Search

Existing Hardware Trojans (HT) detection methods face several critical limitations: logic testing struggles with scalability and coverage for large designs, side-channel analysis requires golden reference chips, and formal verification methods suffer from state-space explosion. The emergence of Large Language Models (LLMs) offers a promising new direction for HT detection by leveraging their natural language understanding and reasoning capabilities. For the first time, this paper explores the potential of general-purpose LLMs in detecting various HTs inserted in Register Transfer Level (RTL) designs, including SRAM, AES, and UART modules. We propose a novel tool for this goal that systematically assesses state-of-the-art LLMs (GPT-4o, Gemini 1.5 pro, and Llama 3.1) in detecting HTs without prior fine-tuning. To address potential training data bias, the tool implements perturbation techniques, i.e., variable name obfuscation, and design restructuring, that make the cases more sophisticated for the used LLMs. Our experimental evaluation demonstrates perfect detection rates by GPT-4o and Gemini 1.5 pro in baseline scenarios (100%/100% precision/recall), with both models achieving better trigger line coverage (TLC: 0.82-0.98) than payload line coverage (PLC: 0.32-0.46). Under code perturbation, while Gemini 1.5 pro maintains perfect detection performance (100%/100%), GPT-4o (100%/85.7%) and Llama 3.1 (66.7%/85.7%) show some degradation in detection rates, and all models experience decreased accuracy in localizing both triggers and payloads. This paper validates the potential of LLM approaches for hardware security applications, highlighting areas for future improvement.

Traditionally, inserting realistic Hardware Trojans (HTs) into complex hardware systems has been a time-consuming and manual process, requiring comprehensive knowledge of the design and navigating intricate Hardware Description Language (HDL) codebases. Machine Learning (ML)-based approaches have attempted to automate this process but often face challenges such as the need for extensive training data, long learning times, and limited generalizability across diverse hardware design landscapes. This paper addresses these challenges by proposing GHOST (Generator for Hardware-Oriented Stealthy Trojans), an automated attack framework that leverages Large Language Models (LLMs) for rapid HT generation and insertion. Our study evaluates three state-of-the-art LLMs - GPT-4, Gemini-1.5-pro, and Llama-3-70B - across three hardware designs: SRAM, AES, and UART. According to our evaluations, GPT-4 demonstrates superior performance, with 88.88% of HT insertion attempts successfully generating functional and synthesizable HTs. This study also highlights the security risks posed by LLM-generated HTs, showing that 100% of GHOST-generated synthesizable HTs evaded detection by an ML-based HT detection tool. These results underscore the urgent need for advanced detection and prevention mechanisms in hardware security to address the emerging threat of LLM-generated HTs. The GHOST HT benchmarks are available at: https://github.com/HSTRG1/GHOSTbenchmarks.git

Recent advances in techniques for monitoring and perturbing neural populations have greatly enhanced our ability to study circuits in the brain. In particular, two-photon holographic optogenetics now enables precise photostimulation of experimenter-specified groups of individual neurons, while simultaneous two-photon calcium imaging enables the measurement of ongoing and induced activity across the neural population. Despite the enormous space of potential photostimulation patterns and the time-consuming nature of photostimulation experiments, very little algorithmic work has been done to determine the most effective photostimulation patterns for identifying the neural population dynamics. Here, we develop methods to efficiently select which neurons to stimulate such that the resulting neural responses will best inform a dynamical model of the neural population activity. Using neural population responses to photostimulation in mouse motor cortex, we demonstrate the efficacy of a low-rank linear dynamical systems model, and develop an active learning procedure which takes advantage of low-rank structure to determine informative photostimulation patterns. We demonstrate our approach on both real and synthetic data, obtaining in some cases as much as a two-fold reduction in the amount of data required to reach a given predictive power. Our active stimulation design method is based on a novel active learning procedure for low-rank regression, which may be of independent interest., Comment: NeurIPS 2024

The growth of ionized hydrogen bubbles in the intergalactic medium around early luminous objects is a fundamental process during the Epoch of Reionization (EoR). In this study, we analyze bubble sizes and their evolution using the state-of-the-art THESAN radiation-hydrodynamics simulation suite, which self-consistently models radiation transport and realistic galaxy formation throughout a large (95.5 cMpc)^3 volume of the Universe. Analogous to the accretion and merger tree histories employed in galaxy formation simulations, we characterize the growth and merger rates of ionized bubbles by focusing on the spatially-resolved redshift of reionization. By tracing the chronological expansion of bubbles, we partition the simulation volume and construct a natural ionization history. We identify three distinct stages of ionized bubble growth: (1) initial slow expansion around the earliest ionizing sources seeding formation sites, (2) accelerated growth through percolation as bubbles begin to merge, and (3) rapid expansion dominated by the largest bubble. Notably, we find that the largest bubble emerges by z=9-10, well before the midpoint of reionization. This bubble becomes dominant during the second growth stage, and defines the third stage by rapidly expanding to eventually encompass the remainder of the simulation volume and becoming one of the few bubbles actively growing. Additionally, we observe a sharp decline in the number of bubbles with radii around ~10 cMpc compared to smaller sizes, indicating a characteristic scale in the final segmented bubble size distribution. Overall, these chronologically sequenced spatial reconstructions offer crucial insights into the physical mechanisms driving ionized bubble growth during the EoR and provide a framework for interpreting the structure and evolution of reionization itself., Comment: 17 pages, 23 figures. Comments welcome. Please visit www.thesan-project.com for more details

The realization of a quantum computer represents a tremendous scientific and technological challenge due to the extreme fragility of quantum information. The physical support of information, namely the quantum bit or qubit, must at the same time be strongly coupled to other qubits by gates to compute information, and well decoupled from its environment to keep its quantum behavior. An interesting physical system for realizing such qubits are magnetic impurities in semiconductors, such as bismuth donors in silicon. Indeed, spins associated to bismuth donors can reach an extremely long coherence time -- of the order of seconds. Yet it is extremely difficult to establish and control efficient gates between these spins. Here we demonstrate a protocol where single bismuth donors can coherently transfer their quantum information to a superconducting flux qubit, which acts as a mediator or quantum bus. This superconducting device allows to connect distant spins on-demand with little impact on their coherent behavior., Comment: 22 pages, 4 figures and 5 extended data figures

Recent advances in large language model (LLM) training have highlighted the need for diverse, high-quality instruction data. Recently, many works are exploring synthetic data generation using LLMs. However, they primarily focus on prompt engineering with standard supervised instruction-finetuned models, which contains a fundamental limitation: these models are optimized for general question-answering/problem-solving rather than data generation. We propose a paradigm shift named \textbf{NOMAD} by investigating how to specifically train models for data generation, demonstrating that this task differs significantly from training a classical LM. We identify two key factors: no-prompt-masked training and proper training set size selection. Our method, NOMAD, shows substantial improvements over baselines, achieving >4\% gains in TriviaQA and >2\% in GSM8K with limited training data. Finally, we offer new insights by interpreting synthetic data through the lenses of "relevance" and "novelty".

In order to mitigate the sample complexity of real-world reinforcement learning, common practice is to first train a policy in a simulator where samples are cheap, and then deploy this policy in the real world, with the hope that it generalizes effectively. Such \emph{direct sim2real} transfer is not guaranteed to succeed, however, and in cases where it fails, it is unclear how to best utilize the simulator. In this work, we show that in many regimes, while direct sim2real transfer may fail, we can utilize the simulator to learn a set of \emph{exploratory} policies which enable efficient exploration in the real world. In particular, in the setting of low-rank MDPs, we show that coupling these exploratory policies with simple, practical approaches -- least-squares regression oracles and naive randomized exploration -- yields a polynomial sample complexity in the real world, an exponential improvement over direct sim2real transfer, or learning without access to a simulator. To the best of our knowledge, this is the first evidence that simulation transfer yields a provable gain in reinforcement learning in settings where direct sim2real transfer fails. We validate our theoretical results on several realistic robotic simulators and a real-world robotic sim2real task, demonstrating that transferring exploratory policies can yield substantial gains in practice as well., Comment: NeurIPS 2024

This work focuses on advancing security research in the hardware design space by formally defining the realistic problem of Hardware Trojan (HT) detection. The goal is to model HT detection more closely to the real world, i.e., describing the problem as The Seeker's Dilemma where a detecting agent is unaware of whether circuits are infected by HTs or not. Using this theoretical problem formulation, we create a benchmark that consists of a mixture of HT-free and HT-infected restructured circuits while preserving their original functionalities. The restructured circuits are randomly infected by HTs, causing a situation where the defender is uncertain if a circuit is infected or not. We believe that our innovative benchmark and methodology of creating benchmarks will help the community judge the detection quality of different methods by comparing their success rates in circuit classification. We use our developed benchmark to evaluate three state-of-the-art HT detection tools to show baseline results for this approach. We use Principal Component Analysis to assess the strength of our benchmark, where we observe that some restructured HT-infected circuits are mapped closely to HT-free circuits, leading to significant label misclassification by detectors., Comment: arXiv admin note: substantial text overlap with arXiv:2402.17918

The Effective Field Theory of Large Scale Structure (EFTofLSS) has found tremendous success as a perturbative framework for the evolution of large scale structure, and it is now routinely used to compare theoretical predictions against cosmological observations. The model for the total matter field includes one nuisance parameter at 1-loop order, the effective sound speed, which can be extracted by matching the EFT to full N-body simulations. In this work we first leverage the Layzer-Irvine cosmic energy equation to show that the equation of state can be exactly computed with knowledge of the fully nonlinear power spectrum. When augmented with separate universe methods, we show one can estimate the effective sound speed. This estimate is in good agreement with simulation results, with errors at the few tens of percent level. We apply our method to investigate the cosmology dependence of the effective sound speed and to shed light on what cosmic structures shape its value., Comment: 17+10 pages, 9 figures. Comments are welcome!

Grading Objective Structured Clinical Examinations (OSCEs) is a time-consuming and expensive process, traditionally requiring extensive manual effort from human experts. In this study, we explore the potential of Large Language Models (LLMs) to assess skills related to medical student communication. We analyzed 2,027 video-recorded OSCE examinations from the University of Texas Southwestern Medical Center (UTSW), spanning four years (2019-2022), and several different medical cases or "stations." Specifically, our focus was on evaluating students' ability to summarize patients' medical history: we targeted the rubric item 'did the student summarize the patients' medical history?' from the communication skills rubric. After transcribing speech audio captured by OSCE videos using Whisper-v3, we studied the performance of various LLM-based approaches for grading students on this summarization task based on their examination transcripts. Using various frontier-level open-source and proprietary LLMs, we evaluated different techniques such as zero-shot chain-of-thought prompting, retrieval augmented generation, and multi-model ensemble methods. Our results show that frontier LLM models like GPT-4 achieved remarkable alignment with human graders, demonstrating a Cohen's kappa agreement of 0.88 and indicating strong potential for LLM-based OSCE grading to augment the current grading process. Open-source models also showed promising results, suggesting potential for widespread, cost-effective deployment. Further, we present a failure analysis identifying conditions where LLM grading may be less reliable in this context and recommend best practices for deploying LLMs in medical education settings.

Spin precession is a textbook example of dynamics of a quantum system that exactly mimics its classical counterpart. Here we challenge this view by certifying the quantumness of exotic states of a nuclear spin through its uniform precession. The key to this result is measuring the positivity, instead of the expectation value, of the $x$-projection of the precessing spin, and using a spin > 1/2 qudit, that is not restricted to semi-classical spin coherent states. The experiment is performed on a single spin-7/2 $^{123}$Sb nucleus, implanted in a silicon nanoelectronic device, amenable to high-fidelity preparation, control, and projective single-shot readout. Using Schr\"odinger cat states and other bespoke states of the nucleus, we violate the classical bound by 19 standard deviations, proving that no classical probability distribution can explain the statistic of this spin precession, and highlighting our ability to prepare quantum resource states with high fidelity in a single atomic-scale qudit., Comment: Main text: 11 pages, 5 figures. Supplementary information: 13 pages, 11 figures

Measurements of the galaxy 4-Point Correlation Function (4PCF) from theSloan Digital Sky Survey Baryon Oscillation Spectroscopic Survey (SDSS BOSS) have recently found strong statistical evidence for parity violation. If this signal is of genuine physical origin, it must stem from beyond-Standard Model physics, most likely during the very early Universe, prior to decoupling ($z$$\sim$$1,020$). Since the Baryon Acoustic Oscillation (BAO) features imprint at decoupling, they are expected in the parity-odd galaxy 4PCF, and so detecting them would be an additional piece of evidence that the signal is genuine. We demonstrate in a toy parity-violating model how the BAO imprint on the parity-odd 4PCF. We then outline how to perform a model-independent search for BAO in the odd 4PCF, desirable since, if the signal is real, we may not know for some time what model of e.g. inflation is producing it. If BAO are detected in the parity-odd sector, they can be used as a standard ruler as is already done in the 2PCF and 3PCF. We derive a simple formula relating the expected precision on the BAO scale to the overall parity-odd detection significance. Pursuing BAO in the odd 4PCF of future redshift surveys such as DESI, Euclid, Spherex, and Roman will be a valuable additional avenue to determine if parity violation in the distribution of galaxies is of genuine cosmological origin., Comment: 19 pages, 7 figures

A search for proton decay into $e^+/\mu^+$ and a $\eta$ meson has been performed using data from a 0.373 Mton$\cdot$year exposure (6050.3 live days) of Super-Kamiokande. Compared to previous searches this work introduces an improved model of the intranuclear $\eta$ interaction cross section, resulting in a factor of two reduction in uncertainties from this source and $\sim$10\% increase in signal efficiency. No significant data excess was found above the expected number of atmospheric neutrino background events resulting in no indication of proton decay into either mode. Lower limits on the proton partial lifetime of $1.4\times\mathrm{10^{34}~years}$ for $p\rightarrow e^+\eta$ and $7.3\times\mathrm{10^{33}~years}$ for $p\rightarrow \mu^+\eta$ at the 90$\%$ C.L. were set. These limits are around 1.5 times longer than our previous study and are the most stringent to date.

We give an algorithm to morph planar graph drawings that achieves small grid size at the expense of allowing a constant number of bends on each edge. The input is an $n$-vertex planar graph and two planar straight-line drawings of the graph on an $O(n) \times O(n)$ grid. The planarity-preserving morph is composed of $O(n)$ linear morphs between successive pairs of drawings, each on an $O(n) \times O(n)$ grid with a constant number of bends per edge. The algorithm to compute the morph runs in $O(n^2)$ time on a word RAM model with standard arithmetic operations -- in particular no square roots or cube roots are required. The first step of the algorithm is to morph each input drawing to a planar orthogonal box drawing where vertices are represented by boxes and each edge is drawn as a horizontal or vertical segment. The second step is to morph between planar orthogonal box drawings. This is done by extending known techniques for morphing planar orthogonal drawings with point vertices., Comment: To appear in the proceedings of the 32nd International Symposium on Graph Drawing and Network Visualization (GD 2024)

Forward Error Correction (FEC) provides reliable data flow in wireless networks despite the presence of noise and interference. However, its processing demands significant fraction of a wireless network's resources, due to its computationally-expensive decoding process. This forces network designers to compromise between performance and implementation complexity. In this paper, we investigate a novel processing architecture for FEC decoding, one based on the quantum approximate optimization algorithm (QAOA), to evaluate the potential of this emerging quantum compute approach in resolving the decoding performance-complexity tradeoff. We present FDeQ, a QAOA-based FEC Decoder design targeting the popular NextG wireless Low Density Parity Check (LDPC) and Polar codes. To accelerate QAOA-based decoding towards practical utility, FDeQ exploits temporal similarity among the FEC decoding tasks. This similarity is enabled by the fixed structure of a particular FEC code, which is independent of any time-varying wireless channel noise, ambient interference, and even the payload data. We evaluate FDeQ at a variety of system parameter settings in both ideal (noiseless) and noisy QAOA simulations, and show that FDeQ achieves successful decoding with error performance at par with state-of-the-art classical decoders at low FEC code block lengths. Furthermore, we present a holistic resource estimation analysis, projecting quantitative targets for future quantum devices in terms of the required qubit count and gate duration, for the application of FDeQ in practical wireless networks, highlighting scenarios where FDeQ may outperform state-of-the-art classical FEC decoders.

We present a field-level emulator for large-scale structure, capturing the cosmology dependence and the time evolution of cosmic structure formation. The emulator maps linear displacement fields to their corresponding nonlinear displacements from N-body simulations at specific redshifts. Designed as a neural network, the emulator incorporates style parameters that encode dependencies on $\Omega_{\rm m}$ and the linear growth factor $D(z)$ at redshift $z$. We train our model on the six-dimensional N-body phase space, predicting particle velocities as the time derivative of the model's displacement outputs. This innovation results in significant improvements in training efficiency and model accuracy. Tested on diverse cosmologies and redshifts not seen during training, the emulator achieves percent-level accuracy on scales of $k\sim~1~{\rm Mpc}^{-1}~h$ at $z=0$, with improved performance at higher redshifts. We compare predicted structure formation histories with N-body simulations via merger trees, finding consistent merger event sequences and statistical properties.

We investigate the problem of carving an $n$-face triangulated three-dimensional polytope using a tool to make cuts modelled by either a half-plane or sweeps from an infinite ray. In the case of half-planes cuts, we present a deterministic algorithm running in $O(n^2)$ time and a randomized algorithm running in $O(n^{3/2+\varepsilon})$ expected time for any $\varepsilon>0$. In the case of cuts defined by sweeps of infinite rays, we present an algorithm running in $O(n^5)$ time., Comment: Appeared in CCCG 2024. 10 pages, 7 figures

The neutron electric dipole moment (EDM) is a sensitive probe for currently undiscovered sources of charge-parity symmetry violation. As part of the TRIUMF Ultracold Advanced Neutron (TUCAN) collaboration, we are developing spin analyzers for ultracold neutrons (UCNs) to be used for a next-generation experiment to measure the neutron EDM with unprecedented precision. Spin-state analysis of UCNs constitutes an essential part of the neutron EDM measurement sequence. Magnetized iron films used as spin filters of UCNs are crucial experimental components, whose performance directly influences the statistical sensitivity of the measurement. To test such iron film spin filters, we propose the use of polarized cold-neutron reflectometry, in addition to conventional UCN transmission experiments. The new method provides information on iron film samples complementary to the UCN tests and accelerates the development cycles. We developed a collaborative effort to produce iron film spin filters and test them with cold and ultracold neutrons available at JRR-3/MINE2 and J-PARC/MLF BL05. In this article, we review the methods of neutron EDM measurements, discuss the complementarity of this new approach to test UCN spin filters, provide an overview of our related activities, and present the first results of polarized cold-neutron reflectometry recently conducted at the MINE2 beamline.

Reinforcement learning with human feedback (RLHF), as a widely adopted approach in current large language model pipelines, is \textit{bottlenecked by the size of human preference data}. While traditional methods rely on offline preference dataset constructions, recent approaches have shifted towards online settings, where a learner uses a small amount of labeled seed data and a large pool of unlabeled prompts to iteratively construct new preference data through self-generated responses and high-quality reward/preference feedback. However, most current online algorithms still focus on preference labeling during policy model updating with given feedback oracles, which incurs significant expert query costs. \textit{We are the first to explore cost-effective proxy reward oracles construction strategies for further labeling preferences or rewards with extremely limited labeled data and expert query budgets}. Our approach introduces two key innovations: (1) on-policy query to avoid OOD and imbalance issues in seed data, and (2) active learning to select the most informative data for preference queries. Using these methods, we train a evaluation model with minimal expert-labeled data, which then effectively labels nine times more preference pairs for further RLHF training. For instance, our model using Direct Preference Optimization (DPO) gains around over 1% average improvement on AlpacaEval2, MMLU-5shot and MMLU-0shot, with only 1.7K query cost. Our methodology is orthogonal to other direct expert query-based strategies and therefore might be integrated with them to further reduce query costs.

Reconfigurable antenna multiple-input multiple-output (MIMO) is a foundational technology for the continuing evolution of cellular systems, including upcoming 6G communication systems. In this paper, we address the problem of flexible/reconfigurable antenna configuration selection for point-to-point MIMO antenna systems by using physics-inspired heuristics. Firstly, we optimize the antenna configuration to maximize the signal-to-noise ratio (SNR) at the receiver by leveraging two basic heuristic solvers, i.e., coherent Ising machines (CIMs), that mimic quantum mechanical dynamics, and quantum annealing (QA), where a real-world QA architecture is considered (D-Wave). A mathematical framework that converts the configuration selection problem into CIM- and QA- compatible unconstrained quadratic formulations is investigated. Numerical and experimental results show that the proposed designs outperform classical counterparts and achieve near-optimal performance (similar to exhaustive search with exponential complexity) while ensuring polynomial complexity. Moreover, we study the optimal antenna configuration that maximizes the end-to-end Shannon capacity. A simulated annealing (SA) heuristic which achieves near-optimal performance through appropriate parameterization is adopted. A modified version of the basic SA that exploits parallel tempering to avoid local maxima is also studied, which provides additional performance gains. Extended numerical studies show that the SA solutions outperform conventional heuristics (which are also developed for comparison purposes), while the employment of the SNR-based solutions is highly sub-optimal., Comment: arXiv admin note: text overlap with arXiv:2403.12571

We introduce the Parity-Odd Power (POP) spectra, a novel set of observables for probing parity violation in cosmological $N$-point statistics. POP spectra are derived from composite fields obtained by applying nonlinear transformations, involving also gradients, curls, and filtering functions, to a scalar field. This compresses the parity-odd trispectrum into a power spectrum. These new statistics offer several advantages: they are computationally fast to construct, estimating their covariance is less demanding compared to estimating that of the full parity-odd trispectrum, and they are simple to model theoretically. We measure the POP spectra on simulations of a scalar field with a specific parity-odd trispectrum shape. We compare these measurements to semi-analytic theoretical calculations and find agreement. We also explore extensions and generalizations of these parity-odd observables., Comment: 18 pages, 5 figures

Whilst RISC-V has grown phenomenally quickly in embedded computing, it is yet to gain significant traction in High Performance Computing (HPC). However, as we move further into the exascale era, the flexibility offered by RISC-V has the potential to be very beneficial in future supercomputers especially as the community places an increased emphasis on decarbonising its workloads. Sophon's SG2042 is the first mass produced, commodity available, high-core count RISC-V CPU designed for high performance workloads. First released in summer 2023, and at the time of writing now becoming widely available, a key question is whether this is a realistic proposition for HPC applications. In this paper we use NASA's NAS Parallel Benchmark (NPB) suite to characterise performance of the SG2042 against other CPUs implementing the RISC-V, x86-64, and AArch64 ISAs. We find that the SG2042 consistently outperforms all other RISC-V solutions, delivering between a 2.6 and 16.7 performance improvement at the single core level. When compared against the x86-64 and AArch64 CPUs, which are commonplace for high performance workloads, we find that the SG2042 performs comparatively well with computationally bound algorithms but decreases in relative performance when the algorithms are memory bandwidth or latency bound. Based on this work, we identify that performance of the SG2042's memory subsystem is the greatest bottleneck., Comment: Preprint of paper submitted to RISC-V for HPC workshop at ISC

We present a novel multimodal preference dataset for creative tasks, consisting of over 250 million human ratings on more than 2.2 million captions, collected through crowdsourcing rating data for The New Yorker's weekly cartoon caption contest over the past eight years. This unique dataset supports the development and evaluation of multimodal large language models and preference-based fine-tuning algorithms for humorous caption generation. We propose novel benchmarks for judging the quality of model-generated captions, utilizing both GPT4 and human judgments to establish ranking-based evaluation strategies. Our experimental results highlight the limitations of current fine-tuning methods, such as RLHF and DPO, when applied to creative tasks. Furthermore, we demonstrate that even state-of-the-art models like GPT4 and Claude currently underperform top human contestants in generating humorous captions. As we conclude this extensive data collection effort, we release the entire preference dataset to the research community, fostering further advancements in AI humor generation and evaluation.

In this paper, we study the non-asymptotic sample complexity for the pure exploration problem in contextual bandits and tabular reinforcement learning (RL): identifying an epsilon-optimal policy from a set of policies with high probability. Existing work in bandits has shown that it is possible to identify the best policy by estimating only the difference between the behaviors of individual policies, which can be substantially cheaper than estimating the behavior of each policy directly. However, the best-known complexities in RL fail to take advantage of this and instead estimate the behavior of each policy directly. Does it suffice to estimate only the differences in the behaviors of policies in RL? We answer this question positively for contextual bandits but in the negative for tabular RL, showing a separation between contextual bandits and RL. However, inspired by this, we show that it almost suffices to estimate only the differences in RL: if we can estimate the behavior of a single reference policy, it suffices to only estimate how any other policy deviates from this reference policy. We develop an algorithm which instantiates this principle and obtains, to the best of our knowledge, the tightest known bound on the sample complexity of tabular RL., Comment: 59 pages, 2 Figures

Data selection has emerged as a core issue for large-scale visual-language model pretaining (e.g., CLIP), particularly with noisy web-curated datasets. Three main data selection approaches are: (1) leveraging external non-CLIP models to aid data selection, (2) training new CLIP-style embedding models that are more effective at selecting high-quality data than the original OpenAI CLIP model, and (3) designing better metrics or strategies universally applicable to any CLIP embedding without requiring specific model properties (e.g., CLIPScore is one popular metric). While the first two approaches have been extensively studied, the third remains under-explored. In this paper, we advance the third approach by proposing two new methods. Firstly, instead of classical CLIP scores that only consider the alignment between two modalities from a single sample, we introduce negCLIPLoss, a CLIP loss-inspired method that adds the alignment between one sample and its contrastive pairs as an extra normalization term for better quality measurement. Secondly, when downstream tasks are known, we propose a new norm-based metric, NormSim, to measure the similarity between pretraining data and target data. We test our methods on the data selection benchmark, DataComp~\cite{gadre2023datacomp}. Compared to the best baseline using only OpenAI's CLIP-L/14, our methods achieve a 5.3\% improvement on ImageNet-1k and a 2.8\% improvement on 38 downstream evaluation tasks. Moreover, both negCLIPLoss and NormSim are compatible with existing techniques. By combining our methods with the current best methods DFN~\cite{fang2023data} and HYPE~\cite{kim2024hype}, we can boost average performance on downstream tasks by 0.9\%, achieving a new state-of-the-art., Comment: This paper supercedes our previous VAS paper (arXiv:2402.02055)

High-dimensional quantum systems are a valuable resource for quantum information processing. They can be used to encode error-correctable logical qubits, for instance in continuous-variable states of oscillators such as microwave cavities or the motional modes of trapped ions. Powerful encodings include 'Schr\"odinger cat' states, superpositions of widely displaced coherent states, which also embody the challenge of quantum effects at the large scale. Alternatively, recent proposals suggest encoding logical qubits in high-spin atomic nuclei, which can host hardware-efficient versions of continuous-variable codes on a finite-dimensional system. Here we demonstrate the creation and manipulation of Schr\"odinger cat states using the spin-7/2 nucleus of a single antimony ($^{123}$Sb) atom, embedded and operated within a silicon nanoelectronic device. We use a coherent multi-frequency control scheme to produce spin rotations that preserve the SU(2) symmetry of the qudit, and constitute logical Pauli operations for logical qubits encoded in the Schr\"odinger cat states. The Wigner function of the cat states exhibits parity oscillations with a contrast up to 0.982(5), and state fidelities up to 0.913(2). These results demonstrate high-fidelity preparation of nonclassical resource states and logical control in a single atomic-scale object, opening up applications in quantum information processing and quantum error correction within a scalable, manufacturable semiconductor platform., Comment: 40 pages including main and supplementary information

The Hardware Trojan (HT) problem can be thought of as a continuous game between attackers and defenders, each striving to outsmart the other by leveraging any available means for an advantage. Machine Learning (ML) has recently played a key role in advancing HT research. Various novel techniques, such as Reinforcement Learning (RL) and Graph Neural Networks (GNNs), have shown HT insertion and detection capabilities. HT insertion with ML techniques, specifically, has seen a spike in research activity due to the shortcomings of conventional HT benchmarks and the inherent human design bias that occurs when we create them. This work continues this innovation by presenting a tool called TrojanForge, capable of generating HT adversarial examples that defeat HT detectors; demonstrating the capabilities of GAN-like adversarial tools for automatic HT insertion. We introduce an RL environment where the RL insertion agent interacts with HT detectors in an insertion-detection loop where the agent collects rewards based on its success in bypassing HT detectors. Our results show that this process helps inserted HTs evade various HT detectors, achieving high attack success percentages. This tool provides insight into why HT insertion fails in some instances and how we can leverage this knowledge in defense.

The Super-Kamiokande and T2K collaborations present a joint measurement of neutrino oscillation parameters from their atmospheric and beam neutrino data. It uses a common interaction model for events overlapping in neutrino energy and correlated detector systematic uncertainties between the two datasets, which are found to be compatible. Using 3244.4 days of atmospheric data and a beam exposure of $19.7(16.3) \times 10^{20}$ protons on target in (anti)neutrino mode, the analysis finds a 1.9$\sigma$ exclusion of CP-conservation (defined as $J_{CP}=0$) and a preference for the normal mass ordering., Comment: 12 pages, 4 figures

5G mmWave roadside networks promise high-speed wireless connectivity, but face significant challenges in maintaining reliable connections for users moving at high speed. Frequent handovers, complex beam alignment, and signal attenuation due to obstacles like car bodies lead to service interruptions and degraded performance. We present Wall-Street, a smart surface installed on vehicles to enhance 5G mmWave connectivity for users inside. Wall-Street improves mobility management by (1) steering outdoor mmWave signals into the vehicle, ensuring coverage for all users; (2) enabling simultaneous serving cell data transfer and candidate handover cell measurement, allowing seamless handovers without service interruption; and (3) combining beams from source and target cells during a handover to increase reliability. Through its flexible signal manipulation capabilities, Wall-Street provides uninterrupted high-speed connectivity for latency-sensitive applications in challenging mobile environments. We have implemented and integrated Wall-Street in the COSMOS testbed and evaluated its real-time performance with three gNBs and multiple mobile clients inside a surface-enabled vehicle, driving on a nearby road. In multi-UE scenarios, Wall-Street doubles the average TCP throughput and reduces delay by 30% over a baseline 5G Standalone handover protocol., Comment: 16 pages, 24 figures, under submission

Membraneless droplets formed through liquid-liquid phase separation (LLPS) play a crucial role in mRNA storage, enabling organisms to swiftly respond to environmental changes. However, the mechanisms underlying mRNA integration and protection within droplets remain unclear. Here, we unravel the role of bacterial aggresomes as stress granules (SGs) in safeguarding mRNA during stress. We discovered that upon stress onset, mobile mRNA molecules selectively incorporate into individual proteinaceous SGs based on length-dependent enthalpic gain over entropic loss. As stress prolongs, SGs undergo compaction facilitated by stronger non-specific RNA-protein interactions, thereby promoting recruitment of shorter RNA chains. Remarkably, mRNA ribonucleases are repelled from bacterial SGs, due to the influence of protein surface charge. This exclusion mechanism ensures the integrity and preservation of mRNA within SGs during stress conditions, explaining how mRNA can be stored and protected from degradation. Following stress removal, SGs facilitate mRNA translation, thereby enhancing cell fitness in changing environments. These droplets maintain mRNA physiological activity during storage, making them an intriguing new candidate for mRNA therapeutics manufacturing.

We present the design and characterization of the first full chain precision timing prototype ASIC, named ETL Readout Chip version 1 (ETROC1) for the CMS MTD endcap timing layer (ETL) upgrade. The ETL utilizes Low Gain Avalanche Diode (LGAD) sensors to detect charged particles, with the goal to achieve a time resolution of 40 - 50 ps per hit, and 30 - 40 ps per track with hits from two detector layers. The ETROC1 is composed of a 5 x 5 pixel array and peripheral circuits. The pixel array includes a 4 x 4 active pixel array with an H-tree shaped network delivering clock and charge injection signals. Each active pixel is composed of various components, including a bump pad, a charge injection circuit, a pre-amplifier, a discriminator, a digital-to-analog converter, and a time-to-digital converter. These components play essential roles as the front-end link in processing LGAD signals and measuring timing-related information. The peripheral circuits provide clock signals and readout functionalities. The size of the ETROC1 chip is 7 mm x 9 mm. ETROC1 has been fabricated in a 65 nm CMOS process, and extensively tested under stimuli of charge injection, infrared laser, and proton beam. The time resolution of bump-bonded ETROC1 + LGAD chipsets reaches 42 - 46 ps per hit in the beam test., Comment: 14 pages, 15 figures

Preceding a core-collapse supernova, various processes produce an increasing amount of neutrinos of all flavors characterized by mounting energies from the interior of massive stars. Among them, the electron antineutrinos are potentially detectable by terrestrial neutrino experiments such as KamLAND and Super-Kamiokande via inverse beta decay interactions. Once these pre-supernova neutrinos are observed, an early warning of the upcoming core-collapse supernova can be provided. In light of this, KamLAND and Super-Kamiokande, both located in the Kamioka mine in Japan, have been monitoring pre-supernova neutrinos since 2015 and 2021, respectively. Recently, we performed a joint study between KamLAND and Super-Kamiokande on pre-supernova neutrino detection. A pre-supernova alert system combining the KamLAND detector and the Super-Kamiokande detector was developed and put into operation, which can provide a supernova alert to the astrophysics community. Fully leveraging the complementary properties of these two detectors, the combined alert is expected to resolve a pre-supernova neutrino signal from a 15 M$_{\odot}$ star within 510 pc of the Earth, at a significance level corresponding to a false alarm rate of no more than 1 per century. For a Betelgeuse-like model with optimistic parameters, it can provide early warnings up to 12 hours in advance., Comment: Resubmitted to ApJ. 22 pages, 16 figures, for more information about the combined pre-supernova alert system, see https://www.lowbg.org/presnalarm/

Neutrinos from very nearby supernovae, such as Betelgeuse, are expected to generate more than ten million events over 10\,s in Super-Kamokande (SK). At such large event rates, the buffers of the SK analog-to-digital conversion board (QBEE) will overflow, causing random loss of data that is critical for understanding the dynamics of the supernova explosion mechanism. In order to solve this problem, two new DAQ modules were developed to aid in the observation of very nearby supernovae. The first of these, the SN module, is designed to save only the number of hit PMTs during a supernova burst and the second, the Veto module, prescales the high rate neutrino events to prevent the QBEE from overflowing based on information from the SN module. In the event of a very nearby supernova, these modules allow SK to reconstruct the time evolution of the neutrino event rate from beginning to end using both QBEE and SN module data. This paper presents the development and testing of these modules together with an analysis of supernova-like data generated with a flashing laser diode. We demonstrate that the Veto module successfully prevents DAQ overflows for Betelgeuse-like supernovae as well as the long-term stability of the new modules. During normal running the Veto module is found to issue DAQ vetos a few times per month resulting in a total dead time less than 1\,ms, and does not influence ordinary operations. Additionally, using simulation data we find that supernovae closer than 800~pc will trigger Veto module resulting in a prescaling of the observed neutrino data., Comment: 28 pages, 18 figures. Submitted to PTEP

The last couple of years have seen an ever-increasing interest in using different Ising solvers, like Quantum annealers, Coherent Ising machines, and Oscillator-based Ising machines, for solving tough computational problems in various domains. Although the simulations predict massive performance improvements for several tough computational problems, the real implementations of the Ising solvers tend to have limited precision, which can cause significant performance deterioration. This paper presents a novel methodology for mapping the problem on the Ising solvers to artificially increase the effective precision. We further evaluate our method for the Multiple-Input-Multiple-Output signal detection problem., Comment: version 1.0

Domain Specific Languages (DSLs) increase programmer productivity and provide high performance. Their targeted abstractions allow scientists to express problems at a high level, providing rich details that optimizing compilers can exploit to target current- and next-generation supercomputers. The convenience and performance of DSLs come with significant development and maintenance costs. The siloed design of DSL compilers and the resulting inability to benefit from shared infrastructure cause uncertainties around longevity and the adoption of DSLs at scale. By tailoring the broadly-adopted MLIR compiler framework to HPC, we bring the same synergies that the machine learning community already exploits across their DSLs (e.g. Tensorflow, PyTorch) to the finite-difference stencil HPC community. We introduce new HPC-specific abstractions for message passing targeting distributed stencil computations. We demonstrate the sharing of common components across three distinct HPC stencil-DSL compilers: Devito, PSyclone, and the Open Earth Compiler, showing that our framework generates high-performance executables based upon a shared compiler ecosystem.

Reconfigurable antenna multiple-input multiple-output (MIMO) is a promising technology for upcoming 6G communication systems. In this paper, we deal with the problem of configuration selection for reconfigurable antenna MIMO by leveraging Coherent Ising Machines (CIMs). By adopting the CIM as a heuristic solver for the Ising problem, the optimal antenna configuration that maximizes the received signal-to-noise ratio is investigated. A mathematical framework that converts the selection problem into a CIM-compatible unconstrained quadratic formulation is presented. Numerical studies show that the proposed CIM-based design outperforms classical counterparts and achieves near-optimal performance (similar to exponentially complex exhaustive searching) while ensuring polynomial complexity.

We present the results of the charge ratio ($R$) and polarization ($P^{\mu}_{0}$) measurements using the decay electron events collected from 2008 September to 2022 June by the Super-Kamiokande detector. Because of its underground location and long operation, we performed high precision measurements by accumulating cosmic-ray muons. We measured the muon charge ratio to be $R=1.32 \pm 0.02$ $(\mathrm{stat.}{+}\mathrm{syst.})$ at $E_{\mu}\cos \theta_{\mathrm{Zenith}}=0.7^{+0.3}_{-0.2}$ $\mathrm{TeV}$, where $E_{\mu}$ is the muon energy and $\theta_{\mathrm{Zenith}}$ is the zenith angle of incoming cosmic-ray muons. This result is consistent with the Honda flux model while this suggests a tension with the $\pi K$ model of $1.9\sigma$. We also measured the muon polarization at the production location to be $P^{\mu}_{0}=0.52 \pm 0.02$ $(\mathrm{stat.}{+}\mathrm{syst.})$ at the muon momentum of $0.9^{+0.6}_{-0.1}$ $\mathrm{TeV}/c$ at the surface of the mountain; this also suggests a tension with the Honda flux model of $1.5\sigma$. This is the most precise measurement ever to experimentally determine the cosmic-ray muon polarization near $1~\mathrm{TeV}/c$. These measurement results are useful to improve the atmospheric neutrino simulations., Comment: 29 pages, 45 figures

The first loading of gadolinium (Gd) into Super-Kamiokande in 2020 was successful, and the neutron capture efficiency on Gd reached 50\%. To further increase the Gd neutron capture efficiency to 75\%, 26.1 tons of $\rm Gd_2(\rm SO_4)_3\cdot \rm 8H_2O$ was additionally loaded into Super-Kamiokande (SK) from May 31 to July 4, 2022. As the amount of loaded $\rm Gd_2(\rm SO_4)_3\cdot \rm 8H_2O$ was doubled compared to the first loading, the capacity of the powder dissolving system was doubled. We also developed new batches of gadolinium sulfate with even further reduced radioactive impurities. In addition, a more efficient screening method was devised and implemented to evaluate these new batches of $\rm Gd_2(\rm SO_4)_3\cdot \rm 8H_2O$. Following the second loading, the Gd concentration in SK was measured to be $333.5\pm2.5$ ppm via an Atomic Absorption Spectrometer (AAS). From the mean neutron capture time constant of neutrons from an Am/Be calibration source, the Gd concentration was independently measured to be 332.7 $\pm$ 6.8(sys.) $\pm$ 1.1(stat.) ppm, consistent with the AAS result. Furthermore, during the loading the Gd concentration was monitored continually using the capture time constant of each spallation neutron produced by cosmic-ray muons,and the final neutron capture efficiency was shown to become 1.5 times higher than that of the first loaded phase, as expected., Comment: 34 pages, 13 figures, submitted to Nuclear Inst. and Methods in Physics Research, A

Among multi-messenger observations of the next galactic core-collapse supernova, Super-Kamiokande (SK) plays a critical role in detecting the emitted supernova neutrinos, determining the direction to the supernova (SN), and notifying the astronomical community of these observations in advance of the optical signal. On 2022, SK has increased the gadolinium dissolved in its water target (SK-Gd) and has achieved a Gd concentration of 0.033%, resulting in enhanced neutron detection capability, which in turn enables more accurate determination of the supernova direction. Accordingly, SK-Gd's real-time supernova monitoring system (Abe te al. 2016b) has been upgraded. SK_SN Notice, a warning system that works together with this monitoring system, was released on December 13, 2021, and is available through GCN Notices (Barthelmy et al. 2000). When the monitoring system detects an SN-like burst of events, SK_SN Notice will automatically distribute an alarm with the reconstructed direction to the supernova candidate within a few minutes. In this paper, we present a systematic study of SK-Gd's response to a simulated galactic SN. Assuming a supernova situated at 10 kpc, neutrino fluxes from six supernova models are used to characterize SK-Gd's pointing accuracy using the same tools as the online monitoring system. The pointing accuracy is found to vary from 3-7$^\circ$ depending on the models. However, if the supernova is closer than 10 kpc, SK_SN Notice can issue an alarm with three-degree accuracy, which will benefit follow-up observations by optical telescopes with large fields of view., Comment: 38 pages, 29 figures, 6 tables

Quantum Annealing (QA)-accelerated MIMO detection is an emerging research approach in the context of NextG wireless networks. The opportunity is to enable large MIMO systems and thus improve wireless performance. The approach aims to leverage QA to expedite the computation required for theoretically optimal but computationally-demanding Maximum Likelihood detection to overcome the limitations of the currently deployed linear detectors. This paper presents X-ResQ, a QA-based MIMO detector system featuring fine-grained quantum task parallelism that is uniquely enabled by the Reverse Annealing (RA) protocol. Unlike prior designs, X-ResQ has many desirable system properties for a parallel QA detector and has effectively improved detection performance as more qubits are assigned. In our evaluations on a state-of-the-art quantum annealer, fully parallel X-ResQ achieves near-optimal throughput (over 10 bits/s/Hz) for $4\times6$ MIMO with 16-QAM using six levels of parallelism with 240 qubits and $220~\mu$s QA compute time, achieving 2.5--5$\times$ gains compared against other tested detectors. For more comprehensive evaluations, we implement and evaluate X-ResQ in the non-quantum digital setting. This non-quantum X-ResQ demonstration showcases the potential to realize ultra-large $1024\times1024$ MIMO, significantly outperforming other MIMO detectors, including the state-of-the-art RA detector classically implemented in the same way., Comment: 22 pages