Search

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: 25 pages, 6 figures

In this paper, we propose an extension to the Propagator algorithm for source bearing estimation by performing root decomposition which eliminates the need for spectral search over angles. Further the propagator spatial spectrum is reused after performing the polynomial rooting to alleviate the observed drawback of an increase in RMSE at lower SNRs for estimating the directions-of-arrival of signals impinging on a linear array of sensors. Our proposed algorithm achieves a 98% reduction in computational complexity compared to Propagator alongside an improved angular resolution suitable for real time DOA estimation for wireless communications., Comment: 5 pages

This research presents a numerical investigation of the flow and heat transfer of a steady dusty flow over a linear horizontal stretching sheet. Transverse force effects have been taken into account. The flow problems formulation comprises of highly nonlinear PDEs that have been transformed into systems of ODEs by using similarity transformation. Then the ODEs has been solved numerically by using Shooting technique followed by Runge-Kutta 4th order method that is incorporated in BVP4C tool of the MATLAB software. The effects various flow parameters, such as the Prandtl number, Eckert number and transverse force on the flow geometry has been investigated. The overall findings are displayed in graphs and tables, and it is discovered that the transverse force reduces the velocity of the particle phase in the flow. Our results has been validated with existing literatures and found agrees with very good way.

In real-world sequential decision making tasks like autonomous driving, robotics, and healthcare, learning from observed state-action trajectories is critical for tasks like imitation, classification, and clustering. For example, self-driving cars must replicate human driving behaviors, while robots and healthcare systems benefit from modeling decision sequences, whether or not they come from expert data. Existing trajectory encoding methods often focus on specific tasks or rely on reward signals, limiting their ability to generalize across domains and tasks. Inspired by the success of embedding models like CLIP and BERT in static domains, we propose a novel method for embedding state-action trajectories into a latent space that captures the skills and competencies in the dynamic underlying decision-making processes. This method operates without the need for reward labels, enabling better generalization across diverse domains and tasks. Our contributions are threefold: (1) We introduce a trajectory embedding approach that captures multiple abilities from state-action data. (2) The learned embeddings exhibit strong representational power across downstream tasks, including imitation, classification, clustering, and regression. (3) The embeddings demonstrate unique properties, such as controlling agent behaviors in IQ-Learn and an additive structure in the latent space. Experimental results confirm that our method outperforms traditional approaches, offering more flexible and powerful trajectory representations for various applications. Our code is available at https://github.com/Erasmo1015/vte., Comment: AAMAS 2025

Purpose: Diabetic retinopathy (DR) is a major cause of vision loss, particularly in India, where access to retina specialists is limited in rural areas. This study aims to evaluate the Artificial Intelligence-based Diabetic Retinopathy Screening System (AIDRSS) for DR detection and prevalence assessment, addressing the growing need for scalable, automated screening solutions in resource-limited settings. Approach: A multicentric, cross-sectional study was conducted in Kolkata, India, involving 5,029 participants and 10,058 macula-centric retinal fundus images. The AIDRSS employed a deep learning algorithm with 50 million trainable parameters, integrated with Contrast Limited Adaptive Histogram Equalization (CLAHE) preprocessing for enhanced image quality. DR was graded using the International Clinical Diabetic Retinopathy (ICDR) Scale, categorizing disease into five stages (DR0 to DR4). Statistical metrics including sensitivity, specificity, and prevalence rates were evaluated against expert retina specialist assessments. Results: The prevalence of DR in the general population was 13.7%, rising to 38.2% among individuals with elevated random blood glucose levels. The AIDRSS achieved an overall sensitivity of 92%, specificity of 88%, and 100% sensitivity for detecting referable DR (DR3 and DR4). These results demonstrate the system's robust performance in accurately identifying and grading DR in a diverse population. Conclusions: AIDRSS provides a reliable, scalable solution for early DR detection in resource-constrained environments. Its integration of advanced AI techniques ensures high diagnostic accuracy, with potential to significantly reduce the burden of diabetes-related vision loss in underserved regions., Comment: 22 pages, 5 figures. arXiv admin note: substantial text overlap with arXiv:1812.07105 by other authors without attribution

Spin states in semiconductors provide exceptionally stable and noise-resistant environments for qubits, positioning them as optimal candidates for reliable quantum computing technologies. The proposal to use nuclear and electronic spins of donor atoms in silicon, introduced by Kane in 1998, sparked a new research field focused on the precise positioning of individual impurity atoms for quantum devices, utilising scanning tunnelling microscopy and ion implantation. This roadmap article reviews the advancements in the 25 years since Kane's proposal, the current challenges, and the future directions in atomic-scale semiconductor device fabrication and measurement. It covers the quest to create a silicon-based quantum computer and expands to include diverse material systems and fabrication techniques, highlighting the potential for a broad range of semiconductor quantum technological applications. Key developments include phosphorus in silicon devices such as single-atom transistors, arrayed few-donor devices, one- and two-qubit gates, three-dimensional architectures, and the development of a toolbox for future quantum integrated circuits. The roadmap also explores new impurity species like arsenic and antimony for enhanced scalability and higher-dimensional spin systems, new chemistry for dopant precursors and lithographic resists, and the potential for germanium-based devices. Emerging methods, such as photon-based lithography and electron beam manipulation, are discussed for their disruptive potential. This roadmap charts the path toward scalable quantum computing and advanced semiconductor quantum technologies, emphasising the critical intersections of experiment, technological development, and theory., Comment: 94 pages

Higher-order skyrmions and antiskyrmions are topologically protected spin textures with an integer topological charge other than $\pm 1$ and nucleate from topological point defects in regular Bloch walls, known as vertical Bloch lines. So far, they have only been observed using Lorentz transmission electron microscopy. In this work, we show that higher-order spin textures coexisting in Co/Ni multilayers at room temperature can be visualized by high-resolution magnetic force microscopy (MFM). The experimental results are supported by micromagnetic simulations confirming that different spin objects give rise to distinct MFM contrast in full agreement to our observations., Comment: 7 pages, 5 figures

Dielectric elastomers have significant potential for new technologies ranging from soft robots to biomedical devices, driven by their ability to display complex shape changes in response to electrical stimulus. However, an important shortcoming of current realizations is that large voltages are required for useful actuation strains. This work proposes, and demonstrates through theory and numerical simulations, a strategy to achieve large and controlled actuation by exploiting the electromechanical analog of the Treloar-Kearsley (TK) instability. The key idea is to use the fact that the TK instability is a symmetry-breaking bifurcation, which implies the existence of a symmetry-driven constant-energy region in the energy landscape. This provides for nonlinear soft modes with large deformations that can be accessed with very small external stimulus, which is achieved here by applying a small in-plane electric field. First, the bifurcation and post-bifurcation behavior of the electromechanical TK instability are established theoretically in the idealized setting of uniform deformation and electric field. Next, building on this, a finite element analysis of a realistic geometry with patterned top and bottom electrodes is applied to demonstrate large and soft shape changes driven by small voltage differences across the electrodes.

Diabetic Retinopathy (DR) is a leading cause of preventable blindness. Early detection at the DR1 stage is critical but is hindered by a scarcity of high-quality fundus images. This study uses StyleGAN3 to generate synthetic DR1 images characterized by microaneurysms with high fidelity and diversity. The aim is to address data scarcity and enhance the performance of supervised classifiers. A dataset of 2,602 DR1 images was used to train the model, followed by a comprehensive evaluation using quantitative metrics, including Frechet Inception Distance (FID), Kernel Inception Distance (KID), and Equivariance with respect to translation (EQ-T) and rotation (EQ-R). Qualitative assessments included Human Turing tests, where trained ophthalmologists evaluated the realism of synthetic images. Spectral analysis further validated image quality. The model achieved a final FID score of 17.29, outperforming the mean FID of 21.18 (95 percent confidence interval - 20.83 to 21.56) derived from bootstrap resampling. Human Turing tests demonstrated the model's ability to produce highly realistic images, though minor artifacts near the borders were noted. These findings suggest that StyleGAN3-generated synthetic DR1 images hold significant promise for augmenting training datasets, enabling more accurate early detection of Diabetic Retinopathy. This methodology highlights the potential of synthetic data in advancing medical imaging and AI-driven diagnostics., Comment: 13 pages, 11 figures

We present OLMo 2, the next generation of our fully open language models. OLMo 2 includes dense autoregressive models with improved architecture and training recipe, pretraining data mixtures, and instruction tuning recipes. Our modified model architecture and training recipe achieve both better training stability and improved per-token efficiency. Our updated pretraining data mixture introduces a new, specialized data mix called Dolmino Mix 1124, which significantly improves model capabilities across many downstream task benchmarks when introduced via late-stage curriculum training (i.e. specialized data during the annealing phase of pretraining). Finally, we incorporate best practices from T\"ulu 3 to develop OLMo 2-Instruct, focusing on permissive data and extending our final-stage reinforcement learning with verifiable rewards (RLVR). Our OLMo 2 base models sit at the Pareto frontier of performance to compute, often matching or outperforming open-weight only models like Llama 3.1 and Qwen 2.5 while using fewer FLOPs and with fully transparent training data, code, and recipe. Our fully open OLMo 2-Instruct models are competitive with or surpassing open-weight only models of comparable size, including Qwen 2.5, Llama 3.1 and Gemma 2. We release all OLMo 2 artifacts openly -- models at 7B and 13B scales, both pretrained and post-trained, including their full training data, training code and recipes, training logs and thousands of intermediate checkpoints. The final instruction model is available on the Ai2 Playground as a free research demo., Comment: Model demo available at playground.allenai.org

In optimizing real-world structures, due to fabrication or budgetary restraints, the design variables may be restricted to a set of standard engineering choices. Such variables, commonly called categorical variables, are discrete and unordered in essence, precluding the utilization of gradient-based optimizers for the problems containing them. In this paper, incorporating the Gumbel-Softmax (GSM) method, we propose a new gradient-based optimizer for handling such variables in the optimal design of large-scale frame structures. The GSM method provides a means to draw differentiable samples from categorical distributions, thereby enabling sensitivity analysis for the variables generated from such distributions. The sensitivity information can greatly reduce the computational cost of traversing high-dimensional and discrete design spaces in comparison to employing gradient-free optimization methods. In addition, since the developed optimizer is gradient-based, it can naturally handle the simultaneous optimization of categorical and continuous design variables. Through three numerical case studies, different aspects of the proposed optimizer are studied and its advantages over population-based optimizers, specifically a genetic algorithm, are demonstrated., Comment: Published in Structural and Multidisciplinary Optimization

Photonics/light localization techniques are important in understanding the structural changes in biological tissues at the nano- to sub-micron scale. It is now known that structural alteration starts at the nanoscale at the beginning of cancer progression. This study examines the molecular-specific nano-structural alterations of chronic alcoholism and probiotic effects on colon cancer using a mouse model of colon cancer. Confocal microscopy and mesoscopic light-scattering analysis are applied to quantify structural changes in DNA (chromatin), cytoskeleton, and ki-67 protein cells with appropriate staining dyes. We assessed alcohol-treated and azoxymethane (AOM) with dextran sulfate sodium (DSS)-induced colitis models, including ethanol (EtOH) and probiotic (L.Casei) treatments separately and together. The inverse participation ratio (IPR) technique was employed to quantify the degree of light localization to access the molecular-specific spatial structural disorder as a biomarker for cancer progression detection. Significant enhancement of cancer progression was observed in the alcohol-treated group, and probiotics treatment with alcohol showed partial reversal of these changes in colon cancer. The results underscore the potential of the IPR technique in detecting early structural changes in colon cancer, offering insights into the mitigating effects of probiotics on alcohol-induced enhancement of colon cancer., Comment: 17 pages, 8 figures

An ethical dilemma describes a choice between two "right" options involving conflicting moral values. We present a comprehensive evaluation of how LLMs navigate ethical dilemmas. Specifically, we investigate LLMs on their (1) sensitivity in comprehending ethical dilemmas, (2) consistency in moral value choice, (3) consideration of consequences, and (4) ability to align their responses to a moral value preference explicitly or implicitly specified in a prompt. Drawing inspiration from a leading ethical framework, we construct a dataset comprising 1,730 ethical dilemmas involving four pairs of conflicting values. We evaluate 20 well-known LLMs from six families. Our experiments reveal that: (1) LLMs exhibit pronounced preferences between major value pairs, and prioritize truth over loyalty, community over individual, and long-term over short-term considerations. (2) The larger LLMs tend to support a deontological perspective, maintaining their choices of actions even when negative consequences are specified. (3) Explicit guidelines are more effective in guiding LLMs' moral choice than in-context examples. Lastly, our experiments highlight the limitation of LLMs in comprehending different formulations of ethical dilemmas.

The growing demand for dynamic, user-centric data analysis and visualization is evident across domains like healthcare, finance, and research. Traditional visualization tools often fail to meet individual user needs due to their static and predefined nature. To address this gap, Text2Insight is introduced as an innovative solution that delivers customized data analysis and visualizations based on user-defined natural language requirements. Leveraging a multi-model architecture, Text2Insight transforms user inputs into actionable insights and dynamic visualizations. The methodology begins with analyzing the input dataset to extract structural details such as columns and values. A pre-trained Llama3 model converts the user's natural language query into an SQL query, which is further refined using a Named Entity Recognition (NER) model for accuracy. A chart predictor determines the most suitable visualization type, while the Llama3 model generates insights based on the SQL query's results. The output is a user-friendly and visually informative chart. To enhance analysis capabilities, the system integrates a question-answering model and a predictive model using the BERT framework. These models provide insights into historical data and predict future trends. Performance evaluation of Text2Insight demonstrates its effectiveness, achieving high accuracy (99%), precision (100%), recall (99%), and F1-score (99%), with a BLEU score of 0.5. The question-answering model attained an accuracy of 89% and the predictive model achieved 70% accuracy. These results validate Text2Insight as a robust and viable solution for transforming natural language text into dynamic, user-specific data analysis and visualizations.

The commercialization prospects of perovskite light emitting diodes depend on its luminescence efficiency under large carrier densities. The decrease in luminescence efficiency under such high injection conditions could lead to an undesired increase in power consumption with associated degradation and stability concerns. Here, through detailed modeling of thermal transport and carrier generation-recombination, we unravel the physical mechanisms that cause luminescence droop under high injection conditions. We show that self-heating leads to a reduction in the radiative recombination (both bimolecular and excitonic). The resultant increase in non-radiative recombination and hence the thermal dissipation acts as a positive feedback mechanism that leads to efficiency droop in perovskites. Our model predictions, well supported by experimental results, could be of broad interest towards the degradation-aware thermal design of perovskite optoelectronics and stability.

Magnons, the quanta of spin angular momentum, can be excited in magnetic insulators by spin-flip scattering processes originated from currents applied to a heavy metal overlayer. The efficiency to generate non-equilibrium magnons across interfaces is parametrized by the spin conductance gs, a phenomenological constant that is considered to be dependent on thermal magnons. Here, we investigate non-linear magnetoresistance phenomena originated in Pt due to current-driven nonequilibrium magnons in Y3Fe5O12 (YIG). Remarkably, we find that spin-flip scattering processes are dominated by subthermal magnons at room temperature, resulting in a large modulation of gs with the magnetic field and YIG thickness. Concretely, we find that decreasing the YIG thickness from 100 to 10 nm increases gs by a factor 30 and observe that the magnetic field exponentially suppresses the magnon generation efficiency. These findings challenge current understanding on gs and indicate that electrically-driven magnonic effects such as damping compensation and magnon condensation can be largely boosted by device miniaturization., Comment: 13 pages, 5 figures

As quantum hardware continues to scale, managing the heterogeneity of resources and applications -- spanning diverse quantum and classical hardware and software frameworks -- becomes increasingly critical. Pilot-Quantum addresses these challenges as a middleware designed to provide unified application-level management of resources and workloads across hybrid quantum-classical environments. It is built on a rigorous analysis of existing quantum middleware systems and application execution patterns. It implements the Pilot Abstraction conceptual model, originally developed for HPC, to manage resources, workloads, and tasks. It is designed for quantum applications that rely on task parallelism, including: (i) Hybrid algorithms, such as variational approaches, and (ii) Circuit cutting systems, used to partition and execute large quantum circuits. Pilot-Quantum facilitates seamless integration of quantum processing units (QPUs), classical CPUs, and GPUs, while supporting high-level programming frameworks like Qiskit and Pennylane. This enables users to design and execute hybrid workflows across diverse computing resources efficiently. The capabilities of Pilot-Quantum are demonstrated through mini-applications -- simplified yet representative kernels focusing on critical performance bottlenecks. We present several mini-apps, including circuit execution across hardware and simulator platforms (e.g., IBM's Eagle QPU), distributed state vector simulation, circuit cutting, and quantum machine learning workflows, demonstrating significant scale (e.g., a 41-qubit simulation on 256 GPUs) and speedups (e.g., 15x for QML, 3.5x for circuit cutting).

Crack segmentation plays a crucial role in ensuring the structural integrity and seismic safety of civil structures. However, existing crack segmentation algorithms encounter challenges in maintaining accuracy with domain shifts across datasets. To address this issue, we propose a novel deep network that employs incremental training with unsupervised domain adaptation (UDA) using adversarial learning, without a significant drop in accuracy in the source domain. Our approach leverages an encoder-decoder architecture, consisting of both domain-invariant and domain-specific parameters. The encoder learns shared crack features across all domains, ensuring robustness to domain variations. Simultaneously, the decoder's domain-specific parameters capture domain-specific features unique to each domain. By combining these components, our model achieves improved crack segmentation performance. Furthermore, we introduce BuildCrack, a new crack dataset comparable to sub-datasets of the well-established CrackSeg9K dataset in terms of image count and crack percentage. We evaluate our proposed approach against state-of-the-art UDA methods using different sub-datasets of CrackSeg9K and our custom dataset. Our experimental results demonstrate a significant improvement in crack segmentation accuracy and generalization across target domains compared to other UDA methods - specifically, an improvement of 0.65 and 2.7 mIoU on source and target domains respectively., Comment: Accepted at ICPR 2024. Details and code can be accessed from https://crackuda.github.io

Organic-inorganic hybrid perovskite solar cells (PSC) have demonstrated impressive performance improvement. Among the various characteristics, the time-dependent current-voltage (J-V) hysteresis allows a direct exploration of various critical phenomena that affect the stability of PSCs. The hysteresis is associated with various spatial heterogeneity-related phenomena, including lifetime, bandgap, and phase segregation. We investigate these phenomena through numerical simulations and quantify how the spatial non-uniformity in the perovskite active layer impacts the hysteresis. Further, we correlate the time dependent device degradation with the hysteresis trends in terms of ion density and effective carrier lifetime.

Evaluating the capability of Large Language Models (LLMs) in following instructions has heavily relied on a powerful LLM as the judge, introducing unresolved biases that deviate the judgments from human judges. In this work, we reevaluate various choices for automatic evaluation on a wide range of instruction-following tasks. We experiment with methods that leverage human-written responses and observe that they enhance the reliability of automatic evaluations across a wide range of tasks, resulting in up to a 3.2% improvement in agreement with human judges. We also discovered that human-written responses offer an orthogonal perspective to model-generated responses in following instructions and should be used as an additional context when comparing model responses. Based on these observations, we develop a new evaluation benchmark, Human Response-Guided Evaluation of Instruction Following (HREF), comprising 4,258 samples across 11 task categories with a composite evaluation setup, employing a composite evaluation setup that selects the most reliable method for each category. In addition to providing reliable evaluation, HREF emphasizes individual task performance and is free from contamination. Finally, we study the impact of key design choices in HREF, including the size of the evaluation set, the judge model, the baseline model, and the prompt template. We host a live leaderboard that evaluates LLMs on the private evaluation set of HREF., Comment: 28 pages, 15 figures

Offline safe reinforcement learning (RL) has emerged as a promising approach for learning safe behaviors without engaging in risky online interactions with the environment. Most existing methods in offline safe RL rely on cost constraints at each time step (derived from global cost constraints) and this can result in either overly conservative policies or violation of safety constraints. In this paper, we propose to learn a policy that generates desirable trajectories and avoids undesirable trajectories. To be specific, we first partition the pre-collected dataset of state-action trajectories into desirable and undesirable subsets. Intuitively, the desirable set contains high reward and safe trajectories, and undesirable set contains unsafe trajectories and low-reward safe trajectories. Second, we learn a policy that generates desirable trajectories and avoids undesirable trajectories, where (un)desirability scores are provided by a classifier learnt from the dataset of desirable and undesirable trajectories. This approach bypasses the computational complexity and stability issues of a min-max objective that is employed in existing methods. Theoretically, we also show our approach's strong connections to existing learning paradigms involving human feedback. Finally, we extensively evaluate our method using the DSRL benchmark for offline safe RL. Empirically, our method outperforms competitive baselines, achieving higher rewards and better constraint satisfaction across a wide variety of benchmark tasks., Comment: AAAI 2025

This summary of the second Terrestrial Very-Long-Baseline Atom Interferometry (TVLBAI) Workshop provides a comprehensive overview of our meeting held in London in April 2024, building on the initial discussions during the inaugural workshop held at CERN in March 2023. Like the summary of the first workshop, this document records a critical milestone for the international atom interferometry community. It documents our concerted efforts to evaluate progress, address emerging challenges, and refine strategic directions for future large-scale atom interferometry projects. Our commitment to collaboration is manifested by the integration of diverse expertise and the coordination of international resources, all aimed at advancing the frontiers of atom interferometry physics and technology, as set out in a Memorandum of Understanding signed by over 50 institutions., Comment: Summary of the second Terrestrial Very-Long-Baseline Atom Interferometry Workshop held at Imperial College London: https://indico.cern.ch/event/1369392/

Orthognathic surgery consultation is essential to help patients understand the changes to their facial appearance after surgery. However, current visualization methods are often inefficient and inaccurate due to limited pre- and post-treatment data and the complexity of the treatment. To overcome these challenges, this study aims to develop a fully automated pipeline that generates accurate and efficient 3D previews of postsurgical facial appearances for patients with orthognathic treatment without requiring additional medical images. The study introduces novel aesthetic losses, such as mouth-convexity and asymmetry losses, to improve the accuracy of facial surgery prediction. Additionally, it proposes a specialized parametric model for 3D reconstruction of the patient, medical-related losses to guide latent code prediction network optimization, and a data augmentation scheme to address insufficient data. The study additionally employs FLAME, a parametric model, to enhance the quality of facial appearance previews by extracting facial latent codes and establishing dense correspondences between pre- and post-surgery geometries. Quantitative comparisons showed the algorithm's effectiveness, and qualitative results highlighted accurate facial contour and detail predictions. A user study confirmed that doctors and the public could not distinguish between machine learning predictions and actual postoperative results. This study aims to offer a practical, effective solution for orthognathic surgery consultations, benefiting doctors and patients., Comment: 9 pages, 5 figures

Public health practitioners often have the goal of monitoring patients and maximizing patients' time spent in "favorable" or healthy states while being constrained to using limited resources. Restless multi-armed bandits (RMAB) are an effective model to solve this problem as they are helpful to allocate limited resources among many agents under resource constraints, where patients behave differently depending on whether they are intervened on or not. However, RMABs assume the reward function is known. This is unrealistic in many public health settings because patients face unique challenges and it is impossible for a human to know who is most deserving of any intervention at such a large scale. To address this shortcoming, this paper is the first to present the use of inverse reinforcement learning (IRL) to learn desired rewards for RMABs, and we demonstrate improved outcomes in a maternal and child health telehealth program. First we allow public health experts to specify their goals at an aggregate or population level and propose an algorithm to design expert trajectories at scale based on those goals. Second, our algorithm WHIRL uses gradient updates to optimize the objective, allowing for efficient and accurate learning of RMAB rewards. Third, we compare with existing baselines and outperform those in terms of run-time and accuracy. Finally, we evaluate and show the usefulness of WHIRL on thousands on beneficiaries from a real-world maternal and child health setting in India. We publicly release our code here: https://github.com/Gjain234/WHIRL.

Large Language Models (LLMs) generating unsafe responses to toxic prompts is a significant issue in their applications. While various efforts aim to address this safety concern, previous approaches often demand substantial human data collection or rely on the less dependable option of using another LLM to generate corrective data. In this paper, we aim to take this problem and overcome limitations of requiring significant high-quality human data. Our method requires only a small set of unsafe responses to toxic prompts, easily obtained from the unsafe LLM itself. By employing a semantic cost combined with a negative Earth Mover Distance (EMD) loss, we guide the LLM away from generating unsafe responses. Additionally, we propose a novel lower bound for EMD loss, enabling more efficient optimization. Our results demonstrate superior performance and data efficiency compared to baselines, and we further examine the nuanced effects of over-alignment and potential degradation of language capabilities when using contrastive data.

The promise of unsupervised multi-view-stereo (MVS) is to leverage large unlabeled datasets, yet current methods underperform when training on difficult data, such as handheld smartphone videos of indoor scenes. Meanwhile, high-quality synthetic datasets are available but MVS networks trained on these datasets fail to generalize to real-world examples. To bridge this gap, we propose a semi-supervised learning framework that allows us to train on real and rendered images jointly, capturing structural priors from synthetic data while ensuring parity with the real-world domain. Central to our framework is a novel set of losses that leverages powerful existing monocular relative-depth estimators trained on the synthetic dataset, transferring the rich structure of this relative depth to the MVS predictions on unlabeled data. Inspired by perceptual image metrics, we compare the MVS and monocular predictions via a deep feature loss and a multi-scale statistical loss. Our full framework, which we call Prism, achieves large quantitative and qualitative improvements over current unsupervised and synthetic-supervised MVS networks. This is a best-case-scenario result, opening the door to using both unlabeled smartphone videos and photorealistic synthetic datasets for training MVS networks., Comment: 11 pages, 6 figures, 3 tables

Cryogenic fluids have extensive applications as fuel for launch vehicles in space applications and research. The physics of cryogenic flows are highly complex due to the sensitive nature of phase transformation from liquid to bubbly liquid and vapor, eventually resulting in cavitating flows at the ambient temperature owing to the very low boiling point of cryogenic fluids, which asserts us to classify such flows under multi-phase flow physics regime. This work elucidates the modeling of bubbly flow for cryogenic fluids such as liquid nitrogen in a converging-diverging venturi-like flow device known as cavitating venturi, a passive flow control metering device. The numerical works in literature are usually limited to modeling iso-thermal bubbly flows such as water devoid of involving energy equations because there is no occurrence of interface heat transfer as latent heat of vaporization of water is higher, unlike cryogenic fluids which are sensitive to phase change at ambient conditions. So, to realize an appropriate model for modeling cryogenic bubbly flows such as liquid nitrogen flow, the effect of heat transfer at the interface and convective heat transfer from the surrounding liquid to the traversing bubble needs to be included. Numerical modeling using an in-house code involving a finite-difference method The numerical results showed the importance of including the heat transport equation due to convection and at the interface of bubble-fluid as a significant source term for the bubble dynamics. The work is supported by computational simulation using a commercial CFD package for 2-dimensional simulations to predict a characterizing parameter, namely cavitation length. A limited flow visualization experiment using a high-speed camera is performed to study the cavitating zone length.

The separation of physical forces acting on the tip of a magnetic force microscope (MFM) is essential for correct magnetic imaging. Electrostatic forces can be modulated by varying the tip-sample potential and minimized to map the local Kelvin potential. However, distinguishing magnetic forces from van der Waals forces typically requires two measurements with opposite tip magnetizations under otherwise identical measurement conditions. Here, we present an inverted magnetic force microscope where the sample is mounted on a flat cantilever for force sensing, and the magnetic tip is attached to a miniaturized electromagnet that periodically flips the tip magnetization. This setup enables the extraction of magnetic tip-sample interactions from the sidebands occurring at the switching rate in the cantilever oscillation spectrum. Our method achieves the separation of magnetic signals from other force contributions in a single-scan mode. Future iterations of this setup may incorporate membrane, trampoline, or string resonators with ultra-high quality factors, potentially improving measurement sensitivity by up to three orders of magnitude compared to the state-of-the-art MFM systems using cantilevers., Comment: 7 pages, 4 figures in Main and 3 pages, 3 figures in supplement

Autoregressive (AR) models have achieved remarkable success in natural language and image generation, but their application to 3D shape modeling remains largely unexplored. Unlike diffusion models, AR models enable more efficient and controllable generation with faster inference times, making them especially suitable for data-intensive domains. Traditional 3D generative models using AR approaches often rely on ``next-token" predictions at the voxel or point level. While effective for certain applications, these methods can be restrictive and computationally expensive when dealing with large-scale 3D data. To tackle these challenges, we introduce 3D-WAG, an AR model for 3D implicit distance fields that can perform unconditional shape generation, class-conditioned and also text-conditioned shape generation. Our key idea is to encode shapes as multi-scale wavelet token maps and use a Transformer to predict the ``next higher-resolution token map" in an autoregressive manner. By redefining 3D AR generation task as ``next-scale" prediction, we reduce the computational cost of generation compared to traditional ``next-token" prediction models, while preserving essential geometric details of 3D shapes in a more structured and hierarchical manner. We evaluate 3D-WAG to showcase its benefit by quantitative and qualitative comparisons with state-of-the-art methods on widely used benchmarks. Our results show 3D-WAG achieves superior performance in key metrics like Coverage and MMD, generating high-fidelity 3D shapes that closely match the real data distribution.

Language model post-training is applied to refine behaviors and unlock new skills across a wide range of recent language models, but open recipes for applying these techniques lag behind proprietary ones. The underlying training data and recipes for post-training are simultaneously the most important pieces of the puzzle and the portion with the least transparency. To bridge this gap, we introduce Tulu 3, a family of fully-open state-of-the-art post-trained models, alongside its data, code, and training recipes, serving as a comprehensive guide for modern post-training techniques. Tulu 3, which builds on Llama 3.1 base models, achieves results surpassing the instruct versions of Llama 3.1, Qwen 2.5, Mistral, and even closed models such as GPT-4o-mini and Claude 3.5-Haiku. The training algorithms for our models include supervised finetuning (SFT), Direct Preference Optimization (DPO), and a novel method we call Reinforcement Learning with Verifiable Rewards (RLVR). With Tulu 3, we introduce a multi-task evaluation scheme for post-training recipes with development and unseen evaluations, standard benchmark implementations, and substantial decontamination of existing open datasets on said benchmarks. We conclude with analysis and discussion of training methods that did not reliably improve performance. In addition to the Tulu 3 model weights and demo, we release the complete recipe -- including datasets for diverse core skills, a robust toolkit for data curation and evaluation, the training code and infrastructure, and, most importantly, a detailed report for reproducing and further adapting the Tulu 3 approach to more domains.

Trilayer nickelates are a rich class of materials exhibiting diverse correlated phenomena, including superconductivity, density wave transitions, non-Fermi liquid behavior along with an unusual metal-to-metal transition around T* ~ 150 K. Understanding the electronic correlations, lattice and charge dynamics are crucial to unreveal the origin of superconductivity and other instabilities in nickelates. Our in-depth Raman measurements shows that trilayer nickelate, La4Ni3O10, shows transition from electron-phonon coupled phase to the electron-electron correlated one below charge density wave transition around T* with an estimated energy gap of ~ 18-20 meV. The transition around T* is also accompanied by the emergence of zone folded phonon modes reflecting the transition into the charge density wave phase. Phonon modes self-energy parameters show anomalous changes around T* attributed to the electron-electron correlations, and renormalization rate of the phonon modes is much slower in the charge-ordered phase compared to the phase above T*. The transition around T* are marked by the suppression of electron-phonon coupling parameter by ~ 70 %, a change of the quasiparticle dynamics from non-Fermi liquid to the Landau-Fermi liquid type behaviour estimated using the low frequency Raman response.

Calculation of the coherent nonlinear response of a system is essential to correctly interpret results from advanced techniques such as two-dimensional coherent spectroscopy (2DCS). Usually, even for the simplest systems, such calculations are either performed for low-intensity excitations where perturbative methods are valid and/or by assuming a simplified pulse envelope, such as a {\delta}-function in time. We present exact calculations using the phase-cycling method without making the aforementioned approximations. We introduce a generalized version of the phase-cycling method to isolate an arbitrary N-wave mixing signal. We then apply this method to model the saturation of the nonlinear signal from excitons in semiconductor quantum wells, which is consistent with 2DCS experiments. We also present simulation results that replicate previously-reported experiments with high-intensity excitation of semiconductor quantum dots. By accurately reproducing a variety of phenomena such as higher-order contributions, switching of coherent signal, and changes in photon-echo transients, we prove the efficacy of the phase-cycling method to calculate the coherent nonlinear signal for high-intensity excitation. This method would be particularly useful for systems with multiple, well-separated peaks and/or large inhomogeneity.

This report provides an initial look at partial results from the TREC 2024 Retrieval-Augmented Generation (RAG) Track. We have identified RAG evaluation as a barrier to continued progress in information access (and more broadly, natural language processing and artificial intelligence), and it is our hope that we can contribute to tackling the many challenges in this space. The central hypothesis we explore in this work is that the nugget evaluation methodology, originally developed for the TREC Question Answering Track in 2003, provides a solid foundation for evaluating RAG systems. As such, our efforts have focused on "refactoring" this methodology, specifically applying large language models to both automatically create nuggets and to automatically assign nuggets to system answers. We call this the AutoNuggetizer framework. Within the TREC setup, we are able to calibrate our fully automatic process against a manual process whereby nuggets are created by human assessors semi-manually and then assigned manually to system answers. Based on initial results across 21 topics from 45 runs, we observe a strong correlation between scores derived from a fully automatic nugget evaluation and a (mostly) manual nugget evaluation by human assessors. This suggests that our fully automatic evaluation process can be used to guide future iterations of RAG systems.

The application of large language models to provide relevance assessments presents exciting opportunities to advance information retrieval, natural language processing, and beyond, but to date many unknowns remain. This paper reports on the results of a large-scale evaluation (the TREC 2024 RAG Track) where four different relevance assessment approaches were deployed in situ: the "standard" fully manual process that NIST has implemented for decades and three different alternatives that take advantage of LLMs to different extents using the open-source UMBRELA tool. This setup allows us to correlate system rankings induced by the different approaches to characterize tradeoffs between cost and quality. We find that in terms of nDCG@20, nDCG@100, and Recall@100, system rankings induced by automatically generated relevance assessments from UMBRELA correlate highly with those induced by fully manual assessments across a diverse set of 77 runs from 19 teams. Our results suggest that automatically generated UMBRELA judgments can replace fully manual judgments to accurately capture run-level effectiveness. Surprisingly, we find that LLM assistance does not appear to increase correlation with fully manual assessments, suggesting that costs associated with human-in-the-loop processes do not bring obvious tangible benefits. Overall, human assessors appear to be stricter than UMBRELA in applying relevance criteria. Our work validates the use of LLMs in academic TREC-style evaluations and provides the foundation for future studies.

Arch filament systems (AFSs) are chromospheric and coronal manifestations of emerging magnetic flux. Using high spatial resolution observations taken at a high cadence by the Extreme Ultraviolet Imager (EUI) on board Solar Orbiter, we identified small-scale elongated brightenings within the AFSs. These brightenings appear as bidirectional flows along the threads of AFSs. For our study, we investigated the coordinated observations of the AFSs acquired by the EUI and the Atmospheric Imaging Assembly (AIA) on board SDO on 2022 March 4 and 17. We analyzed 15 bidirectional propagating brightenings from EUI 174 {\AA} images. These brightenings reached propagating speeds of 100-150 km/s. The event observed on March 17 exhibits blob-like structures, which may be signatures of plasmoids and due to magnetic reconnection. In this case, we also observed counterparts in the running difference slit-jaw images in the 1400 {\AA} passbands taken by the Interface Region Imaging Spectrograph (IRIS). Most events show co-temporal intensity variations in all AIA EUV passbands. Together, this implies that these brightenings in the AFSs are dominated by emission from cool plasma with temperatures well below 1 MK. The magnetograms taken by the Polarimetric and Helioseismic Imager (PHI) on board Solar Orbiter show signatures of flux emergence beneath the brightenings. This suggests that the events in the AFSs are triggered by magnetic reconnection that may occur between the newly emerging magnetic flux and the preexisting magnetic field structures in the middle of the AFSs. This would also give a natural explanation for the bidirectional propagation of the brightenings near the apex of the AFSs. The interaction of the preexisting field and the emerging flux may be important for mass and energy transfer within the AFSs., Comment: Accepted for publication in A&A

Social media has become an integral part of modern life, but it has also brought with it the pervasive issue of cyberbullying a serious menace in today's digital age. Cyberbullying, a form of harassment that occurs on social networks, has escalated alongside the growth of these platforms. Sentiment analysis holds significant potential not only for detecting bullying phrases but also for identifying victims who are at high risk of harm, whether to themselves or others. Our work focuses on leveraging deep learning and natural language understanding techniques to detect traces of bullying in social media posts. We developed a Recurrent Neural Network with Long Short-Term Memory (LSTM) cells, using different embeddings. One approach utilizes BERT embeddings, while the other replaces the embeddings layer with the recently released embeddings API from OpenAI. We conducted a performance comparison between these two approaches to evaluate their effectiveness in sentiment analysis of Formspring Cyberbullying data. Our Code is Available at https://github.com/ppujari/xcs224u

Recent advances have demonstrated that large language models (LLMs) excel as listwise rerankers, but their high computational demands remain a barrier to widespread adoption. Further, the traditional language modeling (LM) objective is not ideally suited for reranking tasks. FIRST is a novel approach that addresses these challenges by integrating a learning-to-rank objective and leveraging the logits of only the first generated token, thereby significantly reducing inference latency compared to traditional LLM rerankers. In this study, we extend the evaluation of FIRST to the TREC Deep Learning datasets (DL19-22), validating its robustness across diverse domains. We investigate the influence of different first-stage retrievers on FIRST rerankers, observing diminishing returns and patterns consistent with traditional LLM rerankers. Through applying the FIRST objective to a broader range of backbone models, we achieve effectiveness surpassing the original implementation. Our experiments confirm that fast reranking with single-token logits does not compromise out-of-domain reranking quality. To better quantify the computational savings in the original study, we measure and compare latency to find a 21%-42% gain across various models and benchmarks. Moreover, while LM training implicitly improves zero-shot single-token reranking, our experiments also raise questions about whether LM pre-training may hinder subsequent fine-tuning with the FIRST objective. These findings pave the way for more efficient and effective listwise reranking in future applications.

We experimentally realise the theoretical proposal for in-situ tunable photonic edge states emerging from qubits coupled to a waveguide with a photonic bandgap. These edge states are directional, exhibiting theoretically zero population in the opposite direction. Our experiment implements a tunable Rice-Mele waveguide configuration, where the directionality of edge states is controlled in-situ by varying the qubit energy. The Rice-Mele waveguide is constructed using lumped resonators coupled to a standard Xmon qubit. We demonstrate the existence of these edge states both actively, via waveguide transmission, and passively, through qubit emission via an edge state. We estimate a 99.4% fidelity in the directionality, constrained by our measurement noise floor. These results hold significant promise for the development of long-range qubit couplers with effectively zero crosstalk.

Multi-objective reinforcement learning (MORL) is used to solve problems involving multiple objectives. An MORL agent must make decisions based on the diverse signals provided by distinct reward functions. Training an MORL agent yields a set of solutions (policies), each presenting distinct trade-offs among the objectives (expected returns). MORL enhances explainability by enabling fine-grained comparisons of policies in the solution set based on their trade-offs as opposed to having a single policy. However, the solution set is typically large and multi-dimensional, where each policy (e.g., a neural network) is represented by its objective values. We propose an approach for clustering the solution set generated by MORL. By considering both policy behavior and objective values, our clustering method can reveal the relationship between policy behaviors and regions in the objective space. This approach can enable decision makers (DMs) to identify overarching trends and insights in the solution set rather than examining each policy individually. We tested our method in four multi-objective environments and found it outperformed traditional k-medoids clustering. Additionally, we include a case study that demonstrates its real-world application.

Recent research has demonstrated that training a linear connector between speech foundation encoders and large language models (LLMs) enables this architecture to achieve strong ASR capabilities. Despite the impressive results, it remains unclear whether these simple approaches are robust enough across different scenarios and speech conditions, such as domain shifts and speech perturbations. In this paper, we address these questions by conducting various ablation experiments using a recent and widely adopted approach called SLAM-ASR. We present novel empirical findings that offer insights on how to effectively utilize the SLAM-ASR architecture across a wide range of settings. Our main findings indicate that SLAM-ASR exhibits poor performance in cross-domain evaluation settings. Additionally, speech perturbations on in-domain data, such as changes in speech rate or additive noise, can significantly degrade performance. Our findings offer critical insights for fine-tuning and configuring robust LLM-based ASR models, tailored to different data characteristics and computational resources., Comment: Accepted in ICASSP 2025 SALMA Workshop

Multi-object multi-part scene segmentation is a challenging task whose complexity scales exponentially with part granularity and number of scene objects. To address the task, we propose a plug-and-play approach termed OLAF. First, we augment the input (RGB) with channels containing object-based structural cues (fg/bg mask, boundary edge mask). We propose a weight adaptation technique which enables regular (RGB) pre-trained models to process the augmented (5-channel) input in a stable manner during optimization. In addition, we introduce an encoder module termed LDF to provide low-level dense feature guidance. This assists segmentation, particularly for smaller parts. OLAF enables significant mIoU gains of $\mathbf{3.3}$ (Pascal-Parts-58), $\mathbf{3.5}$ (Pascal-Parts-108) over the SOTA model. On the most challenging variant (Pascal-Parts-201), the gain is $\mathbf{4.0}$. Experimentally, we show that OLAF's broad applicability enables gains across multiple architectures (CNN, U-Net, Transformer) and datasets. The code is available at olafseg.github.io, Comment: Accepted in The European Conference on Computer Vision (ECCV) 2024

The center-to-limb variations (CLV) of transition region line Gaussian fit parameters in solar plage are reported for the first time. The Si iv 1402.77 A line observed by Interface Region Imaging Spectrograph (IRIS) is used. The spectral intensity increases linearly from the disk center to the solar limb. Similarly, the non-thermal velocity also increases linearly from 23.6 km/s (at the disk center) to 30.9 km/s (at the solar limb). On the other hand, the Doppler velocity decreases from 8.9$\pm$1.0 km/s at the disk center to 0.0 km/s at the limb. This CLV pattern in solar plages is consistent with the CLV pattern reported in the quiet-Sun (QS). However, the average values of the parameters in the solar plage are significantly higher than in the QS. The intensity and non-thermal velocity increase linearly with the magnetic field at the disk center while the Doppler velocity does not depend on the magnetic field. Due to the line-of-sight effect, the plasma column depth increases towards the solar limb which leads to a linear increase in the spectral intensity. Further, the increasing plasma column depth towards the solar limb adds more and more unresolved motions, and as a result, the non-thermal velocity increases from the disk center to the solar limb. In the solar plages, the higher plasma density due to the strong magnetic field leads to the higher intensity and non-thermal velocity compared to QS and coronal hole (CH), Comment: 23 pages, 19 Figures, Accepted for Publication in The Astrophysical Journal (ApJ)

Recent trends in lower precision, e.g. half-precision floating point, training have shown improved system performance and reduced memory usage for Deep Learning while maintaining accuracy. However, current GNN systems cannot achieve such goals for GNN, as our analyses show that they massively underperform while showing abnormal accuracy when using half-precision. These systems suffer from value overflow issues due to lowered precision, under-utilization of hardware resources, and poor training performance. To mitigate this, we introduce HalfGNN, a half-precision based GNN system. HalfGNN proposes novel techniques: new vector operations for half-precision data types that improve data load and reduction performance, and discretized SpMM that overcomes the value overflow and natively provides workload balancing. Such techniques improve hardware utilization, reduce memory usage, and remove atomic writes. Evaluations show that HalfGNN achieves on average of 2.30X speedup in training time over DGL (float-based) for GAT, GCN, and GIN respectively while achieving similar accuracy, and saving 2.67X memory.

Recent years have witnessed the rapid development of Neuro-Symbolic (NeSy) AI systems, which integrate symbolic reasoning into deep neural networks. However, most of the existing benchmarks for NeSy AI fail to provide long-horizon reasoning tasks with complex multi-agent interactions. Furthermore, they are usually constrained by fixed and simplistic logical rules over limited entities, making them far from real-world complexities. To address these crucial gaps, we introduce LogiCity, the first simulator based on customizable first-order logic (FOL) for an urban-like environment with multiple dynamic agents. LogiCity models diverse urban elements using semantic and spatial concepts, such as IsAmbulance(X) and IsClose(X, Y). These concepts are used to define FOL rules that govern the behavior of various agents. Since the concepts and rules are abstractions, they can be universally applied to cities with any agent compositions, facilitating the instantiation of diverse scenarios. Besides, a key feature of LogiCity is its support for user-configurable abstractions, enabling customizable simulation complexities for logical reasoning. To explore various aspects of NeSy AI, LogiCity introduces two tasks, one features long-horizon sequential decision-making, and the other focuses on one-step visual reasoning, varying in difficulty and agent behaviors. Our extensive evaluation reveals the advantage of NeSy frameworks in abstract reasoning. Moreover, we highlight the significant challenges of handling more complex abstractions in long-horizon multi-agent scenarios or under high-dimensional, imbalanced data. With its flexible design, various features, and newly raised challenges, we believe LogiCity represents a pivotal step forward in advancing the next generation of NeSy AI. All the code and data are open-sourced at our website., Comment: 25 pages, 8 figures

We consider the linear causal representation learning setting where we observe a linear mixing of $d$ unknown latent factors, which follow a linear structural causal model. Recent work has shown that it is possible to recover the latent factors as well as the underlying structural causal model over them, up to permutation and scaling, provided that we have at least $d$ environments, each of which corresponds to perfect interventions on a single latent node (factor). After this powerful result, a key open problem faced by the community has been to relax these conditions: allow for coarser than perfect single-node interventions, and allow for fewer than $d$ of them, since the number of latent factors $d$ could be very large. In this work, we consider precisely such a setting, where we allow a smaller than $d$ number of environments, and also allow for very coarse interventions that can very coarsely \textit{change the entire causal graph over the latent factors}. On the flip side, we relax what we wish to extract to simply the \textit{list of nodes that have shifted between one or more environments}. We provide a surprising identifiability result that it is indeed possible, under some very mild standard assumptions, to identify the set of shifted nodes. Our identifiability proof moreover is a constructive one: we explicitly provide necessary and sufficient conditions for a node to be a shifted node, and show that we can check these conditions given observed data. Our algorithm lends itself very naturally to the sample setting where instead of just interventional distributions, we are provided datasets of samples from each of these distributions. We corroborate our results on both synthetic experiments as well as an interesting psychometric dataset. The code can be found at https://github.com/TianyuCodings/iLCS., Comment: NeurIPS 2024

Recent reports indicate that perovskite based light emitting diodes (LEDs) have achieved an external quantum efficiency (EQE) of 32% - rather an internal quantum efficiency close to 100%. Much of this improved performance is attributed to the role of excitons. While the experimental trends are encouraging, the recombination parameters estimated through extensive curve-fitting of photoluminescence (PL) transients are often not amenable to reasonable interpretations. In view of the same, through a detailed analysis of free carrier - exciton dynamics in perovskite optoelectronic materials, here we identify a coherent scheme to unambiguously back extract the relevant parameters. The model predictions compare well with the recent experimental results on perovskite LEDs with record EQE thus quantifying the role of excitons. Importantly, this work identifies a physics aware scheme for the design of experiments tailored towards consistent exploration of underlying physical mechanisms and hence could enable a synergistic optimization of process technology and device performance.

We study the prospect of equivalence principle at the quantum regime by investigating the transition probabilities of a two-level atomic detector in different scenarios. In particular, two specific set-ups are considered. ($i$) {\it Without a boundary}: In one scenario the atom is in uniform acceleration and interacting with Minkowski field modes. While in the other the atom is static and in interaction with Rindler field modes. ($ii$) {\it With a reflecting boundary}: In one scenario the atom is uniformly accelerated and the mirror is static, and in the other scenario the atom is static and the mirror is in uniform acceleration. In these cases the atom is interacting with the field modes, defined in mirror's frame. For both the set-ups, the focus is on the excitation and de-excitation probabilities in $(1+1)$ and $(3+1)$ spacetime dimensions. Our observations affirm that in $(1+1)$ dimensions, for both set-ups the transition probabilities from different scenarios become the same when the atomic and the field frequencies are equal. In contrast, in $(3+1)$ dimensions this equivalence is not observed in general, inspiring us to look for a deeper physical interpretation. Our findings suggest that when the equivalence between different scenarios is concerned, the excitation to de-excitation ratio provides a more consistent measure even in $(3+1)$ dimensions. We discuss the physical interpretation and implications of our findings., Comment: 16 pages, 0 figure, revtex4