Search

Inhomogeneous quantum annealing (IQA), in which transverse fields are turned off one by one rather than simultaneously, has been proposed as an effective way to avoid the first-order phase transitions that impede conventional quantum annealing (QA). Here we explicitly study the dynamics of IQA, rather than merely the thermodynamics, and find that it is appreciably slower than the phase diagram would suggest. Interestingly, this slowdown manifests both when IQA succeeds in circumventing phase transitions and when it fails. Even in the absence of transitions, such as for the mean-field models that have been analyzed previously, IQA is slower than expected by a factor of the number of spins $N$. More significantly, we show that in non-mean-field models, first-order transitions are likely to be quite common, and the gap at such transitions is not merely exponential in $N$ but exactly zero. Thus IQA cannot reach the ground state on any timescale. Both of these results can be understood through the simple observation that a spin's magnetization becomes conserved once its field is turned off during the IQA protocol., Comment: Comments are welcome. 15 pages, 7 figures

Human label variation (HLV) challenges the standard assumption that an example has a single ground truth, instead embracing the natural variation in human labelling to train and evaluate models. While various training methods and metrics for HLV have been proposed, there has been no systematic meta-evaluation of HLV evaluation metrics, contributing to the lack of clarity in the best HLV training method. We propose new evaluation metrics and training methods and empirically meta-evaluate HLV evaluation metrics. We find that training on either disaggregated annotations or soft labels often performs best across metrics, and that our proposed soft metric correlates best with human preference.

Recent cognitive modeling studies have reported that larger language models (LMs) exhibit a poorer fit to human reading behavior, leading to claims of their cognitive implausibility. In this paper, we revisit this argument through the lens of mechanistic interpretability and argue that prior conclusions were skewed by an exclusive focus on the final layers of LMs. Our analysis reveals that next-word probabilities derived from internal layers of larger LMs align with human sentence processing data as well as, or better than, those from smaller LMs. This alignment holds consistently across behavioral (self-paced reading times, gaze durations, MAZE task processing times) and neurophysiological (N400 brain potentials) measures, challenging earlier mixed results and suggesting that the cognitive plausibility of larger LMs has been underestimated. Furthermore, we first identify an intriguing relationship between LM layers and human measures: earlier layers correspond more closely with fast gaze durations, while later layers better align with relatively slower signals such as N400 potentials and MAZE processing times. Our work opens new avenues for interdisciplinary research at the intersection of mechanistic interpretability and cognitive modeling., Comment: 19 pages

For a countable, complete, first-order theory $T$, we study $At$, the class of atomic models of $T$. We develop an analogue of $U$-rank and prove two results. On one hand, if some tp(d/a) is not ranked, then there are $2^{\aleph_1}$ non-isomorphic models in $At$ of size $\aleph_1$. On the other hand, if all types have finite rank, then the rank is fully additive and every finite tuple is dominated by an independent set of realizations of pseudo-minimal types.

The next generation of photonic circuits will require on-chip, ultrafast, compact, and energy-efficient components in a scalable platform to power technologies like quantum and neuromorphic computing. Here we present a significant advancement in ultrafast electrical control of highly nonlinear hybrid light-matter quasi-particles, called waveguide exciton-dipolaritons, with extremely low power consumption. Our device, that was shown to perform as an optical transistor with an optical switching power as low as $\sim$1 fJ/operation, achieves a GHz-rate electrical modulation with a record-low total energy consumption $\sim$3 fJ/bit and a compact active area of down to 25 $\mu$m$^2$. This work establishes waveguide-dipolariton platforms for scalable, electrically reconfigurable, ultra-low power photonic circuits for both classical and quantum computing and communication., Comment: 7 pages, 3 figures, 1 table, Supplementary materials

Leakage errors are unwanted transfer of population outside of a defined computational subspace and they occur in almost every platform for quantum computing. While prevalent, leakage is often overlooked when measuring and reporting the performance of quantum computers with standard randomized benchmarking methods. In fact, when leakage is substantial it can cause a large overestimation of fidelity. We provide several methods for measuring fidelity with randomized benchmarking in the presence of leakage errors that are applicable in different error regimes or under differnt control assumptions. We numerically demonstrate the methods for two-qubit randomized benchmarking, which often have the largest leakage contribution. Finally, we implement the methods on previously shared data from Quantinuum systems., Comment: 20 pages, 10 figures

In this work we study the circumstellar material (CSM) around massive stars, and the mass-loss rates depositing this CSM, using a large sample of radio observations of 325 core-collapse supernovae (CCSNe; only $\sim 22 \%$ of them being detected). This sample comprises both archival data and our new observations of 99 CCSNe conducted with the AMI-LA radio array in a systematic approach devised to constrain the mass-loss at different stages of stellar evolution. In the SN-CSM interaction model, observing the peak of the radio emission of a SN provides the CSM density at a given radius (and therefore mass-loss rate that deposited this CSM). On the other hand, limits on the radio emission, and/or on the peak of the radio emission provide a region in the CSM phase space that can be ruled out. Our analysis shows discrepancy between the values of mass-loss rates derived from radio-detected and radio-non-detected SNe. Furthermore, we rule out mass-loss rates in the range of $2 \times 10^{-6} - 10^{-4} \, \rm M_{\odot} \, yr^{-1}$ for different epochs during the last 1000 years before the explosion (assuming wind velocity of $10 \, \rm km \, s^{-1}$) for the progenitors of $\sim 80\%$ of the type II SNe in our sample. In addition, we rule out the ranges of mass-loss rates suggested for red supergiants for $\sim 50 \%$ of the progenitors of type II SNe in our sample. We emphasize here that these results take a step forward in constraining mass-loss in winds from a statistical point of view., Comment: 22 pages, 9 figures, accepted for publication in the Astrophysical Journal

A dynamic light scattering study of director-layer fluctuations in the antiferroelectric smectic-ZA phase of the ferroelectric nematic liquid crystal DIO is reported. The dynamics are consistent with the distinctive feature of the ZA phase that the smectic layers form parallel to the axis of molecular orientational order (director). A model is developed to describe quantitatively the dispersion of the fluctuation relaxation rates. The model is based on a specialization of the elastic free energy density of the smectic-C phase to the case of 90 degree director tilt, a "first-order" approximation of the viscous stresses by their form for an incompressible uniaxial fluid, and a treatment of the effect of chevron layer structure that develops in planar sample cells due to temperature-dependent layer shrinkage, as documented in previous studies on DIO. From the modeling, the layer compression elastic constant is estimated to be ~100 times lower in the smectic-ZA phase than in an ordinary smectic-A liquid crystal. Possible effects of the antiferroelectric layer polarization on the director splay elasticity and viscosity are described. The temperature dependencies of the splay, twist, and bend elastic constants and associated viscosities in the higher temperature nematic phase are also presented.

Increasingly, recommender systems are tasked with improving users' long-term satisfaction. In this context, we study a content exploration task, which we formalize as a bandit problem with delayed rewards. There is an apparent trade-off in choosing the learning signal: waiting for the full reward to become available might take several weeks, slowing the rate of learning, whereas using short-term proxy rewards reflects the actual long-term goal only imperfectly. First, we develop a predictive model of delayed rewards that incorporates all information obtained to date. Rewards as well as shorter-term surrogate outcomes are combined through a Bayesian filter to obtain a probabilistic belief. Second, we devise a bandit algorithm that quickly learns to identify content aligned with long-term success using this new predictive model. We prove a regret bound for our algorithm that depends on the \textit{Value of Progressive Feedback}, an information theoretic metric that captures the quality of short-term leading indicators that are observed prior to the long-term reward. We apply our approach to a podcast recommendation problem, where we seek to recommend shows that users engage with repeatedly over two months. We empirically validate that our approach significantly outperforms methods that optimize for short-term proxies or rely solely on delayed rewards, as demonstrated by an A/B test in a recommendation system that serves hundreds of millions of users.

HyperK\"ahler spaces, including manifolds, orbifolds and conical singularities play an important role in superstring/$M$-theory and gauge theories as well as in differential and algebraic geometry. In this paper we provide hundreds of new examples of compact hyperK\"ahler orbifolds of Kummer type $T^8/G$, where $T^8$ is the maximal torus of the compact Lie group $E_8$ and $G$ a finite group of isometries whose holonomies form a subgroup of the Weyl group of $E_8$. We show that, out of all of these examples, the only orbifolds whose singularities have a known holomorphic symplectic resolution lead to manifolds diffeomorphic to the two currently known examples of compact hyperK\"ahler 8-manifolds. We also demonstrate that these methods can, when combined with theorems of Joyce, be extended to construct potentially new manifolds of $\operatorname{SU}(4)$- and $\operatorname{Spin}(7)$- holonomy. All of these examples give rise to new vacua of string/$M$-theory in two/three dimensions.

We measure for the first time the differential photoproduction cross section $d\sigma/dt$ of the $a_2(1320)$ meson at an average photon beam energy of 8.5~GeV, using data with an integrated luminosity of 104~pb$^{-1}$ collected by the GlueX experiment. We fully reconstruct the $\gamma p \to \eta\pi^0 p$ reaction and perform a partial-wave analysis in the $a_2(1320)$ mass region with amplitudes that incorporate the linear polarization of the beam. This allows us to separate for the first time the contributions of natural- and unnatural-parity exchanges. These measurements provide novel information about the photoproduction mechanism, which is critical for the search for spin-exotic states., Comment: 8 pages, 4 figures plus supplemental material

This project explores the design and development of a transition edge sensor (TES) spectrometer for resonant soft X- ray scattering (RSXS) measurements developed in collaboration between Argonne National Laboratory (ANL) and the National Institute of Standards and Technology (NIST). Soft X-ray scattering is a powerful technique for studying the electronic and magnetic properties of materials on a microscopic level. However, the lack of high-performance soft X-ray spectrometers has limited the potential of this technique. TES spectrometers have the potential to overcome these limitations due to their high energy resolution, high efficiency, and broad energy range. This project aims to optimize the design of a TES spectrometer for RSXS measurements and more generally soft X-ray spectroscopy at the Advanced Photon Source (APS) 29-ID, leading to improved understanding of advanced materials. We will present a detailed description of the instrument design and implementation. The spectrometer consists of a large array of approximately 250 high-speed and high-resolution pixels. The pixels have saturation energies of approximately 1 keV, sub-ms pulse duration and energy resolution of approximately 1 eV. The array is read out using microwave multiplexing chips with MHz bandwidth per channel, enabling efficient data throughput. To facilitate measurement of samples in situ under ultra-high vacuum conditions at the beamline, the spectrometer is integrated with an approximately 1 m long snout.

There has been a great deal of interest in understanding which knots are characterized by which of their Dehn surgeries. We study a 4-dimensional version of this question: which knots are determined by which of their traces? We prove several results that are in stark contrast with what is known about characterizing surgeries, most notably that the 0-trace detects every L-space knot. Our proof combines tools in Heegaard Floer homology with results about surface homeomorphisms and their dynamics. We also consider nonzero traces, proving for instance that each positive torus knot is determined by its $n$-trace for any $n\leq 0$, whereas no non-positive integer is known to be a characterizing slope for any positive torus knot besides the right-handed trefoil., Comment: 35 pages, 1 figure

To address this gap, we introduce Libra-Leaderboard, a comprehensive framework designed to rank LLMs through a balanced evaluation of performance and safety. Combining a dynamic leaderboard with an interactive LLM arena, Libra-Leaderboard encourages the joint optimization of capability and safety. Unlike traditional approaches that average performance and safety metrics, Libra-Leaderboard uses a distance-to-optimal-score method to calculate the overall rankings. This approach incentivizes models to achieve a balance rather than excelling in one dimension at the expense of some other ones. In the first release, Libra-Leaderboard evaluates 26 mainstream LLMs from 14 leading organizations, identifying critical safety challenges even in state-of-the-art models.

Coupled many-body quantum systems give rise to rich emergent physics and abundance of both stationary and dynamical behaviours. Designing platforms with tunable and distinct forms of coupling gives new insight into the collective behaviour of the dimerised quantum systems. Fundamentally, two systems can exchange particles through either forbidden or allowed channels, underpinning evanescent and ballistic coupling mechanisms, respectively. Based on proximity, the former leads to large spectral splitting that defines, for instance, chemical binding energies, whereas the latter pushes for stringent phase-matching and synchronicity between oscillating degrees-of-freedom, analogous to phase-coupled harmonic oscillators, with a significantly smaller impact on the energy landscape. Here, we demonstrate an all-optically tunable evanescent-ballistic quantum fluid dimer based on hyperbolic exciton-polariton condensates in a photonic crystal waveguide. By changing the angle of the polariton dimer relative to the grating, the system transitions from an evanescently-coupled molecule with large mode-splitting to a ballistic condensate dimer with strict phase-matching conditions and rich interference patterns. We directly measure the condensate spectral features and mass flow subject to a saddle dispersion relation, and connect our results to mean field theory. These results showcase the potential of photonic crystals to study condensed matter phenomena lying at the interface between delay-coupled nonlinear oscillators and tight binding physics.

This paper introduces BiMediX2, a bilingual (Arabic-English) Bio-Medical EXpert Large Multimodal Model (LMM) with a unified architecture that integrates text and visual modalities, enabling advanced image understanding and medical applications. BiMediX2 leverages the Llama3.1 architecture and integrates text and visual capabilities to facilitate seamless interactions in both English and Arabic, supporting text-based inputs and multi-turn conversations involving medical images. The model is trained on an extensive bilingual healthcare dataset consisting of 1.6M samples of diverse medical interactions for both text and image modalities, mixed in Arabic and English. We also propose the first bilingual GPT-4o based medical LMM benchmark named BiMed-MBench. BiMediX2 is benchmarked on both text-based and image-based tasks, achieving state-of-the-art performance across several medical benchmarks. It outperforms recent state-of-the-art models in medical LLM evaluation benchmarks. Our model also sets a new benchmark in multimodal medical evaluations with over 9% improvement in English and over 20% in Arabic evaluations. Additionally, it surpasses GPT-4 by around 9% in UPHILL factual accuracy evaluations and excels in various medical Visual Question Answering, Report Generation, and Report Summarization tasks. The project page including source code and the trained model, is available at https://github.com/mbzuai-oryx/BiMediX2.

A skip graph is a resilient application-layer routing structure that supports range queries of distributed k-dimensional data. By sorting deterministic keys into groups based on locally computed random membership vectors, nodes in a standard skip graph can optimize range query performance in mobile networks such as unmanned aerial vehicle swarms. We propose a skip graph extension that inverts the key and membership vector roles and bases group membership on deterministic vectors derived from the z-ordering of k-dimensional data and sorting within groups is based on locally computed random keys., Comment: archival pre-print of this k-D distributed prefix-free distributed encoding technique ahead of another arXiv release which will cite this work, specifically with respect to using a more extensive prefix-free encoding technique to localize token partitions in an arbitrary input/output context of an LLM, SSM, or other k-D tokenized model

An extreme yet reconfigurable nonlinear response to a single photon by a photonic system is crucial for realizing a universal two-photon gate, an elementary building block for photonic quantum computing. Yet such a response, characterized by the photon blockade effect, has only been achieved in atomic systems or solid states ones that are difficult to scale up. Here we demonstrate electrically tunable partial photon blockade in dipolar waveguide polaritons on a semiconductor chip, measured via photon-correlations. Remarkably, these "dipolar photons" display a two-orders-of-magnitude stronger nonlinearity compared to unpolarized polaritons, with an extracted dipolar blockade radius up to more than 4 $\mu$m, significantly larger than the optical wavelength, and comparable to that of atomic Rydberg polaritons. Furthermore, we show that the dipolar interaction can be electrically switched and locally configured by simply tuning the gate voltage. Finally we show that with a simple modification of the design, a full photon blockade is expected, setting a new route towards scalable, reconfigurable, chip-integrated quantum photonic circuits with strong two-photon nonlinearities., Comment: 30 pages, 11 figures

Termination of the fractional quantum Hall states (FQHSs) and the emergence of Wigner crystal phases at very small Landau level filling factors ($\nu$) have been of continued interest for decades. Recently, in ultra-high-quality, dilute GaAs 2D electron systems (2DESs), strong evidence was reported for FQHSs at $\nu=1/7, 2/13$ and 2/11 which fall in the $\nu = p/(6p\pm1)$ Jain series of FQHSs, interpreted as integer ($p = 1$, 2) QHSs of 6-flux composite fermions ($^6$CFs). These states are surrounded by strongly-insulating phases which are generally believed to be Wigner crystals. Here, we study an ultra-high-quality 2DES confined to an AlAs quantum well where the 2D electrons have a much larger effective mass ($m^*\simeq 0.45 m_e$) and a smaller dielectric constant ($\epsilon\simeq10\epsilon_0$) compared to GaAs 2D electrons ($m^*\simeq 0.067 m_e$ and $\epsilon\simeq13\epsilon_0$). This combination of $m^*$ and $\epsilon$ renders the Landau level mixing parameter $\kappa$, defined as the ratio of the Coulomb and cyclotron energies, $\simeq 9$ times larger in AlAs 2DESs ($\kappa\propto m^*/\epsilon$). Qualitatively similar to the GaAs 2DESs, we observe an insulating behavior reentrant around a strong $\nu=1/5$ FQHS, and extending to $\nu<1/5$. Additionally, we observe a clear minimum in magnetoresistance at $\nu=2/11$, and an inflection point at $\nu=1/7$ which is very reminiscent of the first report of an emerging FQHS at $\nu=1/7$ in GaAs 2DESs. The data provide evidence for developing QHSs of $^6$CFs at very small fillings. This is very surprising because $\kappa$ near $\nu \simeq 1/6$ in our sample is very large ($\simeq4$), and larger $\kappa$ has the tendency to favor Wigner crystal states over FQHSs at small fillings. Our data should inspire calculations that accurately incorporate $\kappa$ in competing many-body phases of $^6$CFs at extremely small fillings near $\nu=1/6$., Comment: 7 pages, 4 figures, journal

High-fidelity mid-circuit measurements, which read out the state of specific qubits in a multiqubit processor without destroying them or disrupting their neighbors, are a critical component for useful quantum computing. They enable fault-tolerant quantum error correction, dynamic circuits, and other paths to solving classically intractable problems. But there are almost no methods to assess their performance comprehensively. We address this gap by introducing the first randomized benchmarking protocol that measures the rate at which mid-circuit measurements induce errors in many-qubit circuits. Using this protocol, we detect and eliminate previously undetected measurement-induced crosstalk in a 20-qubit trapped-ion quantum computer. Then, we use the same protocol to measure the rate of measurement-induced crosstalk error on a 27-qubit IBM Q processor, and quantify how much of that error is eliminated by dynamical decoupling., Comment: 7 pages, 4 figures, plus supplemental notes

In condensed-matter physics, supersolids refer to many-body quantum states breaking translational symmetry while exhibiting off-diagonal long-range order. This combination is debated for commensurate crystals, however it is accessible in lattice potentials fractionally filled by bosons with extended interactions. Here, we report such lattice supersolid with dipolar excitons confined in a sub-wavelength period potential. Excitons implement the Dicke-Hubbard Hamiltonian controlled by spatially extended dipolar and Dicke correlations. At half lattice-filling, these induce both a condensation in a single sub-radiant state and a dipolar quantum order spontaneously breaking the lattice symmetry. This combination signals a lattice supersolid dissipatively prepared across 8x8 sites. Our study underlines that nanoscopic exciton arrays open a route to explore new frontiers of quantum matter, e.g. many-body entanglement, at the interface with quantum nano-photonics., Comment: 12 pages, 7 figures

We present the localization and host galaxy of FRB 20190208A, a repeating source of fast radio bursts (FRBs) discovered using CHIME/FRB. As part of the PRECISE repeater localization program on the EVN, we monitored FRB 20190208A for 65.6 hours at $\sim1.4$ GHz and detected a single burst, which led to its VLBI localization with 260 mas uncertainty (2$\sigma$). Follow-up optical observations with the MMT Observatory ($i\gtrsim 25.7$ mag (AB)) found no visible host at the FRB position. Subsequent deeper observations with the GTC, however, revealed an extremely faint galaxy ($r=27.32 \pm0.16$ mag), very likely ($99.95 \%$) associated with FRB 20190208A. Given the dispersion measure of the FRB ($\sim580$ pc cm$^{-3}$), even the most conservative redshift estimate ($z_{\mathrm{max}}\sim0.83$) implies that this is the lowest-luminosity FRB host to date ($\lesssim10^8L_{\odot}$), even less luminous than the dwarf host of FRB 20121102A. We investigate how localization precision and the depth of optical imaging affect host association, and discuss the implications of such a low-luminosity dwarf galaxy. Unlike the other repeaters with low-luminosity hosts, FRB 20190208A has a modest Faraday rotation measure of a few tens of rad m$^{-2}$, and EVN plus VLA observations reveal no associated compact persistent radio source. We also monitored FRB 20190208A for 40.4 hours over 2 years as part of the \'ECLAT repeating FRB monitoring campaign on the Nan\c{c}ay Radio Telescope, and detected one burst. Our results demonstrate that, in some cases, the robust association of an FRB with a host galaxy will require both high localization precision, as well as deep optical follow-up., Comment: submitted to ApJL

The growing use of large language models (LLMs) has raised concerns regarding their safety. While many studies have focused on English, the safety of LLMs in Arabic, with its linguistic and cultural complexities, remains under-explored. Here, we aim to bridge this gap. In particular, we present an Arab-region-specific safety evaluation dataset consisting of 5,799 questions, including direct attacks, indirect attacks, and harmless requests with sensitive words, adapted to reflect the socio-cultural context of the Arab world. To uncover the impact of different stances in handling sensitive and controversial topics, we propose a dual-perspective evaluation framework. It assesses the LLM responses from both governmental and opposition viewpoints. Experiments over five leading Arabic-centric and multilingual LLMs reveal substantial disparities in their safety performance. This reinforces the need for culturally specific datasets to ensure the responsible deployment of LLMs., Comment: Accepted at NAACL 2025 Main Conference

The paper discusses fast frequency control in bulk power systems using embedded networks of grid-forming energy storage resources. Differing from their traditional roles of regulating reserves, the storage resources in this work operate as fast-acting grid assets shaping transient dynamics. The storage resources in the network are autonomously controlled using local measurements for distributed frequency support during disturbance events. Further, the grid-forming inverter systems interfacing with the storage resources, are augmented with fast-acting safety controls designed to contain frequency transients within a prescribed tolerance band. The control action, derived from the storage network, improves the frequency nadirs in the system and prevents the severity of a disturbance from propagating far from the source. The paper also presents sensitivity studies to evaluate the impacts of storage capacity and inverter controller parameters on the dynamic performance of frequency control and disturbance localization. The performance of the safety-constrained grid-forming control is also compared with the more common grid-following control. The results are illustrated through case studies on an IEEE test system., Comment: accepted at the IEEE PES Electrical Energy Storage Applications and Technologies Conference (EESAT)

Inverter-based storages are poised to play a prominent role in future power grids with massive renewable generation. Grid-forming inverters (GFMs) are emerging as a dominant technology with synchronous generators (SG)-like characteristics through primary control loops. Advanced secondary control schemes, e.g., consensus algorithms, allow GFM-interfaced storage units to participate in frequency regulations and restore nominal frequency following grid disturbances. However, it is imperative to ensure transient frequency excursions do not violate critical safety limits while the grid transitions from pre- to post-disturbance operating point. This paper presents a hierarchical safety-enforced consensus method -- combining a device-layer (decentralized) transient safety filter with a secondary-layer (distributed) consensus coordination -- to achieve three distinct objectives: limiting transient frequency excursions to safe limits, minimizing frequency deviations from nominal, and ensuring coordinated power sharing among GFM-storage units. The proposed hierarchical (two-layered) safety-consensus technique is illustrated using a GFM-interfaced storage network on an IEEE 68-bus system under multiple grid transient scenarios., Comment: accepted for presentation at the IEEE Electrical Energy Storage Applications and Technologies Conference (EESAT)

XML simplifies data exchange among heterogeneous computers, but it is notoriously verbose and has spawned the development of many XML-specific compressors and binary formats. We present an XML test corpus and a combined efficiency metric integrating compression ratio and execution speed. We use this corpus and linear regression to assess 14 general-purpose and XML-specific compressors relative to the proposed metric. We also identify key factors when selecting a compressor. Our results show XMill or WBXML may be useful in some instances, but a general-purpose compressor is often the best choice., Comment: 1. test data at https://web.archive.org/web/20160805043420/http://chris-augeri.com/wp-content/uploads/docs/xml_compress.htm 2. one next step is testing newer compressors, e.g., Brotli, along with Zstandard, which leverages the asymmetric numeral system (ANS) 3. citations at https://scholar.google.com/scholar?cluster=8178011010886368797&hl=en&as_sdt=5,33&sciodt=0,33

As large language models (LLMs) advance, their inability to autonomously execute tasks by directly interacting with external tools remains a critical limitation. Traditional methods rely on inputting tool descriptions as context, which is constrained by context length and requires separate, often inefficient, retrieval mechanisms. We introduce ToolGen, a paradigm shift that integrates tool knowledge directly into the LLM's parameters by representing each tool as a unique token. This enables the LLM to generate tool calls and arguments as part of its next token prediction capabilities, seamlessly blending tool invocation with language generation. Our framework allows the LLM to access and utilize a vast amount of tools with no additional retrieval step, significantly enhancing both performance and scalability. Experimental results with over 47,000 tools show that ToolGen not only achieves superior results in both tool retrieval and autonomous task completion but also sets the stage for a new era of AI agents that can adapt to tools across diverse domains. By fundamentally transforming tool retrieval into a generative process, ToolGen paves the way for more versatile, efficient, and autonomous AI systems. ToolGen enables end-to-end tool learning and opens opportunities for integration with other advanced techniques such as chain-of-thought and reinforcement learning, thereby expanding the practical capabilities of LLMs.

We present our analysis of supernovae serendipitously found to be radio-bright several years after their optical discovery. We used recent observations from the Australian SKA Pathfinder taken as part of the pilot Variables and Slow Transients and Rapid ASKAP Continuum Survey programs. We identified 29 objects by cross-matching sources from these ASKAP observations with known core-collapse supernovae below a declination of $+40^{\circ}$ and with a redshift of $z\leq0.15$. Our results focus on eight cases that show potential late-time radio emission. These supernovae exhibit significantly greater amounts of radio emission than expected from the standard model of a single shockwave propagating through a spherical circumstellar medium, with a constant density structure produced by regular stellar mass-loss. We also discuss how we can learn from future ASKAP surveys about the circumstellar environments and emission mechanisms of supernovae that undergo late-time radio re-brightening. This pilot work tested and confirmed the potential of the Variables and Slow Transients survey to discover and study late-time supernova emission., Comment: Accepted for publication in MNRAS

We introduce Loki, an open-source tool designed to address the growing problem of misinformation. Loki adopts a human-centered approach, striking a balance between the quality of fact-checking and the cost of human involvement. It decomposes the fact-checking task into a five-step pipeline: breaking down long texts into individual claims, assessing their check-worthiness, generating queries, retrieving evidence, and verifying the claims. Instead of fully automating the claim verification process, Loki provides essential information at each step to assist human judgment, especially for general users such as journalists and content moderators. Moreover, it has been optimized for latency, robustness, and cost efficiency at a commercially usable level. Loki is released under an MIT license and is available on GitHub. We also provide a video presenting the system and its capabilities.

We rigorously prove that in nearly arbitrary quantum spin chains with power-law-distributed random fields, namely such that the probability of a field exceeding $h$ scales as $h^{-\alpha}$, it is impossible for any operator evolving in the Heisenberg picture to spread with dynamical exponent less than $1/\alpha$. In particular, ballistic growth is impossible for $\alpha < 1$, diffusive growth is impossible for $\alpha < 1/2$, and any finite dynamical exponent becomes impossible for sufficiently small $\alpha$. This result thus establishes a wide family of models in which the disorder provably prevents conventional transport. We express the result as a tightening of Lieb-Robinson bounds due to random fields -- the proof modifies the standard derivation such that strong fields appear as effective weak interactions, and then makes use of analogous recent results for random-bond spin chains., Comment: Comments welcome!

Heterogeneously integrated hybrid photonic crystal cavities enable strong light-matter interactions with solid-state, optically addressable quantum memories. A key challenge to realizing high quality factor (Q) hybrid photonic crystals is the reduced index contrast on the substrate compared to suspended devices in air. This challenge is particularly acute for color centers in diamond because of diamond's high refractive index, which leads to increased scattering loss into the substrate. Here we develop a design methodology for hybrid photonic crystals utilizing a detailed understanding of substrate-mediated loss, which incorporates sensitivity to fabrication errors as a critical parameter. Using this methodology we design robust, high-Q, GaAs-on-diamond photonic crystal cavities, and by optimizing our fabrication procedure we experimentally realize cavities with Q approaching 30,000 at a resonance wavelength of 955 nm.

In the superionic phase of silver iodide, we observe a distorted tetragonal structure characterized by symmetry breaking in the cation distribution. This phase competes with the well known bcc phase, with a symmetric cation distribution at an energetic cost of only a few meV/atom. The small energy difference suggests that these competing structures may both be thermally accessible near the superionic transition temperature. We also find that the distribution of silver ions depends on the low-temperature parent polymorph, with memory persisting in the superionic phase on the nanosecond time scales accessible in our simulations. Furthermore, simulations on the order 100ns reveal that even at temperatures where the bcc phase is stable, significant fluctuations toward the tetragonal lattice structure remain. Our results are consistent with many "anomalous" experimental observations and offer a molecular mechanism for the "memory effect" in silver iodide.

Feeding the growing human population sustainably amidst climate change is one of the most important challenges in the 21st century. Current practices often lead to the overuse of agronomic inputs, such as synthetic fertilizers and water, resulting in environmental contamination and diminishing returns on crop productivity. The complexity of agricultural systems, involving plant-environment interactions and human management, presents significant scientific and technical challenges for developing sustainable practices. Addressing these challenges necessitates transdisciplinary research, involving intense collaboration among fields such as plant science, engineering, computer science, and social sciences. Here, we present five case studies from two research centers demonstrating successful transdisciplinary approaches toward more sustainable water and fertilizer use. These case studies span multiple scales. Starting from whole-plant signaling, we explore how reporter plants can transform our understanding of plant communication and enable efficient application of water and fertilizers. We then show how new fertilizer technologies could increase the availability of phosphorus in the soil. To accelerate advancements in breeding new cultivars, we discuss robotic technologies for high-throughput plant screening in different environments at a population scale. At the ecosystem scale, we investigate phosphorus recovery from aquatic systems and methods to minimize phosphorus leaching. Finally, as agricultural outputs affect all people, we show how to integrate stakeholder perspectives and needs into the research. With these case studies, we hope to encourage the scientific community to adopt transdisciplinary research and promote cross-training among biologists, engineers, and social scientists to drive discovery and innovation in advancing sustainable agricultural systems.

In non-Hermitian open quantum systems, such as polariton condensates, local tailoring of gains and losses opens up an interesting possibility to realize functional optical elements. Here, we demonstrate that deliberately introducing losses via a photonic defect, realized by reducing the quality factor of a DBR mirror locally within an ultrahigh-quality microcavity, may be utilized to create directed polariton currents towards the defect. We discuss the role of polariton-polariton interactions in the process and how to tailor the effective decay time of a polariton condensate via coupling to the defect. Our results highlight the far-reaching potential of non-Hermitian physics in polaritonics., Comment: 9 pages, 4 figures

The paper presents a theoretical study on small-signal stability and damping in bulk power systems with multiple grid-forming inverter-based storage resources. A detailed analysis is presented, characterizing the impacts of inverter droop gains and storage size on the slower eigenvalues, particularly those concerning inter-area oscillation modes. From these parametric sensitivity studies, a set of necessary conditions are derived that the design of droop gain must satisfy to enhance damping performance. The analytical findings are structured into propositions highlighting potential design considerations for improving system stability. The findings are illustrated via numerical studies on an IEEE 68-bus grid-forming storage network., Comment: accepted for presentation at The 63rd IEEE Conference on Decision and Control

Uncertainty quantification (UQ) is a perspective approach to detecting Large Language Model (LLM) hallucinations and low quality output. In this work, we address one of the challenges of UQ in generation tasks that arises from the conditional dependency between the generation steps of an LLM. We propose to learn this dependency from data. We train a regression model, which target variable is the gap between the conditional and the unconditional generation confidence. During LLM inference, we use this learned conditional dependency model to modulate the uncertainty of the current generation step based on the uncertainty of the previous step. Our experimental evaluation on nine datasets and three LLMs shows that the proposed method is highly effective for uncertainty quantification, achieving substantial improvements over rivaling approaches.

We apply logic-based machine learning techniques to facilitate cellular engineering and drive biological discovery, based on comprehensive databases of metabolic processes called genome-scale metabolic network models (GEMs). Predicted host behaviours are not always correctly described by GEMs. Learning the intricate genetic interactions within GEMs presents computational and empirical challenges. To address these, we describe a novel approach called Boolean Matrix Logic Programming (BMLP) by leveraging boolean matrices to evaluate large logic programs. We introduce a new system, $BMLP_{active}$, which efficiently explores the genomic hypothesis space by guiding informative experimentation through active learning. In contrast to sub-symbolic methods, $BMLP_{active}$ encodes a state-of-the-art GEM of a widely accepted bacterial host in an interpretable and logical representation using datalog logic programs. Notably, $BMLP_{active}$ can successfully learn the interaction between a gene pair with fewer training examples than random experimentation, overcoming the increase in experimental design space. $BMLP_{active}$ enables rapid optimisation of metabolic models and offers a realistic approach to a self-driving lab for microbial engineering., Comment: arXiv admin note: substantial text overlap with arXiv:2405.06724

This work presents the use of graph learning for the prediction of multi-step experimental outcomes for applications across experimental research, including material science, chemistry, and biology. The viability of geometric learning for regression tasks was benchmarked against a collection of linear models through a combination of simulated and real-world data training studies. First, a selection of five arbitrarily designed multi-step surrogate functions were developed to reflect various features commonly found within experimental processes. A graph transformer network outperformed all tested linear models in scenarios that featured hidden interactions between process steps and sequence dependent features, while retaining equivalent performance in sequence agnostic scenarios. Then, a similar comparison was applied to real-world literature data on algorithm guided colloidal atomic layer deposition. Using the complete reaction sequence as training data, the graph neural network outperformed all linear models in predicting the three spectral properties for most training set sizes. Further implementation of graph neural networks and geometric representation of scientific processes for the prediction of experiment outcomes could lead to algorithm driven navigation of higher dimension parameter spaces and efficient exploration of more dynamic systems.

LLM (large language model) practitioners commonly notice that outputs can vary for the same inputs, but we have been unable to find work that evaluates LLM stability as the main objective. In our study of 6 deterministically configured LLMs across 8 common tasks with 5 identical runs, we see accuracy variations up to 10\%. In addition, no LLM consistently delivers repeatable accuracy across all tasks. We also show examples of variation that are not normally distributed and compare configurations with zero-shot/few-shot prompting and fine-tuned examples. To better quantify what is going on, we introduce metrics focused on stability: TARr@N for the total agreement rate at N runs over raw output, and TARa@N for total agreement over parsed-out answers. We suggest that stability metrics be integrated into leader boards and research results going forward.

This paper explores the task of automatic prediction of text spans in a legal problem description that support a legal area label. We use a corpus of problem descriptions written by laypeople in English that is annotated by practising lawyers. Inherent subjectivity exists in our task because legal area categorisation is a complex task, and lawyers often have different views on a problem, especially in the face of legally-imprecise descriptions of issues. Experiments show that training on majority-voted spans outperforms training on disaggregated ones., Comment: Accepted at WASSA 2024