Search

THz amplitude modulators and switches are considered to be the main building blocks of future THz communication systems. Despite rapid progress, modulation and switching devices in this electromagnetic spectrum lag far behind other frequency ranges. Currently, THz modu-lators face major challenges in consistently producing high modulations depths over large frequency bands. Moreover, a convenient integration for practical applications requires that the modulation/switching properties can be electrically controlled. Devices fulfilling all these con-ditions remain to be demonstrated. In this work we show that W-doped VO2 films grown by direct-current magnetron sputtering can be efficiently used for the development reliable, large-area, broadband THz waves modulators. We demonstrate that W doping not only permits to tune the insulator to metal transition (IMT) temperature of VO2, but also, most importantly, to control the topology of the electrically activated transition. In situ / operando X-ray diffraction and Raman spectroscopy characterizations of the devices, coupled with standard resistivi-ty measurements and time-domain THz spectroscopy, unambiguously demonstrate that the changes in the spatial distribution of the IMT is due to structural distortions induced by W doping.

As large language models (LLMs) start interacting with each other and generating an increasing amount of text online, it becomes crucial to better understand how information is transformed as it passes from one LLM to the next. While significant research has examined individual LLM behaviors, existing studies have largely overlooked the collective behaviors and information distortions arising from iterated LLM interactions. Small biases, negligible at the single output level, risk being amplified in iterated interactions, potentially leading the content to evolve towards attractor states. In a series of telephone game experiments, we apply a transmission chain design borrowed from the human cultural evolution literature: LLM agents iteratively receive, produce, and transmit texts from the previous to the next agent in the chain. By tracking the evolution of text toxicity, positivity, difficulty, and length across transmission chains, we uncover the existence of biases and attractors, and study their dependence on the initial text, the instructions, language model, and model size. For instance, we find that more open-ended instructions lead to stronger attraction effects compared to more constrained tasks. We also find that different text properties display different sensitivity to attraction effects, with toxicity leading to stronger attractors than length. These findings highlight the importance of accounting for multi-step transmission dynamics and represent a first step towards a more comprehensive understanding of LLM cultural dynamics., Comment: Code available at https://github.com/jeremyperez2/TelephoneGameLLM. Companion website with a Data Explorer tool at https://sites.google.com/view/telephone-game-llm

The present work aims to determine the source regions of carbonaceous chondrites (CM, CI, CO, CV, CK, CR, CH, CB, or C-ungrouped). We studied 38 individual asteroid families, including young and old ones, and determined their contributions to the NEO populations at metre and kilometre sizes using collisional and orbital models. Our models are in agreement with spectroscopic observations of NEOs, cosmic-ray exposure ages of meteorites, statistics of bolides, infrared emission from dust bands, composition of interplanetary dust particles (IDPs), or abundance of extraterrestrial helium-3. We identified the Veritas, Polana and Eos families as the primary sources of CM/CR, CI and CO/CV/CK chondrites, respectively. Substantial contributions are also expected from CM-like K\"onig and CI-like Clarissa, Misa and Hoffmeister families. The source regions of kilometre-sized bodies are generally different. The Adeona family is by far the main source of CM-like NEOs, whereas the Polana (low-i) and Euphrosyne (high-i) families are at the origin of most CI-like NEOs. The Polana family is the likely source of both Ryugu and Bennu. We were able to link spectroscopically and dynamically several NEOs to the Baptistina family. Finally, it appears that the pre-atmospheric flux of carbonaceous chondrites at metre sizes is about the same as that of ordinary chondrites. Given the difference in fall statistics between the two groups (80\% versus 4.4\%), this implies either substantial atmospheric fragmentation of carbonaceous bodies at the level of ${\sim}0.5\,{\rm MPa}$, or destruction by thermal cracking and water desorption. The source regions of most meteorites and kilometre-sized NEOs have now been determined; including some minor classes like enstatite chondrites and achondrites (Nysa, Hungaria), acapulcoites/lodranites (Iannini)., Comment: 47 pages, 31 figures, accepted in A&A

We show how to compute the purity and entanglement entropy for quantum fields in a systematic perturbative expansion. To that end, we generalize the in-in formalism to non-unitary dynamics (i.e. accounting for the presence of an environment) and to the calculation of quantum information measures, which are not observables in the usual sense. This allows us to reduce the problem to one involving standard correlation functions, and to organize their computation in a diagrammatic expansion for which we construct the corresponding Feynman rules. As an illustration, we apply the formalism to a cosmological setting inspired by the effective field theory of inflation. We find that at late times, non-linear loop corrections share the same time behavior as the linear contribution, and only yield a slight redressing of the purity. In particular, when the environment is heavy compared to the Hubble scale, the phenomenon of recoherence previously encountered is robust to the class of non-linear extensions considered. Bridging the gap between perturbative quantum field theory and open quantum systems paves the way to a better understanding of renormalization and resummation in open effective field theories. It also enables a more systematic exploration of quantum information properties in field theoretic settings., Comment: 43 pages without appendices (66 pages in total), 14 figures, matches published version in JCAP

Centaurs, distinguished by their volatile-rich compositions, play a pivotal role in understanding the formation and evolution of the early solar system, as they represent remnants of the primordial material that populated the outer regions. Stellar occultations offer a means to investigate their physical properties, including shape, rotational state, or the potential presence of satellites and rings. This work aims to conduct a detailed study of the centaur (54598) Bienor through stellar occultations and rotational light curves from photometric data collected during recent years. We successfully predicted three stellar occultations by Bienor, which were observed from Japan, Eastern Europe, and the USA. In addition, we organized observational campaigns from Spain to obtain rotational light curves. At the same time, we develop software to generate synthetic light curves from three-dimensional shape models, enabling us to validate the outcomes through computer simulations. We resolve Bienor's projected ellipse for December 26, 2022, determine a prograde sense of rotation, and confirm an asymmetric rotational light curve. We also retrieve the axes of its triaxial ellipsoid shape as a = (127 $\pm$ 5) km, b = (55 $\pm$ 4) km, and c = (45 $\pm$ 4) km. Moreover, we refine the rotation period to 9.1736 $\pm$ 0.0002 hours and determine a geometric albedo of (6.5 $\pm$ 0.5) %, higher than previously determined by other methods. Finally, by comparing our findings with previous results and simulated rotational light curves, we analyze whether an irregular or contact-binary shape, the presence of an additional element such as a satellite, or significant albedo variations on Bienor's surface, may be present.

When energy is not conserved, imprints of new physics on observable cosmology might not follow the rules of local effective actions. By capturing dissipative and diffusive effects, open effective field theories account for the possibly non-Hamiltonian evolution of cosmological inhomogeneities interacting with an unspecified environment. In this proceeding, we briefly discuss recent progress made towards their implementation in primordial cosmology. Our approach recovers the usual effective field theory of inflation in a certain limit and extends it to account for local dissipation and noises. Non-Gaussianities are generated that peak in the equilateral configuration for large dissipation and in the folded configurations for small dissipation. The construction provides an embedding for local dissipative models of inflation and a framework to study quantum information aspects of the inflationary models., Comment: 8 pages, 2 figures, contribution to the 2024 Cosmology session of the 58th Rencontres de Moriond, based on arXiv:2404.15416

Context. Jupiter-family comets (JFCs), which originate from the Kuiper belt and scattered disk, exhibit low-inclination and chaotic trajectories due to close encounters with Jupiter. Despite their typically short incursions into the inner solar system, a notable number of them are on Earth-crossing orbits, with fireball networks detecting many objects on ``JFC-like'' (2 < TJ < 3) orbits. Aims. This investigation aims to examine the orbital dynamics of JFCs and comet-like fireballs over 10,000 yr timescales, focusing on the trajectories and stability of these objects in the context of gravitational interactions within the solar system. Methods. We employed an extensive fireball dataset from the Desert Fireball Network (DFN), European Fireball Network (EFN), Fireball Recovery and InterPlanetary Observation Network (FRIPON), and Meteorite Observation and Recovery Project (MORP), alongside telescopically observed cometary ephemeris from the NASA HORIZONS database. The study integrates 646 fireball orbits with 661 JFC orbits for a comparative analysis of their orbital stability and evolution. Results. The analysis confirms frequent Jupiter encounters among most JFCs, inducing chaotic orbital behavior with limited predictability and short Lyapunov lifetimes (about 120 years), underscoring Jupiter`s significant dynamical influence. In contrast, ``JFC-like'' meteoroids detected by fireball networks largely exhibit dynamics divergent from genuine JFCs, with 79-92% on ``JFC-like'' orbits shown not to be prone to frequent Jupiter encounters; in particular, only 1-5% of all fireballs detected by the four networks exhibit dynamics similar to that of actual JFCs. In addition, 22% (16 of 72) of near-Earth JFCs are on highly stable orbits, suggesting a potential main belt origin for some of the bodies., Comment: Accepted in Astronomy & Astrophysics (31 pages, 21 figures)

Modern AI systems such as self-driving cars and game-playing agents achieve superhuman performance, but often lack human-like features such as generalization, interpretability and human inter-operability. Inspired by the rich interactions between language and decision-making in humans, we introduce Policy Learning with a Language Bottleneck (PLLB), a framework enabling AI agents to generate linguistic rules that capture the strategies underlying their most rewarding behaviors. PLLB alternates between a rule generation step guided by language models, and an update step where agents learn new policies guided by rules. In a two-player communication game, a maze solving task, and two image reconstruction tasks, we show that PLLB agents are not only able to learn more interpretable and generalizable behaviors, but can also share the learned rules with human users, enabling more effective human-AI coordination., Comment: 18 pages, 13 figures

In our quest to understand the generation of cosmological perturbations, we face two serious obstacles: we do not have direct information about the environment experienced by primordial perturbations during inflation, and our observables are practically limited to correlators of massless fields, heavier fields and derivatives decaying exponentially in the number of e-foldings. The flexible and general framework of open systems has been developed precisely to face similar challenges. Building on previous work, we develop a Schwinger-Keldysh path integral description for an open effective field theory of inflation, describing the possibly dissipative and non-unitary evolution of the Goldstone boson of time translations interacting with an unspecified environment, under the key assumption of locality in space and time. Working in the decoupling limit, we study the linear and interacting theory in de Sitter and derive predictions for the power spectrum and bispectrum that depend on a finite number of effective couplings organised in a derivative expansion. The smoking gun of interactions with the environment is an enhanced but finite bispectrum close to the folded kinematical limit. We demonstrate the generality of our approach by matching our open effective theory to an explicit model. Our construction provides a standard model to simultaneously study phenomenological predictions as well as quantum information aspects of the inflationary dynamics., Comment: 56 pages without appendices (81 pages in total), 12 figures, 2 tables

For rockfall hazard assessment on areas more than several km2 in size, the quantification of runout probability is usually done empirically. Classical methods use statistical distributions of reach or energy angles derived from rockfall databases. However, other topographic descriptors can be derived from the topographic profiles along the rockfall path. Using a database of more than 4,000 profiles of rockfall paths, we determine which topographic descriptors are most appropriate for reach probability estimation, by comparing their statistical distributions for rockfall stopping points, and for points the rocks overtook. We show that the curvilinear length of the propagation path, and the area under the propagation path can improve propagation estimations, especially when they are computed only along the final portion of the propagation path. This is illustrated by comparing an experimental distribution of rockfall stopping points to the estimated distributions.

In recent research, contrastive learning has proven to be a highly effective method for representation learning and is widely used for dense retrieval. However, we identify that relying solely on contrastive learning can lead to suboptimal retrieval performance. On the other hand, despite many retrieval datasets supporting various learning objectives beyond contrastive learning, combining them efficiently in multi-task learning scenarios can be challenging. In this paper, we introduce M3, an advanced recursive Multi-hop dense sentence retrieval system built upon a novel Multi-task Mixed-objective approach for dense text representation learning, addressing the aforementioned challenges. Our approach yields state-of-the-art performance on a large-scale open-domain fact verification benchmark dataset, FEVER. Code and data are available at: https://github.com/TonyBY/M3, Comment: Accepted by LREC-COLING 2024

We compute the rate with which unobserved fields decohere other fields to which they couple, both in flat space and in de Sitter space, for spectator scalar fields prepared in their standard adiabatic vacuum. The process is very efficient in de Sitter space once the modes in question pass outside the Hubble scale, displaying the tell-tale phenomenon of secular growth that indicates the breakdown of perturbative methods on a time scale parameterically long compared with the Hubble time. We show how to match the perturbative evolution valid at early times onto a late-time Lindblad evolution whose domain of validity extends to much later times, thereby allowing a reliable resummation of the perturbative result beyond the perturbative regime. Super-Hubble modes turn out to be dominantly decohered by unobserved modes that are themselves also super-Hubble. Although our calculation is done for spectator fields, if applied to curvature perturbations during inflation our observations here could close a potential loophole in recent calculations of the late-time purity of the observable primordial fluctuations., Comment: 37 pages without appendices (59 pages in total), 7 figures; matches published version in JCAP

Nonlinear photoluminescence (N-PL) is a broadband photon emission arising from non-equilibrium electron distribution generated at the surface of metallic nanostructures by an ultrafast pulsed laser illumination. N-PL is sensitive to surface morphology, local electromagnetic field strength, and electronic band structure making it relevant to probe optically excited nanoscale plasmonic systems. It also has been key to access the complex multiscale time dynamics ruling electron thermalization. Here, we show that the surface plasmons mediated N-PL emitted by a gold nanowire can be modified by an electrical architecture featuring a nanogap. Upon voltage activation, we observe that N-PL becomes dependent to the electrical transport dynamics and can thus be locally modulated. This finding brings an electrical leverage to externally control the photoluminescence generated from metal nanostructures, and constitutes an asset for the development of emerging nanoscale interface devices managing photons and electrons.

Integrated quantum photonic circuits require the efficient coupling of photon sources to photonic waveguides. Hybrid plasmonic/photonic platforms are a promising approach, taking advantage of both plasmon modal confinement for efficient coupling to a nearby emitter and photonic circuitry for optical data transfer and processing. In this work, we established directional quantum dot (QD) emission coupling to a planar TiO$_2$ waveguide assisted by a Yagi-Uda antenna. Antenna on waveguide is first designed by scaling radio frequency dimensions to nano-optics, taking into account the hybrid plasmonic/photonic platform. Design is then optimized by full numerical simulations. We fabricate the antenna on a TiO$_2$ planar waveguide and deposit a few QDs close to the Yagi-Uda antenna. The optical characterization shows clear directional coupling originating from antenna effect. We estimate the coupling efficiency and directivity of the light emitted into the waveguide.

To ensure the proper functioning of the current and future electrical grid, it is necessary for Transmission System Operators (TSOs) to verify that energy providers comply with the grid code and specifications provided by TSOs. A lot of energy production are conntected to the grid through a power electronic inverter. Grid Forming (GFM) and Grid Following (GFL) are the two types of operating modes used to control power electronic converters. The choice of control mode by TSOs to avoid impacting the stability of the grid is crucial, as is the commitment to these choices by energy suppliers. This article proposes a comparison between commonplace machine learning algorithms for converter control mode classification: GFL or GFM. The classification is based on frequency-domain admittance obtained by external measurement methods. Most algorithms are able to classify accurately when the control structure belongs to the training data, but they fail to classify modified control structures with the exception of the random forest algorithm.

We investigate quantum decoherence in a class of models which interpolates between expanding (inflation) and contracting (ekpyrosis) scenarios. For the cases which result in a scale-invariant power spectrum, we find that ekpyrotic universes lead to complete decoherence of the curvature perturbation before the bounce. This is in stark contrast to the inflationary case, where recoherence has been previously observed in some situations. Although the purity can be computed for couplings of all sizes, we also study the purity perturbatively and observe that late-time (secular growth) breakdown of perturbation theory often occurs in these cases. Instead, we establish a simple yet powerful late-time purity resummation which captures the exact evolution to a remarkable level, while maintaining analytical control. We conclude that the cosmological background plays a crucial role in the decoupling of the heavy fields during inflation and alternatives., Comment: 27 pages + appendices, 10 figures

FRIPON is an efficient ground-based network for the detection and characterization of fireballs, which was initiated in France in 2016 with over one hundred cameras and which has been very successfully extended to Europe and Canada with one hundred more stations. After seven successful years of operation in the northern hemisphere, it seems necessary to extend this network towards the southern hemisphere - where the lack of detection is evident - to obtain an exhaustive view of fireball activity. The task of extending the network to any region outside the northern hemisphere presents the challenge of a new installation process, where the recommended and tested version of the several sub-systems that compose a station had to be replaced due to regional availability and compatibility considerations, as well as due to constant software and hardware obsolescence and updates. In Chile, we have a unique geography, with a vast extension in latitude, as well as desert regions, which have generated the need to evaluate the scientific and technical performance of the network under special conditions, prioritizing the optimization of a set of factors related to the deployment process, as well as the feasible and achievable versions of the required components, the geographical location of the stations, and their respective operational, maintenance, safety, and communication conditions. In this talk, we will present the current status of this effort, including a brief report on the obstacles and difficulties encountered and how we have solved them, the current operational status of the network in Northern Chile, as well as the challenges and prospects for the densification of the network over South America., Comment: 4 pages, 2 figures, International Meteor Conference proceedings

The sensitivities of light Dark Matter (DM) particle searches with cryogenic detectors are mostly limited by large backgrounds of events that do not produce ionization signal. The CRYOSEL project develops a new technique where this background in a germanium cryogenic detector is rejected by using the signals from a Superconducting Single Electron Device (SSED) sensor designed to detect the phonons emitted through the Neganov-Trofimov-Luke effect by the e$^-$h$^+$ pairs as they drift in a close-by very high-field region. A tag on signals from this device should suppress the heat-only background. The measurement of the response to IR laser pulses of the first CRYOSEL prototype show the relevance of such sensor technology., Comment: 9 pages, 3 figures, LTD20

A lot of recent machine learning research papers have ``open-ended learning'' in their title. But very few of them attempt to define what they mean when using the term. Even worse, when looking more closely there seems to be no consensus on what distinguishes open-ended learning from related concepts such as continual learning, lifelong learning or autotelic learning. In this paper, we contribute to fixing this situation. After illustrating the genealogy of the concept and more recent perspectives about what it truly means, we outline that open-ended learning is generally conceived as a composite notion encompassing a set of diverse properties. In contrast with previous approaches, we propose to isolate a key elementary property of open-ended processes, which is to produce elements from time to time (e.g., observations, options, reward functions, and goals), over an infinite horizon, that are considered novel from an observer's perspective. From there, we build the notion of open-ended learning problems and focus in particular on the subset of open-ended goal-conditioned reinforcement learning problems in which agents can learn a growing repertoire of goal-driven skills. Finally, we highlight the work that remains to be performed to fill the gap between our elementary definition and the more involved notions of open-ended learning that developmental AI researchers may have in mind.

The ability to invent novel and interesting problems is a remarkable feature of human intelligence that drives innovation, art, and science. We propose a method that aims to automate this process by harnessing the power of state-of-the-art generative models to produce a diversity of challenging yet solvable problems, here in the context of Python programming puzzles. Inspired by the intrinsically motivated literature, Autotelic CodE Search (ACES) jointly optimizes for the diversity and difficulty of generated problems. We represent problems in a space of LLM-generated semantic descriptors describing the programming skills required to solve them (e.g. string manipulation, dynamic programming, etc.) and measure their difficulty empirically as a linearly decreasing function of the success rate of Llama-3-70B, a state-of-the-art LLM problem solver. ACES iteratively prompts a large language model to generate difficult problems achieving a diversity of target semantic descriptors (goal-directed exploration) using previously generated problems as in-context examples. ACES generates problems that are more diverse and more challenging than problems produced by baseline methods and three times more challenging than problems found in existing Python programming benchmarks on average across 11 state-of-the-art code LLMs.

Every population of small bodies in the Solar system contains a sizable fraction of multiple systems. Among these, the Jupiter Trojans have the lowest number of known binary systems and the least characterized. We aim at characterizing the reported binary system (17365) Thymbraeus, one of the only seven multiple systems known among Jupiter Trojans. We conducted light curves observing campaigns in 2013, 2015, and 2021 with ground-based telescopes. We model these lightcurves using dumbbell equilibrium figures. We show that Thymbraeus is unlikely a binary system. Its light curves are fully consistent with a bilobated shape: a dumbbell equilibrium figure. We determine a low density of 830 +/- 50 kg.m-3 , consistent with the reported density of other Jupiter Trojan asteroids and small Kuiper-belt objects. The angular velocity of Thymbraeus is close to fission. If separated, its components would become a similarly-sized double asteroid such as the other Jupiter Trojan (617) Patroclus., Comment: accepted for publication in A&A

Noise generated by wind turbines is significantly impacted by its propagation in the atmosphere. Hence, for annoyance issues an accurate prediction of sound propagation is critical to determine noise levels around wind turbines. This study presents a method to predict wind turbine sound propagation based on linearized Euler equations. We compare this approach to the parabolic equation method, which is widely used since it captures the influence of atmospheric refraction, ground reflection, and sound scattering at a low computational cost. Using the linearized Euler equations is more computationally demanding but can reproduce more physical effects as fewer assumptions are made. An additional benefit of the linearized Euler equations is that they provide a time-domain solution. To compare both approaches, we simulate sound propagation in two distinct scenarios. In the first scenario, a wind turbine is situated on flat terrain; in the second, a turbine is situated on a hilltop. The results show that both methods provide similar noise predictions in the two scenarios. We find that while some differences in the propagation results are observed in the second case, the final predictions for a broadband extended source are similar between the two methods.

Explanations accompanied by a recommendation can assist users in understanding the decision made by recommendation systems, which in turn increases a user's confidence and trust in the system. Recently, research has focused on generating natural language explanations in a human-readable format. Thus far, the proposed approaches leverage item reviews written by users, which are often subjective, sparse in language, and unable to account for new items that have not been purchased or reviewed before. Instead, we aim to generate fact-grounded recommendation explanations that are objectively described with item features while implicitly considering a user's preferences, based on the user's purchase history. To achieve this, we propose a knowledge graph (KG) approach to natural language explainable recommendation. Our approach draws on user-item features through a novel collaborative filtering-based KG representation to produce fact-grounded, personalized explanations, while jointly learning user-item representations for recommendation scoring. Experimental results show that our approach consistently outperforms previous state-of-the-art models on natural language explainable recommendation.

After the discovery of rings around the largest known Centaur object, (10199) Chariklo, we carried out observation campaigns of stellar occultations produced by the second-largest known Centaur object, (2060) Chiron, to better characterize its physical properties and presence of material on its surroundings. We predicted and successfully observed two stellar occultations by Chiron. These observations were used to constrain its size and shape by fitting elliptical limbs with equivalent surface radii in agreement with radiometric measurements. Constraints on the (2060) Chiron shape are reported for the first time. Assuming an equivalent radius of R$_{equiv}$ = 105$^{+6}_{-7}$ km, we obtained a semi-major axis of a = 126 $\pm$ 22 km. Considering Chiron's true rotational light curve amplitude and assuming it has a Jacobi equilibrium shape, we were able to derive a 3D shape with a semi-axis of a = 126 $\pm$ 22 km, b = 109 $\pm$ 19 km, and c = 68 $\pm$ 13 km, implying in a volume-equivalent radius of R$_{vol}$ = 98 $\pm$ 17 km, implying a density of 1119 $\pm$ 4 kg m$^{-3}$. We determined the physical properties of the 2011 secondary events around Chiron, which may then be directly compared with those of Chariklo rings, as the same method was used. Data obtained from SAAO in 2018 do not show unambiguous evidence of the proposed rings, mainly due to the large sampling time. Meanwhile, we discarded the possible presence of a permanent ring similar to (10199) Chariklo's C1R in optical depth and extension. Using the first multi-chord stellar occultation by (2060) Chiron and considering it to have a Jacobi equilibrium shape, we derived its 3D shape. New observations of a stellar occultation by (2060) Chiron are needed to further investigate the material's properties around Chiron, such as the occultation predicted for September 10, 2023.

Knowledge Graph (KG)-to-Text Generation has seen recent improvements in generating fluent and informative sentences which describe a given KG. As KGs are widespread across multiple domains and contain important entity-relation information, and as text simplification aims to reduce the complexity of a text while preserving the meaning of the original text, we propose KGSimple, a novel approach to unsupervised text simplification which infuses KG-established techniques in order to construct a simplified KG path and generate a concise text which preserves the original input's meaning. Through an iterative and sampling KG-first approach, our model is capable of simplifying text when starting from a KG by learning to keep important information while harnessing KG-to-text generation to output fluent and descriptive sentences. We evaluate various settings of the KGSimple model on currently-available KG-to-text datasets, demonstrating its effectiveness compared to unsupervised text simplification models which start with a given complex text. Our code is available on GitHub., Comment: Accepted as a Main Conference Long Paper at CIKM 2023

Recent works in neural knowledge graph inference attempt to combine logic rules with knowledge graph embeddings to benefit from prior knowledge. However, they usually cannot avoid rule grounding, and injecting a diverse set of rules has still not been thoroughly explored. In this work, we propose InjEx, a mechanism to inject multiple types of rules through simple constraints, which capture definite Horn rules. To start, we theoretically prove that InjEx can inject such rules. Next, to demonstrate that InjEx infuses interpretable prior knowledge into the embedding space, we evaluate InjEx on both the knowledge graph completion (KGC) and few-shot knowledge graph completion (FKGC) settings. Our experimental results reveal that InjEx outperforms both baseline KGC models as well as specialized few-shot models while maintaining its scalability and efficiency.

Check-worthy claim detection aims at providing plausible misinformation to downstream fact-checking systems or human experts to check. This is a crucial step toward accelerating the fact-checking process. Many efforts have been put into how to identify check-worthy claims from a small scale of pre-collected claims, but how to efficiently detect check-worthy claims directly from a large-scale information source, such as Twitter, remains underexplored. To fill this gap, we introduce MythQA, a new multi-answer open-domain question answering(QA) task that involves contradictory stance mining for query-based large-scale check-worthy claim detection. The idea behind this is that contradictory claims are a strong indicator of misinformation that merits scrutiny by the appropriate authorities. To study this task, we construct TweetMythQA, an evaluation dataset containing 522 factoid multi-answer questions based on controversial topics. Each question is annotated with multiple answers. Moreover, we collect relevant tweets for each distinct answer, then classify them into three categories: "Supporting", "Refuting", and "Neutral". In total, we annotated 5.3K tweets. Contradictory evidence is collected for all answers in the dataset. Finally, we present a baseline system for MythQA and evaluate existing NLP models for each system component using the TweetMythQA dataset. We provide initial benchmarks and identify key challenges for future models to improve upon. Code and data are available at: https://github.com/TonyBY/Myth-QA, Comment: Accepted by SIGIR 2023

Large Language Models (LLMs) are often misleadingly recognized as having a personality or a set of values. We argue that an LLM can be seen as a superposition of perspectives with different values and personality traits. LLMs exhibit context-dependent values and personality traits that change based on the induced perspective (as opposed to humans, who tend to have more coherent values and personality traits across contexts). We introduce the concept of perspective controllability, which refers to a model's affordance to adopt various perspectives with differing values and personality traits. In our experiments, we use questionnaires from psychology (PVQ, VSM, IPIP) to study how exhibited values and personality traits change based on different perspectives. Through qualitative experiments, we show that LLMs express different values when those are (implicitly or explicitly) implied in the prompt, and that LLMs express different values even when those are not obviously implied (demonstrating their context-dependent nature). We then conduct quantitative experiments to study the controllability of different models (GPT-4, GPT-3.5, OpenAssistant, StableVicuna, StableLM), the effectiveness of various methods for inducing perspectives, and the smoothness of the models' drivability. We conclude by examining the broader implications of our work and outline a variety of associated scientific questions. The project website is available at https://sites.google.com/view/llm-superpositions ., Comment: Preprint

The future Ricochet experiment aims to search for new physics in the electroweak sector by measuring the Coherent Elastic Neutrino-Nucleus Scattering process from reactor antineutrinos with high precision down to the sub-100 eV nuclear recoil energy range. While the Ricochet collaboration is currently building the experimental setup at the reactor site, it is also finalizing the cryogenic detector arrays that will be integrated into the cryostat at the Institut Laue Langevin in early 2024. In this paper, we report on recent progress from the Ge cryogenic detector technology, called the CryoCube. More specifically, we present the first demonstration of a 30~eVee (electron equivalent) baseline ionization resolution (RMS) achieved with an early design of the detector assembly and its dedicated High Electron Mobility Transistor (HEMT) based front-end electronics. This represents an order of magnitude improvement over the best ionization resolutions obtained on similar heat-and-ionization germanium cryogenic detectors from the EDELWEISS and SuperCDMS dark matter experiments, and a factor of three improvement compared to the first fully-cryogenic HEMT-based preamplifier coupled to a CDMS-II germanium detector. Additionally, we discuss the implications of these results in the context of the future Ricochet experiment and its expected background mitigation performance., Comment: 10 pages, 5 figures, 1 table

Humans learn to master open-ended repertoires of skills by imagining and practicing their own goals. This autotelic learning process, literally the pursuit of self-generated (auto) goals (telos), becomes more and more open-ended as the goals become more diverse, abstract and creative. The resulting exploration of the space of possible skills is supported by an inter-individual exploration: goal representations are culturally evolved and transmitted across individuals, in particular using language. Current artificial agents mostly rely on predefined goal representations corresponding to goal spaces that are either bounded (e.g. list of instructions), or unbounded (e.g. the space of possible visual inputs) but are rarely endowed with the ability to reshape their goal representations, to form new abstractions or to imagine creative goals. In this paper, we introduce a language model augmented autotelic agent (LMA3) that leverages a pretrained language model (LM) to support the representation, generation and learning of diverse, abstract, human-relevant goals. The LM is used as an imperfect model of human cultural transmission; an attempt to capture aspects of humans' common-sense, intuitive physics and overall interests. Specifically, it supports three key components of the autotelic architecture: 1)~a relabeler that describes the goals achieved in the agent's trajectories, 2)~a goal generator that suggests new high-level goals along with their decomposition into subgoals the agent already masters, and 3)~reward functions for each of these goals. Without relying on any hand-coded goal representations, reward functions or curriculum, we show that LMA3 agents learn to master a large diversity of skills in a task-agnostic text-based environment.

Decentralized partially observable Markov decision processes (Dec-POMDPs) formalize the problem of designing individual controllers for a group of collaborative agents under stochastic dynamics and partial observability. Seeking a global optimum is difficult (NEXP complete), but seeking a Nash equilibrium -- each agent policy being a best response to the other agents -- is more accessible, and allowed addressing infinite-horizon problems with solutions in the form of finite state controllers. In this paper, we show that this approach can be adapted to cases where only a generative model (a simulator) of the Dec-POMDP is available. This requires relying on a simulation-based POMDP solver to construct an agent's FSC node by node. A related process is used to heuristically derive initial FSCs. Experiment with benchmarks shows that MC-JESP is competitive with exisiting Dec-POMDP solvers, even better than many offline methods using explicit models., Comment: Accepted to UAI 2023, preliminary version

Coherent elastic neutrino-nucleus scattering (CE$\nu$NS) offers valuable sensitivity to physics beyond the Standard Model. The Ricochet experiment will use cryogenic solid-state detectors to perform a precision measurement of the CE$\nu$NS spectrum induced by the high neutrino flux from the Institut Laue-Langevin nuclear reactor. The experiment will employ an array of detectors, each with a mass of $\sim$30 g and a targeted energy threshold of 50 eV. Nine of these detectors (the "Q-Array") will be based on a novel Transition-Edge Sensor (TES) readout style, in which the TES devices are thermally coupled to the absorber using a gold wire bond. We present initial characterization of a Q-Array-style detector using a 1 gram silicon absorber, obtaining a baseline root-mean-square resolution of less than 40 eV., Comment: 18 pages, 8 figures