expected long-spin lifetime. However, whether and how the radical character manifests itself in the charge transport and magnetotransport properties seems to depend on the environment or has not yet been studied. Here, we investigate single-molecule junctions of the Blatter radical molecule in a mechanically controlled break junction device at low temperature. Differential conductance spectroscopy on individual junctions shows two types of zero-bias anomalies attributed to Kondo resonances revealing the ability to retain the open-shell nature of the radical molecule in a two-terminal device. Additionally, a high negative magnetoresistance is also observed in junctions without showing a zero-bias peak. We posit that the high negative magnetoresistance is due to the effect of a singlet-triplet Kondo effect under magnetic field originating from a double-quantum-dot system consisting of a Blatter radical molecule with a strong correlation to a second side-coupled molecule. Our findings not only provide the possibility of using the Blatter radical in a two-terminal system under cryogenic conditions but also reveal the magnetotransport properties emerging from different configurations of the molecule inside a junction., Comment: 17 pages, 5 figures

The relation between magnetic field strength B and gas density n in the interstellar medium is of fundamental importance to many areas of astrophysics, from protostellar disks to galaxy evolution. We present and compare Bayesian analyses of the B - n relation for a comprehensive observational data set, as well as a large body of numerical MHD simulations. We extend the original Zeeman relation of Crutcher et al. (2010) with a large body of magnetic data that includes 700 observations with the Davis-Chandrasekhar-Fermi method. By using a new multiparameter Bayesian analysis we present a new, more general, time-averaged observational relation: B \propto n^{0.27 \pm 0.017} for n \leq n_0 and B \propto n^{0.54 \pm 0.18} for n \geq n_0, with n_0 = 924^(+145-144) cm^-3. We perform a separate analysis on 19 numerical magnetohydrodynamics simulations that cover a wide range of scales, resolutions, initial conditions, and completed with a variety of codes: arepo, flash, pencil, and ramses. The power law exponents derived from the simulations depend on several physical factors including: dynamo effects, time scales, turbulence, and the initial seed field strength. In particular, early-time simulations where the density, velocity and magnetic fields are unevolved do not match the observational scalings. The simulations that trace the observed density range best, the evolved dwarf galaxy and Milky Way like galaxy simulations, settle into a near consistent exponent of = 0.5 in the dense gas, with variability in the diffuse gas exponent., Comment: 23 figures, 28 pages, submitted to MNRAS, Comments welcome

Contextual dueling bandit is used to model the bandit problems, where a learner's goal is to find the best arm for a given context using observed noisy preference feedback over the selected arms for the past contexts. However, existing algorithms assume the reward function is linear, which can be complex and non-linear in many real-life applications like online recommendations or ranking web search results. To overcome this challenge, we use a neural network to estimate the reward function using preference feedback for the previously selected arms. We propose upper confidence bound- and Thompson sampling-based algorithms with sub-linear regret guarantees that efficiently select arms in each round. We then extend our theoretical results to contextual bandit problems with binary feedback, which is in itself a non-trivial contribution. Experimental results on the problem instances derived from synthetic datasets corroborate our theoretical results., Comment: Accepted at ICML 2024 Workshop on Foundations of Reinforcement Learning and Control

Large language models (LLMs) are widely used in decision-making, but their reliability, especially in critical tasks like healthcare, is not well-established. Therefore, understanding how LLMs reason and make decisions is crucial for their safe deployment. This paper investigates how the uncertainty of responses generated by LLMs relates to the information provided in the input prompt. Leveraging the insight that LLMs learn to infer latent concepts during pretraining, we propose a prompt-response concept model that explains how LLMs generate responses and helps understand the relationship between prompts and response uncertainty. We show that the uncertainty decreases as the prompt's informativeness increases, similar to epistemic uncertainty. Our detailed experimental results on real datasets validate our proposed model., Comment: 27 pages, 13 figures

The International Muon Collider Collaboration (IMCC) [1] was established in 2020 following the recommendations of the European Strategy for Particle Physics (ESPP) and the implementation of the European Strategy for Particle Physics-Accelerator R&D Roadmap by the Laboratory Directors Group [2], hereinafter referred to as the the European LDG roadmap. The Muon Collider Study (MuC) covers the accelerator complex, detectors and physics for a future muon collider. In 2023, European Commission support was obtained for a design study of a muon collider (MuCol) [3]. This project started on 1st March 2023, with work-packages aligned with the overall muon collider studies. In preparation of and during the 2021-22 U.S. Snowmass process, the muon collider project parameters, technical studies and physics performance studies were performed and presented in great detail. Recently, the P5 panel [4] in the U.S. recommended a muon collider R&D, proposed to join the IMCC and envisages that the U.S. should prepare to host a muon collider, calling this their "muon shot". In the past, the U.S. Muon Accelerator Programme (MAP) [5] has been instrumental in studies of concepts and technologies for a muon collider., Comment: This document summarises the International Muon Collider Collaboration (IMCC) progress and status of the Muon Collider R&D programme

The rapid evolution of large language models (LLMs) represents a substantial leap forward in natural language understanding and generation. However, alongside these advancements come significant challenges related to the accountability and transparency of LLM responses. Reliable source attribution is essential to adhering to stringent legal and regulatory standards, including those set forth by the General Data Protection Regulation. Despite the well-established methods in source attribution within the computer vision domain, the application of robust attribution frameworks to natural language processing remains underexplored. To bridge this gap, we propose a novel and versatile TRansformer-based Attribution framework using Contrastive Embeddings called TRACE that, in particular, exploits contrastive learning for source attribution. We perform an extensive empirical evaluation to demonstrate the performance and efficiency of TRACE in various settings and show that TRACE significantly improves the ability to attribute sources accurately, making it a valuable tool for enhancing the reliability and trustworthiness of LLMs.

Protecting intellectual property (IP) of text such as articles and code is increasingly important, especially as sophisticated attacks become possible, such as paraphrasing by large language models (LLMs) or even unauthorized training of LLMs on copyrighted text to infringe such IP. However, existing text watermarking methods are not robust enough against such attacks nor scalable to millions of users for practical implementation. In this paper, we propose Waterfall, the first training-free framework for robust and scalable text watermarking applicable across multiple text types (e.g., articles, code) and languages supportable by LLMs, for general text and LLM data provenance. Waterfall comprises several key innovations, such as being the first to use LLM as paraphrasers for watermarking along with a novel combination of techniques that are surprisingly effective in achieving robust verifiability and scalability. We empirically demonstrate that Waterfall achieves significantly better scalability, robust verifiability, and computational efficiency compared to SOTA article-text watermarking methods, and showed how it could be directly applied to the watermarking of code. We also demonstrated that Waterfall can be used for LLM data provenance, where the watermarks of LLM training data can be detected in LLM output, allowing for detection of unauthorized use of data for LLM training and potentially enabling model-centric watermarking of open-sourced LLMs which has been a limitation of existing LLM watermarking works. Our code is available at https://github.com/aoi3142/Waterfall., Comment: Accepted to EMNLP 2024 Main Conference

A top quark and an anti-top quark produced together at colliders have correlated spins. These spins constitute a quantum state that can exhibit entanglement and violate Bell's inequality. In realistic collider experiments, most analyses allow the axes, as well the Lorentz frame to vary event-by-event, thus introducing a dependence on the choice of event-dependent basis leading us to adopt "fictitious states," rather than genuine quantum states. The basis dependence of fictitious states allows for an optimization procedure, which makes the usage of fictitious states advantageous in measuring entanglement and Bell inequality violation. In this work, we show analytically that the basis which diagonalizes the spin-spin correlations is optimal for maximizing spin correlations, entanglement, and Bell inequality violation. We show that the optimal basis is approximately the same as the fixed beam basis (or the rotated beam basis) near the $t\bar t$ production threshold, while it approaches the helicity basis far above threshold. Using this basis, we present the sensitivity for entanglement and Bell inequality violation in $t\bar t$ events at the LHC and a future $e^+e^-$ collider. Since observing Bell inequality violation appears to be quite challenging experimentally, and requires a large dataset in collider experiments, choosing the optimal basis is crucially important to observe Bell inequality violation. Our method and general approach are equally applicable to other systems beyond $t \bar t$, including interactions beyond the Standard Model., Comment: 49 pages, 11 figures, 1 table

Evidence to guide healthcare decisions is often limited by a lack of relevant and trustworthy literature as well as difficulty in contextualizing existing research for a specific patient. Large language models (LLMs) could potentially address both challenges by either summarizing published literature or generating new studies based on real-world data (RWD). We evaluated the ability of five LLM-based systems in answering 50 clinical questions and had nine independent physicians review the responses for relevance, reliability, and actionability. As it stands, general-purpose LLMs (ChatGPT-4, Claude 3 Opus, Gemini Pro 1.5) rarely produced answers that were deemed relevant and evidence-based (2% - 10%). In contrast, retrieval augmented generation (RAG)-based and agentic LLM systems produced relevant and evidence-based answers for 24% (OpenEvidence) to 58% (ChatRWD) of questions. Only the agentic ChatRWD was able to answer novel questions compared to other LLMs (65% vs. 0-9%). These results suggest that while general-purpose LLMs should not be used as-is, a purpose-built system for evidence summarization based on RAG and one for generating novel evidence working synergistically would improve availability of pertinent evidence for patient care., Comment: 28 pages (2 figures, 3 tables) inclusive of 8 pages of supplemental materials (4 supplemental figures and 4 supplemental tables)

Quantum access to arbitrary classical data encoded in unitary black-box oracles underlies interesting data-intensive quantum algorithms, such as machine learning or electronic structure simulation. The feasibility of these applications depends crucially on gate-efficient implementations of these oracles, which are commonly some reversible versions of the boolean circuit for a classical lookup table. We present a general parameterized architecture for quantum circuits implementing a lookup table that encompasses all prior work in realizing a continuum of optimal tradeoffs between qubits, non-Clifford gates, and error resilience, up to logarithmic factors. Our architecture assumes only local 2D connectivity, yet recovers results that previously required all-to-all connectivity, particularly, with the appropriate parameters, poly-logarithmic error scaling. We also identify novel regimes, such as simultaneous sublinear scaling in all parameters. These results enable tailoring implementations of the commonly used lookup table primitive to any given quantum device with constrained resources.

With the widespread applications of neural networks (NNs) trained on personal data, machine unlearning has become increasingly important for enabling individuals to exercise their personal data ownership, particularly the "right to be forgotten" from trained NNs. Since retraining is computationally expensive, we seek approximate unlearning algorithms for NNs that return identical models to the retrained oracle. While Newton's method has been successfully used to approximately unlearn linear models, we observe that adapting it for NN is challenging due to degenerate Hessians that make computing Newton's update impossible. Additionally, we show that when coupled with popular techniques to resolve the degeneracy, Newton's method often incurs offensively large norm updates and empirically degrades model performance post-unlearning. To address these challenges, we propose CureNewton's method, a principle approach that leverages cubic regularization to handle the Hessian degeneracy effectively. The added regularizer eliminates the need for manual finetuning and affords a natural interpretation within the unlearning context. Experiments across different models and datasets show that our method can achieve competitive unlearning performance to the state-of-the-art algorithm in practical unlearning settings, while being theoretically justified and efficient in running time.

This position paper proposes a data-centric viewpoint of AI research, focusing on large language models (LLMs). We start by making the key observation that data is instrumental in the developmental (e.g., pretraining and fine-tuning) and inferential stages (e.g., in-context learning) of LLMs, and yet it receives disproportionally low attention from the research community. We identify four specific scenarios centered around data, covering data-centric benchmarks and data curation, data attribution, knowledge transfer, and inference contextualization. In each scenario, we underscore the importance of data, highlight promising research directions, and articulate the potential impacts on the research community and, where applicable, the society as a whole. For instance, we advocate for a suite of data-centric benchmarks tailored to the scale and complexity of data for LLMs. These benchmarks can be used to develop new data curation methods and document research efforts and results, which can help promote openness and transparency in AI and LLM research., Comment: Preprint

Interleaved practice enhances the memory and problem-solving ability of students in undergraduate courses. We introduce a personalized learning tool built on a Large Language Model (LLM) that can provide immediate and personalized attention to students as they complete homework containing problems interleaved from undergraduate physics courses. Our tool leverages the dimensional analysis method, enhancing students' qualitative thinking and problem-solving skills for complex phenomena. Our approach combines LLMs for symbolic regression with dimensional analysis via prompt engineering and offers students a unique perspective to comprehend relationships between physics variables. This fosters a broader and more versatile understanding of physics and mathematical principles and complements a conventional undergraduate physics education that relies on interpreting and applying established equations within specific contexts. We test our personalized learning tool on the equations from Feynman's lectures on physics. Our tool can correctly identify relationships between physics variables for most equations, underscoring its value as a complementary personalized learning tool for undergraduate physics students.

Fluctuation theorems are generalisations of the second law that describe the relations between work, temperature, and free energy in thermodynamic processes and are used extensively in studies of irreversibility and entropy. Many experiments have verified these relations for different physical systems in the setting of thermodynamics. In this study, we observe the same behavior away from physical thermodynamics, namely for the card game UNO, by performing numerical simulations of the game. As the analog of work, we choose the number of steps one player needs to effect a transition in her deck; the other players and the remaining cards play the role of a finite, non-Markovian bath. We also compare our observation with is expected for a Markovian random walk., Comment: 10 pages, 8 figures

Photo-doped Mott insulators can exhibit novel photocarrier transport and relaxation dynamics and non-equilibrium phases. However, time-resolved real-space imaging of these processes are still lacking. Here, we use scanning ultrafast electron microscopy (SUEM) to directly visualize the spatial-temporal evolution of photoexcited species in a spin-orbit assisted Mott insulator {\alpha}-RuCl3. At low optical fluences, we observe extremely long hot photocarrier transport time over one nanosecond, almost an order of magnitude longer than any known values in conventional semiconductors. At higher optical fluences, we observe nonlinear features suggesting a photo-induced insulator-to-metal transition, which is unusual in a large-gap Mott insulator. Our results demonstrate the rich physics in a photo-doped Mott insulator that can be extracted from spatial-temporal imaging and showcase the capability of SUEM to sensitively probe photoexcitations in strongly correlated electron systems., Comment: Comments are welcome. Please email feedback to bliao@ucsb.edu

Distributed hybrid rendering (DHR) is a real-time rendering approach that incorporates cloud-based ray tracing with locally rasterized graphics for interactive thin client metaverse and game applications. With cloud assistance, DHR can generate high-fidelity ray-traced graphics contents remotely and deliver them to thin clients with low graphics capability, including standalone extended reality devices and mobile phones, while maintaining interactive frame rates for users under adverse network conditions. DHR can already achieve the effect of ray-traced hard shadows that form with the occlusion of direct illumination. We enhance the realism of these shadows by softening their edges with the direction of rays traced and approximating the occlusion of indirect illumination by reconstructing ray-traced ambient occlusion with a modified version of spatiotemporal variance-guided filtering. Our technique uses only 20-30% of the bandwidth of remote rendering and is also tolerant of delays of up to 200 ms with only slight distortion to the shadows along object edges.

The increasing complexity of foundational models underscores the necessity for explainability, particularly for fine-tuning, the most widely used training method for adapting models to downstream tasks. Instance attribution, one type of explanation, attributes the model prediction to each training example by an instance score. However, the robustness of instance scores, specifically towards dataset resampling, has been overlooked. To bridge this gap, we propose a notion of robustness on the sign of the instance score. We theoretically and empirically demonstrate that the popular leave-one-out-based methods lack robustness, while the Shapley value behaves significantly better, but at a higher computational cost. Accordingly, we introduce an efficient fine-tuning-free approximation of the Shapley value (FreeShap) for instance attribution based on the neural tangent kernel. We empirically demonstrate that FreeShap outperforms other methods for instance attribution and other data-centric applications such as data removal, data selection, and wrong label detection, and further generalize our scale to large language models (LLMs). Our code is available at https://github.com/JTWang2000/FreeShap., Comment: Accepted to ICML 2024

Large language models (LLMs) have demonstrated remarkable performances in various tasks. However, the performance of LLMs heavily depends on the input prompt, which has given rise to a number of recent works on prompt optimization. However, previous works often require the availability of a numeric score to assess the quality of every prompt. Unfortunately, when a human user interacts with a black-box LLM, attaining such a score is often infeasible and unreliable. Instead, it is usually significantly easier and more reliable to obtain preference feedback from a human user, i.e., showing the user the responses generated from a pair of prompts and asking the user which one is preferred. Therefore, in this paper, we study the problem of prompt optimization with human feedback (POHF), in which we aim to optimize the prompt for a black-box LLM using only human preference feedback. Drawing inspiration from dueling bandits, we design a theoretically principled strategy to select a pair of prompts to query for preference feedback in every iteration, and hence introduce our algorithm named automated POHF (APOHF). We apply our APOHF algorithm to various tasks, including optimizing user instructions, prompt optimization for text-to-image generative models, and response optimization with human feedback (i.e., further refining the response using a variant of our APOHF). The results demonstrate that our APOHF can efficiently find a good prompt using a small number of preference feedback instances. Our code can be found at \url{https://github.com/xqlin98/APOHF}., Comment: Preprint, 18 pages

Large language models (LLMs) have shown impressive capabilities in real-world applications. The capability of in-context learning (ICL) allows us to adapt an LLM to downstream tasks by including input-label exemplars in the prompt without model fine-tuning. However, the quality of these exemplars in the prompt greatly impacts performance, highlighting the need for an effective automated exemplar selection method. Recent studies have explored retrieval-based approaches to select exemplars tailored to individual test queries, which can be undesirable due to extra test-time computation and an increased risk of data exposure. Moreover, existing methods fail to adequately account for the impact of exemplar ordering on the performance. On the other hand, the impact of the instruction, another essential component in the prompt given to the LLM, is often overlooked in existing exemplar selection methods. To address these challenges, we propose a novel method named EASE, which leverages the hidden embedding from a pre-trained language model to represent ordered sets of exemplars and uses a neural bandit algorithm to optimize the sets of exemplars while accounting for exemplar ordering. Our EASE can efficiently find an ordered set of exemplars that performs well for all test queries from a given task, thereby eliminating test-time computation. Importantly, EASE can be readily extended to jointly optimize both the exemplars and the instruction. Through extensive empirical evaluations (including novel tasks), we demonstrate the superiority of EASE over existing methods, and reveal practical insights about the impact of exemplar selection on ICL, which may be of independent interest. Our code is available at https://github.com/ZhaoxuanWu/EASE-Prompt-Optimization., Comment: 28 pages, 1 figure, 35 tables

In-context learning (ICL) allows transformer-based language models that are pre-trained on general text to quickly learn a specific task with a few "task demonstrations" without updating their parameters, significantly boosting their flexibility and generality. ICL possesses many distinct characteristics from conventional machine learning, thereby requiring new approaches to interpret this learning paradigm. Taking the viewpoint of recent works showing that transformers learn in context by formulating an internal optimizer, we propose an influence function-based attribution technique, DETAIL, that addresses the specific characteristics of ICL. We empirically verify the effectiveness of our approach for demonstration attribution while being computationally efficient. Leveraging the results, we then show how DETAIL can help improve model performance in real-world scenarios through demonstration reordering and curation. Finally, we experimentally prove the wide applicability of DETAIL by showing our attribution scores obtained on white-box models are transferable to black-box models in improving model performance.

Two main mechanisms have classically been proposed for the formation of runaway stars. In the binary supernova scenario (BSS), a massive star in a binary explodes as a supernova, ejecting its companion. In the dynamical ejection scenario, a star is ejected during a strong dynamical encounter between multiple stars. We propose a third mechanism for the formation of runaway stars: the subcluster ejection scenario (SCES), where a subset of stars from an infalling subcluster is ejected out of the cluster via a tidal interaction with the contracting gravitational potential of the assembling cluster. We demonstrate the SCES in a star-by-star simulation of the formation of a young massive cluster from a $10^6\rm~M_\odot$ gas cloud using the Torch framework. This star cluster forms hierarchically through a sequence of subcluster mergers determined by the initial turbulent, spherical conditions of the gas. We find that these mergers drive the formation of runaway stars in our model. Late-forming subclusters fall into the central potential, where they are tidally disrupted, forming tidal tails of runaway stars that are distributed highly anisotropically. Runaways formed in the same SCES have similar ages, velocities, and ejection directions. Surveying observations, we identify several SCES candidate groups with anisotropic ejection directions. The SCES is capable of producing runaway binaries: two wide dynamical binaries in infalling subclusters were tightened through ejection. This allows for another velocity kick via subsequent via a subsequent BSS ejection. An SCES-BSS ejection is a possible avenue for the creation of hypervelocity stars unbound to the Galaxy. We expect nonspherical initial gas distributions to increase the number of calculated runaway stars. The observation of groups of runaway stars formed via the SCES can thus reveal the assembly history of their natal clusters., Comment: 11 pages, 10 figures. Accepted for publication in A&A. Abstract abridged for arXiv

We study the design of risk-sensitive online algorithms, in which risk measures are used in the competitive analysis of randomized online algorithms. We introduce the CVaR$_\delta$-competitive ratio ($\delta$-CR) using the conditional value-at-risk of an algorithm's cost, which measures the expectation of the $(1-\delta)$-fraction of worst outcomes against the offline optimal cost, and use this measure to study three online optimization problems: continuous-time ski rental, discrete-time ski rental, and one-max search. The structure of the optimal $\delta$-CR and algorithm varies significantly between problems: we prove that the optimal $\delta$-CR for continuous-time ski rental is $2-2^{-\Theta(\frac{1}{1-\delta})}$, obtained by an algorithm described by a delay differential equation. In contrast, in discrete-time ski rental with buying cost $B$, there is an abrupt phase transition at $\delta = 1 - \Theta(\frac{1}{\log B})$, after which the classic deterministic strategy is optimal. Similarly, one-max search exhibits a phase transition at $\delta = \frac{1}{2}$, after which the classic deterministic strategy is optimal; we also obtain an algorithm that is asymptotically optimal as $\delta \downarrow 0$ that arises as the solution to a delay differential equation., Comment: Accepted for presentation at the Conference on Learning Theory (COLT) 2024. Updated with an additional reference and minor edits