Search

The gravitational effects of a primordial black hole (PBH) passing through the human body are examined, with the goal of determining the minimum mass necessary to produce significant injury or death. Two effects are examined: the damage caused by a shock wave propagating outward from the black hole trajectory, and the dissociation of brain cells from tidal forces produced by the black hole on its passage through the human body. It is found that the former is the dominant effect, with a cutoff mass for serious injury or death of approximately $M_{PBH} > 1.4 \times 10^{17} {\rm g}$. The number density of primordial black holes with a mass above this cutoff is far too small to produce any observable effects on the human population., Comment: 3 pages, no figures, references and discussion added

Identifying closely related languages at sentence level is difficult, in particular because it is often impossible to assign a sentence to a single language. In this paper, we focus on multi-label sentence-level Scandinavian language identification (LID) for Danish, Norwegian Bokm\r{a}l, Norwegian Nynorsk, and Swedish. We present the Scandinavian Language Identification and Evaluation, SLIDE, a manually curated multi-label evaluation dataset and a suite of LID models with varying speed-accuracy tradeoffs. We demonstrate that the ability to identify multiple languages simultaneously is necessary for any accurate LID method, and present a novel approach to training such multi-label LID models.

While Large Language Models require more and more data to train and scale, rather than looking for any data to acquire, we should consider what types of tasks are more likely to benefit from data scaling. We should be intentional in our data acquisition. We argue that the topology of data itself informs which tasks to prioritize in data scaling, and shapes the development of the next generation of compute paradigms for tasks where data scaling is inefficient, or even insufficient.

In the rapidly evolving landscape of autonomous mobile robots, the emphasis on seamless human-robot interactions has shifted towards autonomous decision-making. This paper delves into the intricate challenges associated with robotic autonomy, focusing on navigation in dynamic environments shared with humans. It introduces an embedded real-time tracking pipeline, integrated into a navigation planning framework for effective person tracking and avoidance, adapting a state-of-the-art 2D LiDAR-based human detection network and an efficient multi-object tracker. By addressing the key components of detection, tracking, and planning separately, the proposed approach highlights the modularity and transferability of each component to other applications. Our tracking approach is validated on a quadruped robot equipped with 270{\deg} 2D-LiDAR against motion capture system data, with the preferred configuration achieving an average MOTA of 85.45% in three newly recorded datasets, while reliably running in real-time at 20 Hz on the NVIDIA Jetson Xavier NX embedded GPU-accelerated platform. Furthermore, the integrated tracking and avoidance system is evaluated in real-world navigation experiments, demonstrating how accurate person tracking benefits the planner in optimizing the generated trajectories, enhancing its collision avoidance capabilities. This paper contributes to safer human-robot cohabitation, blending recent advances in human detection with responsive planning to navigate shared spaces effectively and securely., Comment: Accepted by SAS 2024

We examine several dark energy models with a time-varying equation of state parameter, $w(z)$, to determine what information can be derived by fitting the distance modulus in such models to a constant equation of state parameter, $w_*$. We derive $w_*$ as a function of the model parameters for the Chevallier-Polarski-Linder (CPL) parametrization, and for the Dutta-Scherrer approximation to hilltop quintessence models. We find that all of the models examined here can be well-described by a pivot-like redshift, $z_{pivot}$ at which the value of $w(z)$ in the model is equal to $w_*$. However, the exact value of $z_{pivot}$ is a model-dependent quantity; it varies from $z_{pivot} = 0.22-0.25$ for the CPL models to $z_{pivot} = 0.17-0.20$ for the hilltop quintessence models. Hence, for all of the models considered here, a constant-$w$ fit gives the value of $w$ for $z$ near 0.2. However, given the fairly wide variation in $z_{pivot}$ over even this restricted set of models, the information gained by fitting to a constant value of $w$ seems rather limited., Comment: 7 pages, 4 figures, quality of figures improved, matches published version

Contacts between particles in dense, sheared suspensions are believed to underpin much of their rheology. Roughness and adhesion are known to constrain the relative motion of particles, and thus globally affect the shear response, but an experimental description of how they microscopically influence the transmission of forces and relative displacements within contacts is lacking. Here we show that an innovative colloidal-probe atomic force microscopy technique allows the simultaneous measurement of normal and tangential forces exchanged between tailored surfaces and microparticles while tracking their relative sliding and rolling, unlocking the direct measurement of coefficients of rolling friction, as well as of sliding friction. We demonstrate that, in the presence of sufficient traction, particles spontaneously roll, reducing dissipation and promoting longer-lasting contacts. Conversely, when rolling is prevented, friction is greatly enhanced for rough and adhesive surfaces, while smooth particles coated by polymer brushes maintain well-lubricated contacts. We find that surface roughness induces rolling due to load-dependent asperity interlocking, leading to large off-axis particle rotations. In contrast, smooth, adhesive surfaces promote rolling along the principal axis of motion. Our results offer direct values of friction coefficients for numerical studies and an interpretation of the onset of discontinuous shear thickening based on them, opening up new ways to tailor rheology via contact engineering., Comment: 7 Figures

Certain holographic dark energy (HDE) models allow for the possibility of a ``long freeze,'' in which the scale factor evolves to a constant in the long-time limit. Here we extend previous calculations by adding a nonrelativistic matter component. The addition of a matter component tends to destroy the long freeze behavior, driving the universe to recollapse. Long freeze evolution is still possible, but only for a limited set of HDE models., Comment: v3, 7 pages, minor name corrections

In this work, we present a new unsupervised anomaly (outlier) detection (AD) method using the sliced-Wasserstein metric. This filtering technique is conceptually interesting for MLOps pipelines deploying machine learning models in critical sectors, e.g., energy, as it offers a conservative data selection. Additionally, we open the first dataset showcasing localized critical peak rebate demand response in a northern climate. We demonstrate the capabilities of our method on synthetic datasets as well as standard AD datasets and use it in the making of a first benchmark for our open-source localized critical peak rebate dataset.

Self-interested individuals often fail to cooperate, posing a fundamental challenge for multi-agent learning. How can we achieve cooperation among self-interested, independent learning agents? Promising recent work has shown that in certain tasks cooperation can be established between learning-aware agents who model the learning dynamics of each other. Here, we present the first unbiased, higher-derivative-free policy gradient algorithm for learning-aware reinforcement learning, which takes into account that other agents are themselves learning through trial and error based on multiple noisy trials. We then leverage efficient sequence models to condition behavior on long observation histories that contain traces of the learning dynamics of other agents. Training long-context policies with our algorithm leads to cooperative behavior and high returns on standard social dilemmas, including a challenging environment where temporally-extended action coordination is required. Finally, we derive from the iterated prisoner's dilemma a novel explanation for how and when cooperation arises among self-interested learning-aware agents.

We show that some holographic dark energy models can lead to a future evolution of the universe in which the scale factor $a$ is asymptotically constant, while $\dot a \rightarrow 0$ and the corresponding energy and pressure densities also vanish. We provide specific examples of such models and general conditions that can lead to an asymptotically static universe, which we have called the ``long freeze." We show that in some cases, such evolution can follow an arbitrarily long exponential expansion essentially identical to the asymptotic evolution of $\Lambda$CDM., Comment: v2, references added, 5 pages with no figures

Classifiers are used throughout industry to enforce policies, ranging from the detection of toxic content to age-appropriate content filtering. While these classifiers serve important functions, it is also essential that they are built in ways that minimize unfair biases for users. One such fairness consideration is called group fairness, which desires that different sub-population of users receive equal treatment. This is a well-studied problem in the context of 'classical' classifiers. However, the emergence of prompt-based language model (LM) decision making has created new opportunities to solve text-based classification tasks, and the fairness properties of these new classifiers are not yet well understood. Further, the `remediation toolkit' is incomplete for LM-based decision makers and little is understood about how to improve decision maker group fairness while maintaining classifier performance. This work sets out to add more tools to that toolbox. We introduce adaptations of existing effective approaches from the classical classifier fairness to the prompt-based classifier space. We also devise simple methods that take advantage of the new structure of prompt-based decision makers and operate at the prompt level. We compare these approaches empirically on real data. Our results suggest that adaptations of approaches that are effective for classical classifiers remain effective in the LM-based classifier environment. However, there is room for further exploration of prompt-based remediation methods (and other remediation methods that take advantage of LM structure).

We use contextualized word definitions generated by large language models as semantic representations in the task of diachronic lexical semantic change detection (LSCD). In short, generated definitions are used as `senses', and the change score of a target word is retrieved by comparing their distributions in two time periods under comparison. On the material of five datasets and three languages, we show that generated definitions are indeed specific and general enough to convey a signal sufficient to rank sets of words by the degree of their semantic change over time. Our approach is on par with or outperforms prior non-supervised sense-based LSCD methods. At the same time, it preserves interpretability and allows to inspect the reasons behind a specific shift in terms of discrete definitions-as-senses. This is another step in the direction of explainable semantic change modeling., Comment: Findings of ACL 2024

We examine the power spectrum $P_n$ of the $n$th harmonic for Garfinkle-Vachaspati cosmic strings, which correspond to planar rectangular loops. While these loop tractories are self-intersecting, a slightly perturbed non-self-intersecting form of these trajectories has been suggested as a generic end state of cosmic string loop fragmentation. We find that $P_n$ scales as $n^{-2} \ln n$, rather than the expected $n^{-2}$ for kink-kink collisions. This result is demonstrated analytically for even $n$ in square loops and numerically for all other cases. At lowest order, the effect of loop decays is to further enhance $P_n$ at large $n$ relative to its value at small $n$. The consequences for relic stochastic background radiation along with the caveats pertaining to our results are discussed., Comment: 12 pages, 7 figures, typos corrected

A recent advance in networking is the deployment of path-aware multipath network architectures, where network endpoints are given multiple network paths to send their data on. In this work, we tackle the challenge of selecting paths for latency-sensitive applications. Even today's path-aware networks, which are much smaller than the current Internet, already offer dozens and in several cases over a hundred paths to a given destination, making it impractical to measure all path latencies to find the lowest latency path. Furthermore, for short flows, performing latency measurements may not provide benefits as the flow may finish before completing the measurements. To overcome these issues, we argue that endpoints should be provided with a latency estimate before sending any packets, enabling latency-aware path choice for the first packet sent. As we cannot predict the end-to-end latency due to dynamically changing queuing delays, we measure and disseminate the propagation latency, enabling novel use cases and solving concrete problems in current network protocols. We present the Global Latency Information Dissemination System (GLIDS), which is a step toward global latency transparency through the dissemination of propagation latency information., Comment: Extended version of the 7-page conference paper "Toward Global Latency Transparency" published at the IFIP Networking 2024 Conference

Forensic authorship profiling uses linguistic markers to infer characteristics about an author of a text. This task is paralleled in dialect classification, where a prediction is made about the linguistic variety of a text based on the text itself. While there have been significant advances in recent years in variety classification, forensic linguistics rarely relies on these approaches due to their lack of transparency, among other reasons. In this paper we therefore explore the explainability of machine learning approaches considering the forensic context. We focus on variety classification as a means of geolinguistic profiling of unknown texts based on social media data from the German-speaking area. For this, we identify the lexical items that are the most impactful for the variety classification. We find that the extracted lexical features are indeed representative of their respective varieties and note that the trained models also rely on place names for classifications.

This paper introduces v0.5 of the AI Safety Benchmark, which has been created by the MLCommons AI Safety Working Group. The AI Safety Benchmark has been designed to assess the safety risks of AI systems that use chat-tuned language models. We introduce a principled approach to specifying and constructing the benchmark, which for v0.5 covers only a single use case (an adult chatting to a general-purpose assistant in English), and a limited set of personas (i.e., typical users, malicious users, and vulnerable users). We created a new taxonomy of 13 hazard categories, of which 7 have tests in the v0.5 benchmark. We plan to release version 1.0 of the AI Safety Benchmark by the end of 2024. The v1.0 benchmark will provide meaningful insights into the safety of AI systems. However, the v0.5 benchmark should not be used to assess the safety of AI systems. We have sought to fully document the limitations, flaws, and challenges of v0.5. This release of v0.5 of the AI Safety Benchmark includes (1) a principled approach to specifying and constructing the benchmark, which comprises use cases, types of systems under test (SUTs), language and context, personas, tests, and test items; (2) a taxonomy of 13 hazard categories with definitions and subcategories; (3) tests for seven of the hazard categories, each comprising a unique set of test items, i.e., prompts. There are 43,090 test items in total, which we created with templates; (4) a grading system for AI systems against the benchmark; (5) an openly available platform, and downloadable tool, called ModelBench that can be used to evaluate the safety of AI systems on the benchmark; (6) an example evaluation report which benchmarks the performance of over a dozen openly available chat-tuned language models; (7) a test specification for the benchmark.

We explore the asymptotic future evolution of holographic dark energy (HDE) models, in which the density of the dark energy is a function of a cutoff scale $L$. We develop a general methodology to determine which models correspond to future big rip, little rip, and pseudo-rip (de Sitter) evolution, and we apply this methodology to a variety of well-studied HDE models. None of these HDE models display little rip evolution, and we are able to show, under very general assumptions, that HDE models with a Granda-Oliveros cutoff almost never evolve toward a future little rip. We extend these results to HDE models with nonstandard Friedman equations and show that a similar conclusion applies: little rip evolution is a very special case that is almost never realized in such models., Comment: v3, matches the published version in PRD, minor clarifications and typos corrected, 9 pages with no figures

We examine big bang nucleosynthesis (BBN) in models with a time-varying gravitational constant $G$, when this time variation is rapid on the scale of the expansion rate $H$, i.e, $\dot G/G \gg H$. Such models can arise naturally in the context of scalar-tensor theories of gravity and result in additional terms in the Friedman equation. We examine two representative models: a step-function evolution for $G$ and a rapidly-oscillating $G$. In the former case, the additional terms in the Friedman equation tend to cancel the effects of an initial value of $G$ that differs from the present-day value. In the case of deuterium, this effect is large enough to reverse the sign of the change in (D/H) for a given change in the initial value of $G$. For rapidly-oscillating $G$, the effect on the Friedman equation is similar to that of adding a vacuum energy density, and BBN allows upper limits to be placed on the product of the oscillation frequency and amplitude. The possibility that a rapidly oscillating $G$ could mimic a cosmological constant is briefly discussed., Comment: 8 pages, 8 figures, discussion and clarification added, including expanded discussion of the lithium problem, matches version to appear in Phys. Rev. D

FinanceBench is a first-of-its-kind test suite for evaluating the performance of LLMs on open book financial question answering (QA). It comprises 10,231 questions about publicly traded companies, with corresponding answers and evidence strings. The questions in FinanceBench are ecologically valid and cover a diverse set of scenarios. They are intended to be clear-cut and straightforward to answer to serve as a minimum performance standard. We test 16 state of the art model configurations (including GPT-4-Turbo, Llama2 and Claude2, with vector stores and long context prompts) on a sample of 150 cases from FinanceBench, and manually review their answers (n=2,400). The cases are available open-source. We show that existing LLMs have clear limitations for financial QA. Notably, GPT-4-Turbo used with a retrieval system incorrectly answered or refused to answer 81% of questions. While augmentation techniques such as using longer context window to feed in relevant evidence improve performance, they are unrealistic for enterprise settings due to increased latency and cannot support larger financial documents. We find that all models examined exhibit weaknesses, such as hallucinations, that limit their suitability for use by enterprises., Comment: Dataset is available at: https://huggingface.co/datasets/PatronusAI/financebench

The past year has seen rapid acceleration in the development of large language models (LLMs). However, without proper steering and safeguards, LLMs will readily follow malicious instructions, provide unsafe advice, and generate toxic content. We introduce SimpleSafetyTests (SST) as a new test suite for rapidly and systematically identifying such critical safety risks. The test suite comprises 100 test prompts across five harm areas that LLMs, for the vast majority of applications, should refuse to comply with. We test 11 open-access and open-source LLMs and four closed-source LLMs, and find critical safety weaknesses. While some of the models do not give a single unsafe response, most give unsafe responses to more than 20% of the prompts, with over 50% unsafe responses in the extreme. Prepending a safety-emphasising system prompt substantially reduces the occurrence of unsafe responses, but does not completely stop them from happening. Trained annotators labelled every model response to SST (n = 3,000). We use these annotations to evaluate five AI safety filters (which assess whether a models' response is unsafe given a prompt) as a way of automatically evaluating models' performance on SST. The filters' performance varies considerably. There are also differences across the five harm areas, and on the unsafe versus safe responses. The widely-used Perspective API has 72% accuracy and a newly-created zero-shot prompt to OpenAI's GPT-4 performs best with 89% accuracy. Content Warning: This paper contains prompts and responses that relate to child abuse, suicide, self-harm and eating disorders, scams and fraud, illegal items, and physical harm.

Photonic integrated circuits are paving the way for novel on-chip functionalities with diverse applications in communication, computing, and beyond. The integration of on-chip light sources, especially single-mode lasers, is crucial for advancing those photonic chips to their full potential. Recently, novel concepts involving topological designs introduced a variety of options for tuning device properties such as the desired single mode emission. Here we introduce a novel cavity design that allows to amplify the topological interface mode by deterministic placement of gain material within the topological lattice. The proposed design is experimentally implemented by a selective epitaxy process resulting in Si and InGaAs nanorods embedded within the same topological lattice. This results in the first demonstration of a single-mode laser in the telecom band using the concept of amplified topological modes.

Some autoregressive models exhibit in-context learning capabilities: being able to learn as an input sequence is processed, without undergoing any parameter changes, and without being explicitly trained to do so. The origins of this phenomenon are still poorly understood. Here we analyze a series of Transformer models trained to perform synthetic sequence prediction tasks, and discover that standard next-token prediction error minimization gives rise to a subsidiary learning algorithm that adjusts the model as new inputs are revealed. We show that this process corresponds to gradient-based optimization of a principled objective function, which leads to strong generalization performance on unseen sequences. Our findings explain in-context learning as a product of autoregressive loss minimization and inform the design of new optimization-based Transformer layers.

This paper presents a case study on the design, administration, post-processing, and evaluation of surveys on large language models (LLMs). It comprises two components: (1) A statistical method for eliciting beliefs encoded in LLMs. We introduce statistical measures and evaluation metrics that quantify the probability of an LLM "making a choice", the associated uncertainty, and the consistency of that choice. (2) We apply this method to study what moral beliefs are encoded in different LLMs, especially in ambiguous cases where the right choice is not obvious. We design a large-scale survey comprising 680 high-ambiguity moral scenarios (e.g., "Should I tell a white lie?") and 687 low-ambiguity moral scenarios (e.g., "Should I stop for a pedestrian on the road?"). Each scenario includes a description, two possible actions, and auxiliary labels indicating violated rules (e.g., "do not kill"). We administer the survey to 28 open- and closed-source LLMs. We find that (a) in unambiguous scenarios, most models "choose" actions that align with commonsense. In ambiguous cases, most models express uncertainty. (b) Some models are uncertain about choosing the commonsense action because their responses are sensitive to the question-wording. (c) Some models reflect clear preferences in ambiguous scenarios. Specifically, closed-source models tend to agree with each other.

Modern DDoS defense systems rely on probabilistic monitoring algorithms to identify flows that exceed a volume threshold and should thus be penalized. Commonly, classic sketch algorithms are considered sufficiently accurate for usage in DDoS defense. However, as we show in this paper, these algorithms achieve poor detection accuracy under burst-flood attacks, i.e., volumetric DDoS attacks composed of a swarm of medium-rate sub-second traffic bursts. Under this challenging attack pattern, traditional sketch algorithms can only detect a high share of the attack bursts by incurring a large number of false positives. In this paper, we present ALBUS, a probabilistic monitoring algorithm that overcomes the inherent limitations of previous schemes: ALBUS is highly effective at detecting large bursts while reporting no legitimate flows, and therefore improves on prior work regarding both recall and precision. Besides improving accuracy, ALBUS scales to high traffic rates, which we demonstrate with an FPGA implementation, and is suitable for programmable switches, which we showcase with a P4 implementation., Comment: Accepted at the 42nd International Symposium on Reliable Distributed Systems (SRDS 2023)

This report presents the results of the shared tasks organized as part of the VarDial Evaluation Campaign 2023. The campaign is part of the tenth workshop on Natural Language Processing (NLP) for Similar Languages, Varieties and Dialects (VarDial), co-located with EACL 2023. Three separate shared tasks were included this year: Slot and intent detection for low-resource language varieties (SID4LR), Discriminating Between Similar Languages -- True Labels (DSL-TL), and Discriminating Between Similar Languages -- Speech (DSL-S). All three tasks were organized for the first time this year.

Internet service providers (ISPs) have a variety of quality attributes that determine their attractiveness for data transmission, ranging from quality-of-service metrics such as jitter to security properties such as the presence of DDoS defense systems. ISPs should optimize these attributes in line with their profit objective, i.e., maximize revenue from attracted traffic while minimizing attribute-related cost, all in the context of alternative offers by competing ISPs. However, this attribute optimization is difficult not least because many aspects of ISP competition are barely understood on a systematic level, e.g., the multi-dimensional and cost-driving nature of path quality, and the distributed decision making of ISPs on the same path. In this paper, we improve this understanding by analyzing how ISP competition affects path quality and ISP profits. To that end, we develop a game-theoretic model in which ISPs (i) affect path quality via multiple attributes that entail costs, (ii) are on paths together with other selfish ISPs, and (iii) are in competition with alternative paths when attracting traffic. The model enables an extensive theoretical analysis, surprisingly showing that competition can have both positive and negative effects on path quality and ISP profits, depending on the network topology and the cost structure of ISPs. However, a large-scale simulation, which draws on real-world data to instantiate the model, shows that the positive effects will likely prevail in practice: If the number of selectable paths towards any destination increases from 1 to 5, the prevalence of quality attributes increases by at least 50%, while 75% of ISPs improve their profit., Comment: Accepted at the International Symposium on Computer Performance, Modeling, Measurements and Evaluation (PERFORMANCE) 2023

We report on a study of the Multiview Observatory for Solar Terrestrial Science (MOST) mission that will provide comprehensive imagery and time series data needed to understand the magnetic connection between the solar interior and the solar atmosphere/inner heliosphere. MOST will build upon the successes of SOHO and STEREO missions with new views of the Sun and enhanced instrument capabilities. This article is based on a study conducted at NASA Goddard Space Flight Center that determined the required instrument refinement, spacecraft accommodation, launch configuration, and flight dynamics for mission success. MOST is envisioned as the next generation great observatory positioned to obtain three-dimensional information of large-scale heliospheric structures such as coronal mass ejections, stream interaction regions, and the solar wind itself. The MOST mission consists of 2 pairs of spacecraft located in the vicinity of Sun-Earth Lagrange points L4 (MOST1, MOST3) and L5 (MOST2 and MOST4). The spacecraft stationed at L4 (MOST1) and L5 (MOST2) will each carry seven remote-sensing and three in-situ instrument suites, including a novel radio package known as the Faraday Effect Tracker of Coronal and Heliospheric structures (FETCH). MOST3 and MOST4 will carry only the FETCH instruments and are positioned at variable locations along the Earth orbit up to 20{\deg} ahead of L4 and 20{\deg} behind L5, respectively. FETCH will have polarized radio transmitters and receivers on all four spacecraft to measure the magnetic content of solar wind structures propagating from the Sun to Earth using the Faraday rotation technique. The MOST mission will be able to sample the magnetized plasma throughout the Sun-Earth connected space during the mission lifetime over a solar cycle., Comment: 42 pages, 19 figures, 8 tables, to appear in J. Atmospheric and Solar Terrestrial Physics