Search

We present a scalable, robust approach to creating quantum programs of arbitrary size and complexity. The approach is based on the true abstraction of the problem. The quantum program is expressed in terms of a high-level model together with constraints and objectives on the final program. Advanced synthesis algorithms transform the model into a low-level quantum program that meets the user's specification and is directed at a stipulated hardware. This separation of description from implementation is essential for scale. The technology adapts electronic design automation methods to quantum computing, finding feasible implementations in a virtually unlimited functional space. The results show clear superiority over the compilation and transpilation methods used today. We expect that this technological approach will take over and prevail as quantum software become more demanding, complex, and essential.

We investigate the dynamic spreading of a dense suspension of particles under power law repulsive potentials decaying as $1/r^k$. We find analytically that the particles spread in a self-similar form, where the radius grows with time as $t^{1/(k+2)}$. Our results closely align with experiments of paramagnetic colloidal particles under an external magnetic field interacting via a $1/r^3$ potential. We further substantiate the theory by molecular dynamic simulations of thousands of particles in one and two dimensions. The simulations reveal a rich diversity of behaviors contingent on the value of $k$. Specifically, for $k>D-2$ (where $D$ is the dimension), the density is centered in the middle, for $k = D-2$, density is uniform, and for $k

Multi-agent systems, where specialized agents collaborate to solve a shared task hold great potential, from increased modularity to simulating complex environments. However, they also have a major caveat -- a single agent can cause the entire system to fail. Consider a simple game where the knowledge to solve the task is distributed between agents, which share information in a communication channel. At each round, any of the agents can terminate the game and make the final prediction, even if they are uncertain about the outcome of their action. Detection of such rogue agents $\textit{before they act}$ may prevent the system's failure. In this work, we propose to $\textit{monitor}$ agents during action prediction and $\textit{intervene}$ when a future error is likely to occur. To test our approach, we introduce WhoDunitEnv, a multi-agent collaboration environment that allows modular control over task complexity and communication structure. Experiments on two variants of WhoDunitEnv and the GovSim environment for resource sustainability show that our approach leads to substantial performance gains up to 17.4% and 20%, respectively. Moreover, a thorough analysis shows that our monitors successfully identify critical points of agent confusion and our interventions effectively stop agent errors from propagating.

We use a two sorted equational theory of algebraic effects to model concurrent shared state with preemptive interleaving, recovering Brookes's seminal 1996 trace-based model precisely. The decomposition allows us to analyse Brookes's model algebraically in terms of separate but interacting components. The multiple sorts partition terms into layers. We use two sorts: a "hold" sort for layers that disallow interleaving of environment memory accesses, analogous to holding a global lock on the memory; and a "cede" sort for the opposite. The algebraic signature comprises of independent interlocking components: two new operators that switch between these sorts, delimiting the atomic layers, thought of as acquiring and releasing the global lock; non-deterministic choice; and state-accessing operators. The axioms similarly divide cleanly: the delimiters behave as a closure pair; all operators are strict, and distribute over non-empty non-deterministic choice; and non-deterministic global state obeys Plotkin and Power's presentation of global state. Our representation theorem expresses the free algebras over a two-sorted family of variables as sets of traces with suitable closure conditions. When the held sort has no variables, we recover Brookes's trace semantics., Comment: to be published in the conference proceedings of Foundations of Software Science and Computation Structures (FoSSaCS) 2025

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai., Comment: 26 pages, 6 figures

In this work we estimate the explosion and progenitor properties of six Type II supernovae (SNe) at 0.675 <= z <= 3.61 discovered by the JWST Advanced Deep Extragalactic Survey (JADES) transient survey by modeling their light curves. This high-redshift Type II SN sample allows us to compare low-redshift Type II SNe to their high-redshift counterparts. Two Type II SNe are found to have high explosion energies of 3e51 erg, while the other four Type II SNe are estimated to have typical explosion energies found in the local Universe [(0.5-2)e51 erg]. The fraction of Type II SNe with high explosion energies might be higher at high redshifts because of, e.g., lower metallicity, but it is still difficult to draw a firm conclusion because of the small sample size and potential observational biases. We found it difficult to constrain the progenitor masses for Type II SNe in our sample because of the sparse light-curve data. We found two Type II SN light curves can be better reproduced by introducing confined, dense circumstellar matter. Thus, the confined, dense circumstellar matter frequently observed in nearby Type II SNe is likely to exist in Type II SNe at high redshifts as well. Two Type II SNe are estimated to have high host galaxy extinctions, showing the ability of JWST to discover dust-obscured SNe at high redshifts. More high-redshift Type II SNe are required to investigate the differences in the properties of Type II SNe near and far, but here we show the first glimpse into the high-redshift population of Type II SNe., Comment: 12 pages, 12 figures, 2 tables, submitted to Publications of the Astronomical Society of Japan

Competition between species and genotypes is a dominant factor in a variety of ecological and evolutionary processes. Biological dynamics are typically highly stochastic, and therefore, analyzing a competitive system requires accounting for the random nature of birth and death processes (demographic stochasticity) as well as the variability of external conditions (environmental stochasticity). Recent studies have highlighted the importance of species life history, showing that differences in life history lead competing species to experience different levels of demographic stochasticity. Here, we propose a simple model of two-species competition with different life histories and derive analytical expressions for various properties (fixation probability, fixation time, absorption time, probability density) under a wide range of conditions, including migration, selection, and environmental stochasticity. These properties provide insights into the long-term outcomes of competition, such as species persistence, extinction risks, and the influence of environmental variability on the evenness of the community.

We introduce LTX-Video, a transformer-based latent diffusion model that adopts a holistic approach to video generation by seamlessly integrating the responsibilities of the Video-VAE and the denoising transformer. Unlike existing methods, which treat these components as independent, LTX-Video aims to optimize their interaction for improved efficiency and quality. At its core is a carefully designed Video-VAE that achieves a high compression ratio of 1:192, with spatiotemporal downscaling of 32 x 32 x 8 pixels per token, enabled by relocating the patchifying operation from the transformer's input to the VAE's input. Operating in this highly compressed latent space enables the transformer to efficiently perform full spatiotemporal self-attention, which is essential for generating high-resolution videos with temporal consistency. However, the high compression inherently limits the representation of fine details. To address this, our VAE decoder is tasked with both latent-to-pixel conversion and the final denoising step, producing the clean result directly in pixel space. This approach preserves the ability to generate fine details without incurring the runtime cost of a separate upsampling module. Our model supports diverse use cases, including text-to-video and image-to-video generation, with both capabilities trained simultaneously. It achieves faster-than-real-time generation, producing 5 seconds of 24 fps video at 768x512 resolution in just 2 seconds on an Nvidia H100 GPU, outperforming all existing models of similar scale. The source code and pre-trained models are publicly available, setting a new benchmark for accessible and scalable video generation.

Solving complex real-world tasks requires cycles of actions and observations. This is particularly true in science, where tasks require many cycles of analysis, tool use, and experimentation. Language agents are promising for automating intellectual tasks in science because they can interact with tools via natural language or code. Yet their flexibility creates conceptual and practical challenges for software implementations, since agents may comprise non-standard components such as internal reasoning, planning, tool usage, as well as the inherent stochasticity of temperature-sampled language models. Here, we introduce Aviary, an extensible gymnasium for language agents. We formalize agents as policies solving language-grounded partially observable Markov decision processes, which we term language decision processes. We then implement five environments, including three challenging scientific environments: (1) manipulating DNA constructs for molecular cloning, (2) answering research questions by accessing scientific literature, and (3) engineering protein stability. These environments were selected for their focus on multi-step reasoning and their relevance to contemporary biology research. Finally, with online training and scaling inference-time compute, we show that language agents backed by open-source, non-frontier LLMs can match and exceed both frontier LLM agents and human experts on multiple tasks at up to 100x lower inference cost.

Recent advancements in generative modeling now enable the creation of 4D content (moving 3D objects) controlled with text prompts. 4D generation has large potential in applications like virtual worlds, media, and gaming, but existing methods provide limited control over the appearance and geometry of generated content. In this work, we introduce a method for animating user-provided 3D objects by conditioning on textual prompts to guide 4D generation, enabling custom animations while maintaining the identity of the original object. We first convert a 3D mesh into a ``static" 4D Neural Radiance Field (NeRF) that preserves the visual attributes of the input object. Then, we animate the object using an Image-to-Video diffusion model driven by text. To improve motion realism, we introduce an incremental viewpoint selection protocol for sampling perspectives to promote lifelike movement and a masked Score Distillation Sampling (SDS) loss, which leverages attention maps to focus optimization on relevant regions. We evaluate our model in terms of temporal coherence, prompt adherence, and visual fidelity and find that our method outperforms baselines that are based on other approaches, achieving up to threefold improvements in identity preservation measured using LPIPS scores, and effectively balancing visual quality with dynamic content.

In this paper we identify the source of a singularity in the training loss of key denoising models, that causes the denoiser's predictions to collapse towards the mean of the source or target distributions. This degeneracy creates false basins of attraction, distorting the denoising trajectories and ultimately increasing the number of steps required to sample these models. We circumvent this artifact by leveraging the deterministic ODE-based samplers, offered by certain denoising diffusion and score-matching models, which establish a well-defined change-of-variables between the source and target distributions. Given this correspondence, we propose a new probability flow model, the Lines Matching Model (LMM), which matches globally straight lines interpolating the two distributions. We demonstrate that the flow fields produced by the LMM exhibit notable temporal consistency, resulting in trajectories with excellent straightness scores. Beyond its sampling efficiency, the LMM formulation allows us to enhance the fidelity of the generated samples by integrating domain-specific reconstruction and adversarial losses, and by optimizing its training for the sampling procedure used. Overall, the LMM achieves state-of-the-art FID scores with minimal NFEs on established benchmark datasets: 1.57/1.39 (NFE=1/2) on CIFAR-10, 1.47/1.17 on ImageNet 64x64, and 2.68/1.54 on AFHQ 64x64. Finally, we provide a theoretical analysis showing that the use of optimal transport to relate the two distributions suffers from a curse of dimensionality, where the pairing set size (mini-batch) must scale exponentially with the signal dimension.

The BrowserGym ecosystem addresses the growing need for efficient evaluation and benchmarking of web agents, particularly those leveraging automation and Large Language Models (LLMs) for web interaction tasks. Many existing benchmarks suffer from fragmentation and inconsistent evaluation methodologies, making it challenging to achieve reliable comparisons and reproducible results. BrowserGym aims to solve this by providing a unified, gym-like environment with well-defined observation and action spaces, facilitating standardized evaluation across diverse benchmarks. Combined with AgentLab, a complementary framework that aids in agent creation, testing, and analysis, BrowserGym offers flexibility for integrating new benchmarks while ensuring consistent evaluation and comprehensive experiment management. This standardized approach seeks to reduce the time and complexity of developing web agents, supporting more reliable comparisons and facilitating in-depth analysis of agent behaviors, and could result in more adaptable, capable agents, ultimately accelerating innovation in LLM-driven automation. As a supporting evidence, we conduct the first large-scale, multi-benchmark web agent experiment and compare the performance of 6 state-of-the-art LLMs across all benchmarks currently available in BrowserGym. Among other findings, our results highlight a large discrepancy between OpenAI and Anthropic's latests models, with Claude-3.5-Sonnet leading the way on almost all benchmarks, except on vision-related tasks where GPT-4o is superior. Despite these advancements, our results emphasize that building robust and efficient web agents remains a significant challenge, due to the inherent complexity of real-world web environments and the limitations of current models.

The generalized distance matrix of a graph is a matrix in which the $(i,j)$th entry is a function, $f$, of the distance between vertex $i$ and vertex $j$. Depending on the choice of $f$, this family of matrices includes both the adjacency matrix and the traditional distance matrix. We present a cospectral construction for the generalized distance matrix akin to Godsil-McKay Switching. We also investigate a special case of the generalized distance matrix: the exponential distance matrix, which is a matrix where every entry is a value $q$ raised to the power of the distance between the vertices. We give an upper bound on the values of $q$ needed to show a pair of graphs is cospectral for all values of $q$ corresponding to the diameter of the graphs. We also give cospectral constructions unique to value $q=1/2$.

We consider two-dimensional periodically driven systems of fermions with particle-hole symmetry. Such systems support non-trivial topological phases, including ones that cannot be realized in equilibrium. We show that a space-time defect in the driving Hamiltonian, dubbed a ``time vortex,'' can bind $\pi$ Majorana modes. A time vortex is a point in space around which the phase lag of the Hamiltonian changes by a multiple of $2\pi$. We demonstrate this behavior on a periodically driven version of Kitaev's honeycomb spin model, where $\mathbb{Z}_2$ fluxes and time vortices can realize any combination of $0$ and $\pi$ Majorana modes. We show that a time vortex can be created using Clifford gates, simplifying its realization in near-term quantum simulators., Comment: 11 pages, 6 figures

Dust from core-collapse supernovae (CCSNe), specifically Type IIP SNe, has been suggested to be a significant source of the dust observed in high-redshift galaxies. CCSNe eject large amounts of newly formed heavy elements, which can condense into dust grains in the cooling ejecta. However, infrared (IR) observations of typical CCSNe generally measure dust masses that are too small to account for the dust production needed at high redshifts. Type IIn SNe, classified by their dense circumstellar medium (CSM), are also known to exhibit strong IR emission from warm dust, but the dust origin and heating mechanism have generally remained unconstrained because of limited observational capabilities in the mid-IR. Here, we present a JWST/MIRI Medium Resolution Spectrograph (MRS) spectrum of the Type IIn SN 2005ip nearly 17 years post-explosion. The Type IIn SN 2005ip is one of the longest-lasting and most well-studied SNe observed to date. Combined with a Spitzer mid-IR spectrum of SN 2005ip obtained in 2008, this data set provides a rare 15-year baseline, allowing for a unique investigation of the evolution of dust. The JWST spectrum shows a new high-mass dust component ($\gtrsim0.08$ M$_{\odot}$) that is not present in the earlier Spitzer spectrum. Our analysis shows dust likely formed over the past 15 years in the cold, dense shell (CDS), between the forward and reverse shocks. There is also a smaller mass of carbonaceous dust ($\gtrsim0.005$ M$_{\odot}$) in the ejecta. These observations provide new insights into the role of SN dust production, particularly within the CDS, and its potential contribution to the rapid dust enrichment of the early Universe.

Racial segregation is a widespread social and physical phenomenon present in every city across the United States. Although prevalent nationwide, each city has a unique history of racial segregation, resulting in distinct "shapes" of segregation. We use persistent homology, a technique from applied algebraic topology, to investigate whether common patterns of racial segregation exist among U.S. cities. We explore two methods of constructing simplicial complexes that preserve geospatial data, applying them to White, Black, Asian, and Hispanic demographic data from the U.S. census for 112 U.S. cities. Using these methods, we cluster the cities based on their persistence to identify groups with similar segregation "shapes". Finally, we apply cluster analysis techniques to explore the characteristics of our clusters. This includes calculating the mean cluster statistics to gain insights into the demographics of each cluster and using the Adjusted Rand Index to compare our results with other clustering methods., Comment: 27 pages, 12 figures

JWST observations of the young Galactic supernova remnant Cassiopeia A revealed an unexpected structure seen as a green emission feature in colored composite MIRI F1130W and F1280W images - hence dubbed the Green Monster - that stretches across the central parts of the remnant in projection. Combining the kinematic information from NIRSpec and MIRI MRS with the multi-wavelength imaging from NIRCam and MIRI, we associate the Green Monster with circumstellar material that was lost during an asymmetric mass-loss phase. MIRI images are dominated by dust emission but its spectra show emission lines from Ne, H and Fe with low radial velocities indicative of a CSM nature. An X-ray analysis of this feature in a companion paper (Vink et al. 2024) supports its CSM nature and detects significant blue shifting, thereby placing the Green Monster on the near side, in front of the Cas A SN remnant. The most striking features of the Green Monster are dozens of almost perfectly circular 1" - 3" sized holes, most likely created by interaction between high-velocity SN ejecta material and the CSM. Further investigation is needed to understand whether these holes were formed by small 8000-10500 km/s N-rich ejecta knots that penetrated and advanced out ahead of the remnant's 5000 - 6000 km/s outer blastwave, or by narrow ejecta fingers that protrude into the forward-shocked CSM. The detection of the Green Monster provides further evidence of the highly asymmetric mass-loss that Cas A's progenitor star underwent prior to explosion., Comment: 28 pages, 12 figures, resubmitted to ApJL after minor revision, comments welcome

Autonomous systems for software engineering are now capable of fixing bugs and developing features. These systems are commonly evaluated on SWE-bench (Jimenez et al., 2024a), which assesses their ability to solve software issues from GitHub repositories. However, SWE-bench uses only Python repositories, with problem statements presented predominantly as text and lacking visual elements such as images. This limited coverage motivates our inquiry into how existing systems might perform on unrepresented software engineering domains (e.g., front-end, game development, DevOps), which use different programming languages and paradigms. Therefore, we propose SWE-bench Multimodal (SWE-bench M), to evaluate systems on their ability to fix bugs in visual, user-facing JavaScript software. SWE-bench M features 617 task instances collected from 17 JavaScript libraries used for web interface design, diagramming, data visualization, syntax highlighting, and interactive mapping. Each SWE-bench M task instance contains at least one image in its problem statement or unit tests. Our analysis finds that top-performing SWE-bench systems struggle with SWE-bench M, revealing limitations in visual problem-solving and cross-language generalization. Lastly, we show that SWE-agent's flexible language-agnostic features enable it to substantially outperform alternatives on SWE-bench M, resolving 12% of task instances compared to 6% for the next best system.

Information retrieval (IR) methods, like retrieval augmented generation, are fundamental to modern applications but often lack statistical guarantees. Conformal prediction addresses this by retrieving sets guaranteed to include relevant information, yet existing approaches produce large-sized sets, incurring high computational costs and slow response times. In this work, we introduce a score refinement method that applies a simple monotone transformation to retrieval scores, leading to significantly smaller conformal sets while maintaining their statistical guarantees. Experiments on various BEIR benchmarks validate the effectiveness of our approach in producing compact sets containing relevant information., Comment: 6 pages

JWST has now revealed a population of broad-line AGN at $z>4$ characterized by a distinctive SED shape, with very red rest-frame optical and very blue rest-frame UV continuum. While the optical continuum is thought to originate from the accretion disk, the origin of the UV continuum has been largely unclear. We report the detection of the strong rest-frame UV emission lines of CIII]$\lambda\lambda$1907,1909 and CIV$\lambda\lambda$1549,1551 in a "little red dot" AGN, COS-66964. Spectroscopically confirmed at $z=7.0371$, COS-66964 exhibits broad H$\alpha$ emission (FWHM $\sim 2000$ km s$^{-1}$), and weak broad H$\beta$, implying significant dust attenuation to the BLR ($A_V = 3.9^{+1.7}_{-0.9}$). The H$\alpha$ line width implies a central SMBH mass of $M_{\rm BH} = \left(1.9^{+1.6}_{-0.7}\right)\times10^{7}$ M$_\odot$, and an Eddington ratio $\lambda\sim0.3$-$0.5$. While marginal HeII$\lambda4687$ and [FeX]$\lambda6376$ detections further indicate that the AGN dominates in the rest-frame optical, the non-detection of HeII$\lambda1640$ in the UV despite high EW CIII] and CIV ($\sim 35$ {\AA}) is more consistent with photoionization by massive stars. The non-detection of MgII$\lambda\lambda$2800 is similarly inconsistent with an AGN scattered light interpretation. Assuming the rest-frame UV is dominated by stellar light, we derive a stellar mass of $\log M_\star/M_\odot\sim8.5$, implying an elevated $M_{\rm BH}/M_\star$ ratio $\sim2$ orders of magnitude above the local relation, but consistent with other high-$z$ AGN discovered by JWST. The source is unresolved in all bands, implying a very compact size $\lesssim200$ pc in the UV. This suggests that the simultaneous buildup of compact stellar populations (i.e., galaxy bulges) and the central SMBH is ongoing even at $z>7$., Comment: 14 pages, 6 figures; submitted to ApJL

We conduct an incentivized lab experiment to test participants' ability to understand the DA matching mechanism and the strategyproofness property, conveyed in different ways. We find that while many participants can (using a novel GUI) learn DA's mechanics and calculate its outcomes, such understanding does not imply understanding of strategyproofness (as measured by specially designed tests). However, a novel menu description of strategyproofness conveys this property significantly better than other treatments. While behavioral effects are small on average, participants with levels of strategyproofness understanding above a certain threshold play the classical dominant strategy at very high rates.

Renormalization of physical quantities for quantum field theories in curved spacetimes can be achieved via the consistent subtraction of counterterms within a regularization scheme such as a point-splitting method. Pragmatic mode-sum regularization (PMR) is a point-splitting method which is particularly suitable for rotating black hole spacetimes. We extend and tailor the t-splitting variant of PMR specifically for the interior of a Kerr black hole on the axis of rotation, focusing on a minimally-coupled massless scalar field in the physically-motivated Unruh state. The method addresses unique challenges in the black hole interior that do not occur outside. In particular, while the infinite sum over multipolar number l converges in the black hole exterior, it diverges in the interior, necessitating the subtraction of a so-called intermediate divergence which includes introducing an additional "small" split in the direction of the polar angle. This procedure is outlined and justified, along with the standard PMR method's counterterms subtraction. We apply this method to calculate the renormalized energy-momentum fluxes $\langle T_{uu}\rangle^U_\text{ren}$, $\langle T_{vv}\rangle^U_\text{ren}$ (where u and v are the standard Eddington coordinates) and the renormalized field square $\langle \Phi^2\rangle^U_\text{ren}$ throughout the black hole interior, spanning from (just off) the event horizon to (just off) the inner horizon. Special emphasis is placed on the inner horizon vicinity, where our t-splitting results for the fluxes asymptote to those obtained directly at the inner horizon using a different method in a previous work. In an Appendix, we develop an alternative t-splitting PMR variant which does not include the intermediate divergence subtraction. We utilize it for independent computations that are used to verify the standard t-splitting variant presented in the main text., Comment: 57 pages, 19 figures. Including a supplemental Mathematica notebook

Training large language models (LLMs) for external tool usage is a rapidly expanding field, with recent research focusing on generating synthetic data to address the shortage of available data. However, the absence of systematic data quality checks poses complications for properly training and testing models. To that end, we propose two approaches for assessing the reliability of data for training LLMs to use external tools. The first approach uses intuitive, human-defined correctness criteria. The second approach uses a model-driven assessment with in-context evaluation. We conduct a thorough evaluation of data quality on two popular benchmarks, followed by an extrinsic evaluation that showcases the impact of data quality on model performance. Our results demonstrate that models trained on high-quality data outperform those trained on unvalidated data, even when trained with a smaller quantity of data. These findings empirically support the significance of assessing and ensuring the reliability of training data for tool-using LLMs.

We study open-ended evolution by focusing on computational and information-processing dynamics underlying major evolutionary transitions. In doing so, we consider biological organisms as hierarchical dynamical systems that generate regularities in their phase-spaces through interactions with their environment. These emergent information patterns can then be encoded within the organism's components, leading to self-modelling "tangled hierarchies". Our main conjecture is that when macro-scale patterns are encoded within micro-scale components, it creates fundamental tensions (computational inconsistencies) between what is encodable at a particular evolutionary stage and what is potentially realisable in the environment. A resolution of these tensions triggers an evolutionary transition which expands the problem-space, at the cost of generating new tensions in the expanded space, in a continual process. We argue that biological complexification can be interpreted computation-theoretically, within the G\"odel--Turing--Post recursion-theoretic framework, as open-ended generation of computational novelty. In general, this process can be viewed as a meta-simulation performed by higher-order systems that successively simulate the computation carried out by lower-order systems. This computation-theoretic argument provides a basis for hypothesising the biological arrow of time., Comment: 30 pages, 13 figures

Russian Internet Trolls use fake personas to spread disinformation through multiple social media streams. Given the increased frequency of this threat across social media platforms, understanding those operations is paramount in combating their influence. Using Twitter content identified as part of the Russian influence network, we created a predictive model to map the network operations. We classify accounts type based on their authenticity function for a sub-sample of accounts by introducing logical categories and training a predictive model to identify similar behavior patterns across the network. Our model attains 88% prediction accuracy for the test set. Validation is done by comparing the similarities with the 3 million Russian troll tweets dataset. The result indicates a 90.7% similarity between the two datasets. Furthermore, we compare our model predictions on a Russian tweets dataset, and the results state that there is 90.5% correspondence between the predictions and the actual categories. The prediction and validation results suggest that our predictive model can assist with mapping the actors in such networks., Comment: 17 pages, 08 figures, and 04 tables. Further, the paper is published in https://www.lhscientificpublishing.com/Journals/articles/DOI-10.5890-JVTSD.2023.06.001.aspx

In 2016, a network of social media accounts animated by Russian operatives attempted to divert political discourse within the American public around the presidential elections. This was a coordinated effort, part of a Russian-led complex information operation. Utilizing the anonymity and outreach of social media platforms Russian operatives created an online astroturf that is in direct contact with regular Americans, promoting Russian agenda and goals. The elusiveness of this type of adversarial approach rendered security agencies helpless, stressing the unique challenges this type of intervention presents. Building on existing scholarship on the functions within influence networks on social media, we suggest a new approach to map those types of operations. We argue that pretending to be legitimate social actors obliges the network to adhere to social expectations, leaving a social footprint. To test the robustness of this social footprint we train artificial intelligence to identify it and create a predictive model. We use Twitter data identified as part of the Russian influence network for training the artificial intelligence and to test the prediction. Our model attains 88% prediction accuracy for the test set. Testing our prediction on two additional models results in 90.7% and 90.5% accuracy, validating our model. The predictive and validation results suggest that building a machine learning model around social functions within the Russian influence network can be used to map its actors and functions., Comment: 28 pages, 03 figures, 06 tables. Further, this paper has been published in Social Network Analysis and Mining

We extend the concept of loss landscape mode connectivity to the input space of deep neural networks. Mode connectivity was originally studied within parameter space, where it describes the existence of low-loss paths between different solutions (loss minimizers) obtained through gradient descent. We present theoretical and empirical evidence of its presence in the input space of deep networks, thereby highlighting the broader nature of the phenomenon. We observe that different input images with similar predictions are generally connected, and for trained models, the path tends to be simple, with only a small deviation from being a linear path. Our methodology utilizes real, interpolated, and synthetic inputs created using the input optimization technique for feature visualization. We conjecture that input space mode connectivity in high-dimensional spaces is a geometric effect that takes place even in untrained models and can be explained through percolation theory. We exploit mode connectivity to obtain new insights about adversarial examples and demonstrate its potential for adversarial detection. Additionally, we discuss applications for the interpretability of deep networks.

Multi-document summarization (MDS) is a challenging task, often decomposed to subtasks of salience and redundancy detection, followed by text generation. In this context, alignment of corresponding sentences between a reference summary and its source documents has been leveraged to generate training data for some of the component tasks. Yet, this enabling alignment step has usually been applied heuristically on the sentence level on a limited number of subtasks. In this paper, we propose extending the summary-source alignment framework by (1) applying it at the more fine-grained proposition span level, (2) annotating alignment manually in a multi-document setup, and (3) revealing the great potential of summary-source alignments to yield several datasets for at least six different tasks. Specifically, for each of the tasks, we release a manually annotated test set that was derived automatically from the alignment annotation. We also release development and train sets in the same way, but from automatically derived alignments. Using the datasets, each task is demonstrated with baseline models and corresponding evaluation metrics to spur future research on this broad challenge., Comment: Accepted to ACL-Findings 2024

Programs written in C/C++ often include inline assembly: a snippet of architecture-specific assembly code used to access low-level functionalities that are impossible or expensive to simulate in the source language. Although inline assembly is widely used, its semantics has not yet been formally studied. In this paper, we overcome this deficiency by investigating the effect of inline assembly on the consistency semantics of C/C++ programs. We propose the first memory model of the C++ Programming Language with support for inline assembly for Intel's x86 including non-temporal stores and store fences. We argue that previous provably correct compiler optimizations and correct compiler mappings should remain correct under such an extended model and we prove that this requirement is met by our proposed model.

Analytic continuations of holographic entanglement entropy in which the boundary subregion extends along a timelike direction have brought a promise of a novel, time-centric probe of the emergence of spacetime. We propose that the bulk carriers of this holographic timelike entanglement entropy are boundary-anchored extremal surfaces probing analytic continuation of holographic spacetimes into complex coordinates. This proposal not only provides a geometric interpretation of all the known cases obtained by direct analytic continuation of closed-form expressions of holographic entanglement entropy of a strip subregion but crucially also opens a window to study holographic timelike entanglement entropy in full generality. We initialize the investigation of complex extremal surfaces anchored on a timelike strip at the boundary of anti-de Sitter black branes. We find multiple complex extremal surfaces and discuss possible principles singling out the physical contribution., Comment: 8 pages, 6 figures

We present JWST/MIRI observations of the Type I superluminous supernova (SLSN) 2017gci taken over 2000 rest-frame days after the supernova (SN) exploded, which represent the latest phase images taken of any known SLSN. We find that archival \WISE detections of SN\,2017gci taken 70 to 200 days after explosion are most likely explained by an IR dust echo from a $\sim 3 \times 10^{-4}$ M$_\odot$ shell of pre-existing dust, as opposed to freshly-formed dust. New JWST observations reveal IR emission in the field of SN\,2017gci, which we determine is most likely dominated by the host galaxy of the SN, based on the expected flux of the galaxy and the measurable separation between said emission and the location of the SN. Based on models for IR emission of carbonate dust, we place a $3\sigma$ upper limit of $0.83$ M$_\odot$ of dust formed in SN\,2017gci, with a lowest $1\sigma$ limit of $0.44$ M$_\odot$. Infrared (IR) detections of other SLSNe have suggested that SLSNe could be among the most efficient dust producers in the universe. Our results suggest that SLSNe do not necessarily form more dust than other types of SNe, but instead might have a more accelerated dust formation process. More IR observations of a larger sample of SLSNe will be required to determine how efficient dust production is in SLSNe., Comment: 11 pages, 5 figures, submitted to ApJ

Reasoning about hyperproperties of concurrent implementations, such as the guarantees these implementations provide to randomized client programs, has been a long-standing challenge. Standard linearizability enables the use of atomic specifications for reasoning about standard properties, but not about hyperproperties. A stronger correctness criterion, called strong linearizability, enables such reasoning, but is rarely achievable, leaving various useful implementations with no means for reasoning about their hyperproperties. In this paper, we focus on registers and devise non-atomic specifications that capture a wide-range of well-studied register implementations and enable reasoning about their hyperproperties. First, we consider the class of write strong-linearizable implementations, a recently proposed useful weakening of strong linearizability, which allows more intricate implementations, such as the well-studied single-writer ABD distributed implementation. We introduce a simple shared-memory register specification that can be used for reasoning about hyperproperties of programs that use write strongly-linearizable implementations. Second, we introduce a new linearizability class, which we call decisive linearizability, that is weaker than write strong-linearizability and includes multi-writer ABD, and develop a second shared-memory register specification for reasoning about hyperproperties of programs that use register implementations of this class. These results shed light on the hyperproperties guaranteed when simulating shared memory in a crash-resilient message-passing system.

The James Webb Space Telescope (JWST) has opened up a new window to study highly reddened explosive transients. We present results from late-time (1421 days post-explosion) JWST follow-up spectroscopic observations with NIRSpec and MIRI LRS of the intermediate luminosity red transient (ILRT) AT 2019abn located in the nearby Messier 51 galaxy (8.6 Mpc). ILRTs represent a mysterious class of transients which exhibit peak luminosities between those of classical novae and supernovae and which are known to be highly dust obscured. Similar to the prototypical examples of this class of objects, NGC 300 2008-OT and SN 2008S, AT 2019abn has an extremely red and dusty progenitor detected only in pre-explosion Spitzer/IRAC imaging at 3.6 and 4.5 micron and not in deep optical or near-infrared HST images. We find that late time observations of AT 2019abn from NEOWISE and JWST are consistent with the late time evolution of SN 2008S. In part because they are so obscured by dust, it is unknown what produces an ILRT with hypotheses ranging from high mass stellar merger events, non-terminal stellar outbursts, or terminal supernovae explosions through electron-capture in super-AGB stars. Our JWST observations show strong mid-IR Class C PAH features at 6.3 and 8.25 micron typical of carbon-rich post-AGB sources. These features suggest the dust around AT 2019abn, either pre-existing or newly formed in the ejecta, is composed of carbonaceous grains which are not typically observed around red supergiants. However, depending on the strength and temperature of hot bottom burning, SAGBs may be expected to exhibit a carbon-rich chemistry. Thus our JWST observations are consistent with AT 2019abn having an SAGB progenitor., Comment: 12 pages, 4 figures, submitted to ApJL

CXL (Compute Express Link) is an emerging open industry-standard interconnect between processing and memory devices that is expected to revolutionize the way systems are designed in the near future. It enables cache-coherent shared memory pools in a disaggregated fashion at unprecedented scales, allowing algorithms to interact with a variety of storage devices using simple loads and stores. Alongside unleashing unique opportunities for a wide range of applications, CXL introduces new challenges of data management and crash consistency. Alas, CXL lacks an adequate programming model, which makes reasoning about the correctness and expected behaviors of algorithms and systems on top of it nearly impossible. In this work, we present CXL0, the first programming model for concurrent programs running on top of CXL. We propose a high-level abstraction for CXL memory accesses and formally define operational semantics on top of that abstraction. We perform initial measurements that provide practical insight into CXL0. We provide a set of general transformations that adapt concurrent algorithms to the new disruptive technology. These transformations enhance linearizable algorithms with durability under a general partial-failure model. We provide an additional transformation for algorithms designed for persistent main memory and full-system crashes. We believe that this work will serve as a stepping stone for systems design and modeling on top of CXL, and support the development of future models as software and hardware evolve.

Language agents, built on top of language models (LMs), are systems that can interact with complex environments, such as the open web. In this work, we examine whether such agents can perform realistic and time-consuming tasks on the web, e.g., monitoring real-estate markets or locating relevant nearby businesses. We introduce AssistantBench, a challenging new benchmark consisting of 214 realistic tasks that can be automatically evaluated, covering different scenarios and domains. We find that AssistantBench exposes the limitations of current systems, including language models and retrieval-augmented language models, as no model reaches an accuracy of more than 26 points. While closed-book LMs perform well in terms of accuracy, they exhibit low precision and tend to hallucinate facts. State-of-the-art web agents reach a score of near zero. Additionally, we introduce SeePlanAct (SPA), a new web agent that significantly outperforms previous agents, and an ensemble of SPA and closed-book models reaches the best overall performance. Moreover, we analyze failures of current systems and highlight that open web navigation remains a major challenge.

Thousands of new scientific papers are published each month. Such information overload complicates researcher efforts to stay current with the state-of-the-art as well as to verify and correctly attribute claims. We pose the following research question: Given a text excerpt referencing a paper, could an LM act as a research assistant to correctly identify the referenced paper? We advance efforts to answer this question by building a benchmark that evaluates the abilities of LMs in citation attribution. Our benchmark, CiteME, consists of text excerpts from recent machine learning papers, each referencing a single other paper. CiteME use reveals a large gap between frontier LMs and human performance, with LMs achieving only 4.2-18.5% accuracy and humans 69.7%. We close this gap by introducing CiteAgent, an autonomous system built on the GPT-4o LM that can also search and read papers, which achieves an accuracy of 35.3\% on CiteME. Overall, CiteME serves as a challenging testbed for open-ended claim attribution, driving the research community towards a future where any claim made by an LM can be automatically verified and discarded if found to be incorrect.

We present the most comprehensive catalog to date of Type I Superluminous Supernovae (SLSNe), a class of stripped envelope supernovae (SNe) characterized by exceptionally high luminosities. We have compiled a sample of 262 SLSNe reported through 2022 December 31. We verified the spectroscopic classification of each SLSN and collated an exhaustive data set of UV, optical and IR photometry from both publicly available data and our own FLEET observational follow-up program, totaling over 30,000 photometric detections. Using these data we derive observational parameters such as the peak absolute magnitudes, rise and decline timescales, as well as bolometric luminosities, temperature and photospheric radius evolution for all SLSNe. Additionally, we model all light curves using a hybrid model that includes contributions from both a magnetar central engine and the radioactive decay of $^{56}$Ni. We explore correlations among various physical and observational parameters, and recover the previously found relation between ejecta mass and magnetar spin, as well as the overall progenitor pre-explosion mass distribution with a peak at $\approx 6.5$ M$_\odot$. We find no significant redshift dependence for any parameter, and no evidence for distinct sub-types of SLSNe. We find that $< 3$\% of SLSNe are best fit with a significant contribution from radioactive decay $\gtrsim 50$\%, representing a set of relatively dim and slowly declining SNe. We provide several analytical tools designed to simulate typical SLSN light curves across a broad range of wavelengths and phases, enabling accurate K-corrections, bolometric scaling calculations, and inclusion of SLSNe in survey simulations or future comparison works. The complete catalog, including all of the photometry, models, and derived parameters, is made available as an open-source resource on GitHub., Comment: 59 pages, 22 Figures, Submitted to MNRAS

Large language models (LLMs) often exhibit undesirable behaviors, such as hallucinations and sequence repetitions. We propose to view these behaviors as fallbacks that models exhibit under epistemic uncertainty, and investigate the connection between them. We categorize fallback behaviors - sequence repetitions, degenerate text, and hallucinations - and extensively analyze them in models from the same family that differ by the amount of pretraining tokens, parameter count, or the inclusion of instruction-following training. Our experiments reveal a clear and consistent ordering of fallback behaviors, across all these axes: the more advanced an LLM is (i.e., trained on more tokens, has more parameters, or instruction-tuned), its fallback behavior shifts from sequence repetitions, to degenerate text, and then to hallucinations. Moreover, the same ordering is observed during the generation of a single sequence, even for the best-performing models; as uncertainty increases, models shift from generating hallucinations to producing degenerate text and finally sequence repetitions. Lastly, we demonstrate that while common decoding techniques, such as random sampling, alleviate unwanted behaviors like sequence repetitions, they increase harder-to-detect hallucinations., Comment: NeurIPS Workshop on Attributing Model Behavior at Scale (ATTRIB 2024)

Existing studies of training state-of-the-art Contrastive Language-Image Pretraining (CLIP) models on large-scale data involve hundreds of or even thousands of GPUs due to the requirement of a large batch size. However, such a large amount of resources is not accessible to most people. While advanced compositional optimization techniques for optimizing global contrastive losses have been demonstrated effective for removing the requirement of large batch size, their performance on large-scale data remains underexplored and not optimized. To bridge the gap, this paper explores several aspects of CLIP training with limited resources (e.g., up to tens of GPUs). First, we introduce FastCLIP, a general CLIP training framework built on advanced compositional optimization techniques while designed and optimized for the distributed setting. Our framework is equipped with an efficient gradient reduction strategy to reduce communication overhead. Second, to further boost training efficiency, we investigate three components of the framework from an optimization perspective: the schedule of the inner learning rate, the update rules of the temperature parameter and the model parameters, respectively. Experiments on different strategies for each component shed light on how to conduct CLIP training more efficiently. Finally, we benchmark the performance of FastCLIP and the state-of-the-art training baseline (OpenCLIP) on different compute scales up to 32 GPUs on 8 nodes, and three data scales ranging from 2.7 million, 9.1 million to 315 million image-text pairs to demonstrate the significant improvement of FastCLIP in the resource-limited setting. We release the code of FastCLIP at https://github.com/Optimization-AI/fast_clip ., Comment: 29 pages

In many reinforcement learning (RL) applications one cannot easily let the agent act in the world; this is true for autonomous vehicles, healthcare applications, and even some recommender systems, to name a few examples. Offline RL provides a way to train agents without real-world exploration, but is often faced with biases due to data distribution shifts, limited coverage, and incomplete representation of the environment. To address these issues, practical applications have tried to combine simulators with grounded offline data, using so-called hybrid methods. However, constructing a reliable simulator is in itself often challenging due to intricate system complexities as well as missing or incomplete information. In this work, we outline four principal challenges for combining offline data with imperfect simulators in RL: simulator modeling error, partial observability, state and action discrepancies, and hidden confounding. To help drive the RL community to pursue these problems, we construct ``Benchmarks for Mechanistic Offline Reinforcement Learning'' (B4MRL), which provide dataset-simulator benchmarks for the aforementioned challenges. Our results suggest the key necessity of such benchmarks for future research.

This study explores the efficacy of thermal splicing conditions between silica and zirconium-fluoride fibers, focusing on achieving mechanical strength between the two fibers. A comprehensive characterization of the thermal profile in the hot zone of the filament splicer was conducted using a fiber Bragg grating, providing valuable insights into its stability and overall performance. Results demonstrate mechanically strong joints and suggest a very narrow temperature window to achieve strong connection between the two materials. Moreover, we characterize the surface composition of the ZrF4 fiber using energy dispersive spectroscopy following splicing at ideal temperatures, as well as at higher and lower temperatures. This work paves the way towards future implementation of silica and fluoride fibers splicing using alternative splicing solutions such as CO2 laser system while raising interesting facts for further studies in the specific field., Comment: * Authors contributed equally

In "VC density in some theories without the independence property" the authors asked whether any partial order of finite width has the VC1 property (i.e. every formula in one variable has UDTFS in one parameter). We give a negative answer and some related remarks.

Various tasks, such as summarization, multi-hop question answering, or coreference resolution, are naturally phrased over collections of real-world documents. Such tasks present a unique set of challenges, revolving around the lack of coherent narrative structure across documents, which often leads to contradiction, omission, or repetition of information. Despite their real-world application and challenging properties, there is currently no benchmark which specifically measures the abilities of large language models (LLMs) on multi-document tasks. To bridge this gap, we present SEAM (a Stochastic Evaluation Approach for Multi-document tasks), a conglomerate benchmark over a diverse set of multi-document datasets, setting conventional evaluation criteria, input-output formats, and evaluation protocols. In particular, SEAM addresses the sensitivity of LLMs to minor prompt variations through repeated evaluations, where in each evaluation we sample uniformly at random the values of arbitrary factors (e.g., the order of documents). We evaluate different LLMs on SEAM finding that multi-document tasks pose a significant challenge for LLMs, even for state-of-the-art models with 70B parameters. In addition, we show that the stochastic approach uncovers underlying statistical trends which cannot be observed in a static benchmark. We hope that SEAM will spur progress via consistent and meaningful evaluation of multi-document tasks.

A principled approach to cyclicality and intransitivity in cardinal paired comparison data is developed within the framework of graphical linear models. Fundamental to our developments is a detailed understanding and study of the parameter space which accommodates cyclicality and intransitivity. In particular, the relationships between the reduced, completely transitive model, the full, not necessarily transitive model, and all manner of intermediate models are explored for both complete and incomplete paired comparison graphs. It is shown that identifying cyclicality and intransitivity reduces to a model selection problem and a new method for model selection employing geometrical insights, unique to the problem at hand, is proposed. The large sample properties of the estimators as well as guarantees on the selected model are provided. It is thus shown that in large samples all cyclicalities and intransitivities can be identified. The method is exemplified using simulations and the analysis of an illustrative example., Comment: 60 pages, 2 figures, 8 tables

The JWST Advanced Deep Extragalactic Survey (JADES) is a multi-cycle JWST program that has taken among the deepest near-/mid-infrared images to date (down to $\sim$30 ABmag) over $\sim$25 arcmin$^2$ in the GOODS-S field in two sets of observations with one year of separation. This presented the first opportunity to systematically search for transients, mostly supernovae (SNe), out to $z$$>$2. We found 79 SNe: 38 at $z$$<$2, 23 at 2$<$$z$$<$3, 8 at 3$<$$z$$<$4, 7 at 4$<$$z$$<$5, and 3 with undetermined redshifts, where the redshifts are predominantly based on spectroscopic or highly reliable JADES photometric redshifts of the host galaxies. At this depth, the detection rate is $\sim$1-2 per arcmin$^2$ per year, demonstrating the power of JWST as a supernova discovery machine. We also conducted multi-band follow-up NIRCam observations of a subset of the SNe to better constrain their light curves and classify their types. Here, we present the survey, sample, search parameters, spectral energy distributions (SEDs), light curves, and classifications. Even at $z$$\geq$2, the NIRCam data quality is high enough to allow SN classification via multi-epoch light-curve fitting with confidence. The multi-epoch SN sample includes a Type Ia SN at $z_{\mathrm{spec}}$$=$2.90, Type IIP SN at $z_{\mathrm{spec}}$$=$3.61, and a Type Ic-BL SN at $z_{\mathrm{spec}}$$=$2.83. We also found that two $z$$\sim$16 galaxy candidates from the first imaging epoch were actually transients that faded in the second epoch, illustrating the possibility that moderate/high-redshift SNe could mimic high-redshift dropout galaxies., Comment: 46 pages, 16 figures, 16 tables. Accepted by ApJ. Appendix A (64 MB) is available at https://drive.google.com/file/d/1xs5jXUVOvdDPgdghK72KR1FMGvPcK7dv/view?usp=sharing . Appendix B (81 MB) is available at https://drive.google.com/file/d/18ImLT80pQdPzXCZA-KEy21DaE2CQiGz1/view?usp=sharing