Search

Large language models (LLMs) can store a significant amount of factual knowledge in their parameters. However, their parametric knowledge may conflict with the information provided in the context. Such conflicts can lead to undesirable model behaviour, such as reliance on outdated or incorrect information. In this work, we investigate whether LLMs can identify knowledge conflicts and whether it is possible to know which source of knowledge the model will rely on by analysing the residual stream of the LLM. Through probing tasks, we find that LLMs can internally register the signal of knowledge conflict in the residual stream, which can be accurately detected by probing the intermediate model activations. This allows us to detect conflicts within the residual stream before generating the answers without modifying the input or model parameters. Moreover, we find that the residual stream shows significantly different patterns when the model relies on contextual knowledge versus parametric knowledge to resolve conflicts. This pattern can be employed to estimate the behaviour of LLMs when conflict happens and prevent unexpected answers before producing the answers. Our analysis offers insights into how LLMs internally manage knowledge conflicts and provides a foundation for developing methods to control the knowledge selection processes., Comment: Foundation Model Interventions Workshop @ NeurIPS 2024

Large language models (LLMs) can store a significant amount of factual knowledge in their parameters. However, their parametric knowledge may conflict with the information provided in the context -- this phenomenon, known as \emph{context-memory knowledge conflicts}, can lead to undesirable model behaviour, such as reliance on outdated or incorrect information. Analysing the internal activations of LLMs, we find that they can internally register the signals of knowledge conflict at mid-layers. Such signals allow us to detect whether a knowledge conflict occurs and use \emph{inference-time} intervention strategies to resolve it. In this work, we propose \textsc{SpARE}, a \emph{training-free} representation engineering method that uses pre-trained sparse auto-encoders (SAEs) to control the knowledge selection behaviour of LLMs. \textsc{SpARE} identifies the functional features that control the knowledge selection behaviours and applies them to edit the internal activations of LLMs at inference time. Our experimental results show that \textsc{SpARE} can effectively control the usage of either knowledge source to resolve knowledge conflict in open-domain question-answering tasks, surpassing existing representation engineering methods ($+10\%$) as well as contrastive decoding methods ($+15\%$).

DarkSide-20k (DS-20k) is a dark matter detection experiment under construction at the Laboratori Nazionali del Gran Sasso (LNGS) in Italy. It utilises ~100 t of low radioactivity argon from an underground source (UAr) in its inner detector, with half serving as target in a dual-phase time projection chamber (TPC). The UAr cryogenics system must maintain stable thermodynamic conditions throughout the experiment's lifetime of >10 years. Continuous removal of impurities and radon from the UAr is essential for maximising signal yield and mitigating background. We are developing an efficient and powerful cryogenics system with a gas purification loop with a target circulation rate of 1000 slpm. Central to its design is a condenser operated with liquid nitrogen which is paired with a gas heat exchanger cascade, delivering a combined cooling power of >8 kW. Here we present the design choices in view of the DS-20k requirements, in particular the condenser's working principle and the cooling control, and we show test results obtained with a dedicated benchmarking platform at CERN and LNGS. We find that the thermal efficiency of the recirculation loop, defined in terms of nitrogen consumption per argon flow rate, is 95 % and the pressure in the test cryostat can be maintained within $\pm$(0.1-0.2) mbar. We further detail a 5-day cool-down procedure of the test cryostat, maintaining a cooling rate typically within -2 K/h, as required for the DS-20k inner detector. Additionally, we assess the circuit's flow resistance, and the heat transfer capabilities of two heat exchanger geometries for argon phase change, used to provide gas for recirculation. We conclude by discussing how our findings influence the finalisation of the system design, including necessary modifications to meet requirements and ongoing testing activities., Comment: 45 pages, 24 figures

Recently, foundation models based on Vision Transformers (ViTs) have become widely available. However, their fine-tuning process is highly resource-intensive, and it hinders their adoption in several edge or low-energy applications. To this end, in this paper we introduce an efficient fine-tuning method for ViTs called $\textbf{ALaST}$ ($\textit{Adaptive Layer Selection Fine-Tuning for Vision Transformers}$) to speed up the fine-tuning process while reducing computational cost, memory load, and training time. Our approach is based on the observation that not all layers are equally critical during fine-tuning, and their importance varies depending on the current mini-batch. Therefore, at each fine-tuning step, we adaptively estimate the importance of all layers and we assign what we call ``compute budgets'' accordingly. Layers that were allocated lower budgets are either trained with a reduced number of input tokens or kept frozen. Freezing a layer reduces the computational cost and memory usage by preventing updates to its weights, while discarding tokens removes redundant data, speeding up processing and reducing memory requirements. We show that this adaptive compute allocation enables a nearly-optimal schedule for distributing computational resources across layers, resulting in substantial reductions in training time (up to 1.5x), FLOPs (up to 2x), and memory load (up to 2x) compared to traditional full fine-tuning approaches. Additionally, it can be successfully combined with other parameter-efficient fine-tuning methods, such as LoRA.

The experiments at the Large Hadron Collider at CERN generate vast amounts of complex data from high-energy particle collisions. This data presents significant challenges due to its volume and complex reconstruction, necessitating the use of advanced analysis techniques for analysis. Recent advancements in deep learning, particularly Graph Neural Networks, have shown promising results in addressing the challenges but remain computationally expensive. The study presented in this paper uses a simulated particle collision dataset to integrate influence analysis inside the graph classification pipeline aiming at improving the accuracy and efficiency of collision event prediction tasks. By using a Graph Neural Network for initial training, we applied a gradient-based data influence method to identify influential training samples and then we refined the dataset by removing non-contributory elements: the model trained on this new reduced dataset can achieve good performances at a reduced computational cost. The method is completely agnostic to the specific influence method: different influence modalities can be easily integrated into our methodology. Moreover, by analyzing the discarded elements we can provide further insights about the event classification task. The novelty of integrating data attribution techniques together with Graph Neural Networks in high-energy physics tasks can offer a robust solution for managing large-scale data problems, capturing critical patterns, and maximizing accuracy across several high-data demand domains., Comment: 10 pages, 6 figures, 2 tables

The dual-phase liquid argon time projection chamber is presently one of the leading technologies to search for dark matter particles with masses below 10 GeV/c$^2$. This was demonstrated by the DarkSide-50 experiment with approximately 50 kg of low-radioactivity liquid argon as target material. The next generation experiment DarkSide-20k, currently under construction, will use 1,000 times more argon and is expected to start operation in 2027. Based on the DarkSide-50 experience, here we assess the DarkSide-20k sensitivity to models predicting light dark matter particles, including Weakly Interacting Massive Particles (WIMPs) and sub-GeV/c$^2$ particles interacting with electrons in argon atoms. With one year of data, a sensitivity improvement to dark matter interaction cross-sections by at least one order of magnitude with respect to DarkSide-50 is expected for all these models. A sensitivity to WIMP--nucleon interaction cross-sections below $1\times10^{-42}$ cm$^2$ is achievable for WIMP masses above 800 MeV/c$^2$. With 10 years exposure, the neutrino fog can be reached for WIMP masses around 5 GeV/c$^2$., Comment: submitted to Nature Communications

The deployment of large language models (LLMs) is often hindered by the extensive memory requirements of the Key-Value (KV) cache, especially as context lengths increase. Existing approaches to reduce the KV cache size involve either fine-tuning the model to learn a compression strategy or leveraging attention scores to reduce the sequence length. We analyse the attention distributions in decoder-only Transformers-based models and observe that attention allocation patterns stay consistent across most layers. Surprisingly, we find a clear correlation between the $L_2$ and the attention scores over cached KV pairs, where a low $L_2$ of a key embedding usually leads to a high attention score during decoding. This finding indicates that the influence of a KV pair is potentially determined by the key embedding itself before being queried. Based on this observation, we compress the KV cache based on the $L_2$ of key embeddings. Our experimental results show that this simple strategy can reduce the KV cache size by 50% on language modelling and needle-in-a-haystack tasks and 90% on passkey retrieval tasks without losing accuracy. Moreover, without relying on the attention scores, this approach remains compatible with FlashAttention, enabling broader applicability., Comment: This is an extended version of a paper published in the proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP 2024); this version was presented at the 4th NeurIPS Workshop on Efficient Natural Language and Speech Processing (ENLSP-IV)

Maybe not. We identify and analyse errors in the popular Massive Multitask Language Understanding (MMLU) benchmark. Even though MMLU is widely adopted, our analysis demonstrates numerous ground truth errors that obscure the true capabilities of LLMs. For example, we find that 57% of the analysed questions in the Virology subset contain errors. To address this issue, we introduce a comprehensive framework for identifying dataset errors using a novel error taxonomy. Then, we create MMLU-Redux, which is a subset of 3,000 manually re-annotated questions across 30 MMLU subjects. Using MMLU-Redux, we demonstrate significant discrepancies with the model performance metrics that were originally reported. Our results strongly advocate for revising MMLU's error-ridden questions to enhance its future utility and reliability as a benchmark. Therefore, we open up MMLU-Redux for additional annotation https://huggingface.co/datasets/edinburgh-dawg/mmlu-redux.

We present the two-loop mixed strong-electroweak virtual corrections to the charged current Drell-Yan process. The final-state collinear singularities are regularised by the lepton mass. The evaluation of all the relevant Feynman integrals, including those with up to two different internal massive lines, has been worked out relying on semi-analytical techniques, using complex-valued masses. We can provide, at any arbitrary phase-space point, the solution as a power series in the $W$-boson mass, around a reference value. Starting from these expansions, we can prepare a numerical grid for any value of the $W$-boson mass within their radius of convergence in a negligible amount of time., Comment: 26 pages, 7 figures. arXiv admin note: text overlap with arXiv:2201.01754

Experiments aimed at direct searches for WIMP dark matter require highly effective reduction of backgrounds and control of any residual radioactive contamination. In particular, neutrons interacting with atomic nuclei represent an important class of backgrounds due to the expected similarity of a WIMP-nucleon interaction, so that such experiments often feature a dedicated neutron detector surrounding the active target volume. In the context of the development of DarkSide-20k detector at INFN Gran Sasso National Laboratory (LNGS), several R&D projects were conceived and developed for the creation of a new hybrid material rich in both hydrogen and gadolinium nuclei to be employed as an essential element of the neutron detector. Thanks to its very high cross-section for neutron capture, gadolinium is one of the most widely used elements in neutron detectors, while the hydrogen-rich material is instrumental in efficiently moderating the neutrons. In this paper results from one of the R&Ds are presented. In this effort the new hybrid material was obtained as a poly(methyl methacrylate) (PMMA) matrix, loaded with gadolinium oxide in the form of nanoparticles. We describe its realization, including all phases of design, purification, construction, characterization, and determination of mechanical properties of the new material.

In this paper, we propose a novel design for AI-native goal-oriented communications, exploiting transformer neural networks under dynamic inference constraints on bandwidth and computation. Transformers have become the standard architecture for pretraining large-scale vision and text models, and preliminary results have shown promising performance also in deep joint source-channel coding (JSCC). Here, we consider a dynamic model where communication happens over a channel with variable latency and bandwidth constraints. Leveraging recent works on conditional computation, we exploit the structure of the transformer blocks and the multihead attention operator to design a trainable semantic token selection mechanism that learns to select relevant tokens (e.g., image patches) from the input signal. This is done dynamically, on a per-input basis, with a rate that can be chosen as an additional input by the user. We show that our model improves over state-of-the-art token selection mechanisms, exhibiting high accuracy for a wide range of latency and bandwidth constraints, without the need for deploying multiple architectures tailored to each constraint. Last, but not least, the proposed token selection mechanism helps extract powerful semantics that are easy to understand and explain, paving the way for interpretable-by-design models for the next generation of AI-native communication systems., Comment: 5 pages

This article summarizes principles and ideas from the emerging area of applying \textit{conditional computation} methods to the design of neural networks. In particular, we focus on neural networks that can dynamically activate or de-activate parts of their computational graph conditionally on their input. Examples include the dynamic selection of, e.g., input tokens, layers (or sets of layers), and sub-modules inside each layer (e.g., channels in a convolutional filter). We first provide a general formalism to describe these techniques in an uniform way. Then, we introduce three notable implementations of these principles: mixture-of-experts (MoEs) networks, token selection mechanisms, and early-exit neural networks. The paper aims to provide a tutorial-like introduction to this growing field. To this end, we analyze the benefits of these modular designs in terms of efficiency, explainability, and transfer learning, with a focus on emerging applicative areas ranging from automated scientific discovery to semantic communication.

Humans are capable of acquiring new knowledge and transferring learned knowledge into different domains, incurring a small forgetting. The same ability, called Continual Learning, is challenging to achieve when operating with neural networks due to the forgetting affecting past learned tasks when learning new ones. This forgetting can be mitigated by replaying stored samples from past tasks, but a large memory size may be needed for long sequences of tasks; moreover, this could lead to overfitting on saved samples. In this paper, we propose a novel regularisation approach and a novel incremental classifier called, respectively, Margin Dampening and Cascaded Scaling Classifier. The first combines a soft constraint and a knowledge distillation approach to preserve past learned knowledge while allowing the model to learn new patterns effectively. The latter is a gated incremental classifier, helping the model modify past predictions without directly interfering with them. This is achieved by modifying the output of the model with auxiliary scaling functions. We empirically show that our approach performs well on multiple benchmarks against well-established baselines, and we also study each component of our proposal and how the combinations of such components affect the final results., Comment: Previously called "Cascaded Scaling Classifier: class incremental learning with probability scaling ". The official code is available https://github.com/jaryP/CIL-Margin-Dampening-Gated-Classifier

The ATLAS and CMS collaborations at the LHC have recently announced evidence for the rare Higgs boson decay into a $Z$ boson and a photon. We analyze the interference between the process $gg\! \to \! H \! \to \! Z \gamma$ induced by loops of heavy particles, which is by far the dominant contribution to the signal, and the continuum $gg \to Z \gamma$ QCD background process mediated by light quark loops. This interference modifies the event yield, the resonance line-shape and the apparent mass of the Higgs boson. We calculate the radiative corrections to this interference beyond the leading-order approximation in perturbative QCD and find that, while differing numerically from the corresponding effects on the more studied $gg \! \to \! \gamma \gamma$ signal, they are generally rather small. As such, they do not impact significantly the interpretation of the present measurements of the $H \to Z \gamma$ decay mode., Comment: v1: 6 pages, 3 figures; v2: version published in journal

The computational cost of transformer models makes them inefficient in low-latency or low-power applications. While techniques such as quantization or linear attention can reduce the computational load, they may incur a reduction in accuracy. In addition, globally reducing the cost for all inputs may be sub-optimal. We observe that for each layer, the full width of the layer may be needed only for a small subset of tokens inside a batch and that the "effective" width needed to process a token can vary from layer to layer. Motivated by this observation, we introduce the Adaptive Computation Module (ACM), a generic module that dynamically adapts its computational load to match the estimated difficulty of the input on a per-token basis. An ACM consists of a sequence of learners that progressively refine the output of their preceding counterparts. An additional gating mechanism determines the optimal number of learners to execute for each token. We also describe a distillation technique to replace any pre-trained model with an "ACMized" variant. The distillation phase is designed to be highly parallelizable across layers while being simple to plug-and-play into existing networks. Our evaluation of transformer models in computer vision and speech recognition demonstrates that substituting layers with ACMs significantly reduces inference costs without degrading the downstream accuracy for a wide interval of user-defined budgets.

The stability of a dark matter detector on the timescale of a few years is a key requirement due to the large exposure needed to achieve a competitive sensitivity. It is especially crucial to enable the detector to potentially detect any annual event rate modulation, an expected dark matter signature. In this work, we present the performance history of the DarkSide-50 dual-phase argon time projection chamber over its almost three-year low-radioactivity argon run. In particular, we focus on the electroluminescence signal that enables sensitivity to sub-keV energy depositions. The stability of the electroluminescence yield is found to be better than 0.5%. Finally, we show the temporal evolution of the observed event rate around the sub-keV region being consistent to the background prediction., Comment: 13 pages, 5 figures

We compute all helicity amplitudes for the scattering of five partons in two-loop QCD in all the relevant flavor configurations, retaining all contributing color structures. We employ tensor projection to obtain helicity amplitudes in the 't Hooft-Veltman scheme starting from a set of primitive amplitudes. Our analytic results are expressed in terms of massless pentagon functions, and are easy to evaluate numerically. These amplitudes provide important input to investigations of collinear-factorization breaking and to studies of the multi-Regge kinematics regime., Comment: 6 pages, one appendix, 2 tables. Results available at https://www.ph.nat.tum.de/ttpmath/research/published-results/fivepartons/

We describe how the nested soft-collinear subtraction scheme [1] can be used to compute the next-to-next-to-leading order (NNLO) QCD corrections to the production of an arbitrary number of gluonic jets in hadron collisions. We show that the infrared subtraction terms can be combined into recurring structures that in many cases are simple iterations of those terms known from next-to-leading order. The way that these recurring structures are identified and computed is fairly general, and can be applied to any partonic process. As an example, we explicitly demonstrate the cancellation of all singularities in the fully-differential cross section for the $q\bar{q} \to X + N g $ process at NNLO in QCD. The finite remainder of the NNLO QCD contribution, which arises upon cancellation of all $\epsilon$-poles, is expressed via relatively simple formulas, which can be implemented in a numerical code in a straightforward way. Our approach can be extended to describe arbitrary processes at NNLO in QCD; the largest remaining challenge at this point is the combinatorics of quark and gluon collinear limits., Comment: 116 pages, 9 appendices, 1 figure, alphabetic index

In this contribution we present our recent computation of the NNLO QCD corrections to the production of a Higgs boson associated with a top-antitop quark pair. This process is of great importance since it allows for a direct measurement of the top-quark Yukawa coupling, and the inclusion of NNLO corrections is crucial in order to provide theory predictions with an uncertainty competitive with the projected accuracy of the experimental measurements at the end of the high-luminosity phase of the LHC., Comment: Contribution to the 34th Rencontres de Blois on Particle Physics and Cosmology (Blois 2023)

We consider extensions of the soft-gluon effective coupling that generalize the Catani--Marchesini--Webber (CMW) coupling in the context of soft-gluon resummation beyond the next-to-leading logarithmic accuracy. Starting from the probability density of correlated soft emission in d dimensions we introduce a class of soft couplings relevant for resummed QCD calculations of hard-scattering observables. We show that at the conformal point, where the d-dimensional QCD $\beta$ function vanishes, all these effective couplings are equal and they are also equal to the cusp anomalous dimension. We present explicit results in d dimensions for the soft-emission probability density and the soft couplings at the second-order in the QCD coupling $\alpha_s$. In d=4 dimensions we obtain the explicit relation between the soft couplings at ${\cal O}(\alpha_s^3).$ Finally, we study the structure of the soft coupling in the large-$n_F$ limit and we present explicit expressions to all orders in perturbation theory. We also check that, at the conformal point, our large-$n_F$ results agree with the known result of the cusp anomalous dimension., Comment: 24 pages

The direct search for dark matter in the form of weakly interacting massive particles (WIMP) is performed by detecting nuclear recoils (NR) produced in a target material from the WIMP elastic scattering. A promising experimental strategy for direct dark matter search employs argon dual-phase time projection chambers (TPC). One of the advantages of the TPC is the capability to detect both the scintillation and charge signals produced by NRs. Furthermore, the existence of a drift electric field in the TPC breaks the rotational symmetry: the angle between the drift field and the momentum of the recoiling nucleus can potentially affect the charge recombination probability in liquid argon and then the relative balance between the two signal channels. This fact could make the detector sensitive to the directionality of the WIMP-induced signal, enabling unmistakable annual and daily modulation signatures for future searches aiming for discovery. The Recoil Directionality (ReD) experiment was designed to probe for such directional sensitivity. The TPC of ReD was irradiated with neutrons at the INFN Laboratori Nazionali del Sud, and data were taken with 72 keV NRs of known recoil directions. The direction-dependent liquid argon charge recombination model by Cataudella et al. was adopted and a likelihood statistical analysis was performed, which gave no indications of significant dependence of the detector response to the recoil direction. The aspect ratio R of the initial ionization cloud is estimated to be 1.037 +/- 0.027 and the upper limit is R < 1.072 with 90% confidence level, Comment: 20 pages, 10 figures, submitted to Eur. Phys. J. C

Dark matter induced event rate in an Earth-based detector is predicted to show an annual modulation as a result of the Earth's orbital motion around the Sun. We searched for this modulation signature using the ionization signal of the DarkSide-50 liquid argon time projection chamber. No significant signature compatible with dark matter is observed in the electron recoil equivalent energy range above $40~{\rm eV_{ee}}$, the lowest threshold ever achieved in such a search., Comment: 8 pages, 4 figures

The production of a top-antitop quark pair in association with a $W$ boson ($t\bar tW$) is one of the heaviest signatures currently probed at the Large Hadron Collider (LHC). Since the first observation reported in 2015 the corresponding rates have been found to be consistently higher than the Standard Model predictions, which are based on next-to-leading order~(NLO) calculations in the QCD and electroweak (EW) interactions. We present the first next-to-next-to-leading order (NNLO) QCD computation of $t\bar tW$ production at hadron colliders. The calculation is exact, except for the finite part of the two-loop virtual corrections, which is estimated using two different approaches that lead to consistent results within their uncertainties. We combine the newly computed NNLO QCD corrections with the complete NLO QCD+EW results, thus obtaining the most advanced perturbative prediction available to date for the \ttW inclusive cross section. The tension with the latest ATLAS and CMS results remains at the $1\sigma-2\sigma$ level., Comment: 7 pages, 3 figures, 1 table. Minor changes, version published on PRL

Stigma toward people who use substances (PWUS) is a leading barrier to seeking treatment.Further, those in treatment are more likely to drop out if they experience higher levels of stigmatization. While related concepts of hate speech and toxicity, including those targeted toward vulnerable populations, have been the focus of automatic content moderation research, stigma and, in particular, people who use substances have not. This paper explores stigma toward PWUS using a data set of roughly 5,000 public Reddit posts. We performed a crowd-sourced annotation task where workers are asked to annotate each post for the presence of stigma toward PWUS and answer a series of questions related to their experiences with substance use. Results show that workers who use substances or know someone with a substance use disorder are more likely to rate a post as stigmatizing. Building on this, we use a supervised machine learning framework that centers workers with lived substance use experience to label each Reddit post as stigmatizing. Modeling person-level demographics in addition to comment-level language results in a classification accuracy (as measured by AUC) of 0.69 -- a 17% increase over modeling language alone. Finally, we explore the linguist cues which distinguish stigmatizing content: PWUS substances and those who don't agree that language around othering ("people", "they") and terms like "addict" are stigmatizing, while PWUS (as opposed to those who do not) find discussions around specific substances more stigmatizing. Our findings offer insights into the nature of perceived stigma in substance use. Additionally, these results further establish the subjective nature of such machine learning tasks, highlighting the need for understanding their social contexts., Comment: Accepted for publication the 2024 International AAAI Conference on Web and Social Media (ICWSM)

We present a novel approach for the search of dark matter in the DarkSide-50 experiment, relying on Bayesian Networks. This method incorporates the detector response model into the likelihood function, explicitly maintaining the connection with the quantity of interest. No assumptions about the linearity of the problem or the shape of the probability distribution functions are required, and there is no need to morph signal and background spectra as a function of nuisance parameters. By expressing the problem in terms of Bayesian Networks, we have developed an inference algorithm based on a Markov Chain Monte Carlo to calculate the posterior probability. A clever description of the detector response model in terms of parametric matrices allows us to study the impact of systematic variations of any parameter on the final results. Our approach not only provides the desired information on the parameter of interest, but also potential constraints on the response model. Our results are consistent with recent published analyses and further refine the parameters of the detector response model., Comment: 24 pages, 12 figures, 1 table

We consider QCD radiative corrections to the production of a heavy-quark pair in hadronic collisions. We present the computation of the soft-parton contributions at low transverse momentum of the heavy-quark pair up to second order in the QCD coupling $\alpha_s$. These results complete the evaluation at the next-to-next-to-leading order (NNLO) of the transverse-momentum resummation formula for this process. Moreover, they give all the ingredients that are needed for the NNLO implementation of the qT subtraction formalism for heavy-quark production. We discuss the details of the computation and we provide a code that can be used to obtain the relevant results in numerical form., Comment: 67 pages, 3 figures. Numerical code for the evaluation of the results included in the submission

The Lunar Orbital Platform-Gateway (LOP-Gateway, or simply Gateway) is a crewed platform that will be assembled and operated in the vicinity of the Moon by NASA and international partner organizations, including ESA, starting from the mid-2020s. It will offer new opportunities for fundamental and applied scientific research. The Moon is a unique location to study the deep space plasma environment. Moreover, the lunar surface and the surface-bounded exosphere are interacting with this environment, constituting a complex multi-scale interacting system. This paper examines the opportunities provided by externally mounted payloads on the Gateway in the field of space plasma physics, heliophysics and space weather, but also examines the impact of the space environment on an inhabited platform in the vicinity of the Moon. It then presents the conceptual design of a model payload, required to perform these space plasma measurements and observations. It results that the Gateway is very well-suited for space plasma physics research. It allows a series of scientific objectives with a multidisciplinary dimension to be addressed.

In this paper we consider signal-background interference effects in Higgs-mediated diphoton production at the LHC. After reviewing earlier works that show how to use these effects to constrain the Higgs boson total decay width, we provide predictions beyond NLO accuracy for the interference and related observables, and study the impact of QCD radiative corrections on the Higgs width determination. In particular, we use the so-called soft-virtual approximation to estimate interference effects at NNLO in QCD. The inclusion of these effects reduce the NNLO prediction for the total Higgs cross-section in the diphoton channel by about 1.7%. We study in detail the impact of QCD corrections on the Higgs-boson line-shape and its implications for the Higgs width extraction. Assuming an experimental resolution of about 150~MeV on interference-induced modifications of the Higgs-boson line-shape, our NNLO analysis shows that one could constrain the Higgs-boson total width to about 10-20 times its Standard Model value., Comment: 16 pages, 6 figures, 3 tables

We present the first calculation for the hadroproduction of a $W$ boson in association with a massive bottom ($b$) quark-antiquark pair at next-to-next-to-leading order (NNLO) in QCD perturbation theory. We exploit the hierarchy between the $b$ quark mass and the characteristic energy scale of the process to obtain a reliable analytic expression for the two-loop virtual amplitude with three massive legs, starting from the corresponding result available for massless bottom quarks. The use of massive $b$ quarks avoids the ambiguities associated with the correct flavour assignment in massless calculations, paving the way to a more realistic comparison with experimental data. We present phenomenological results considering proton-proton collisions at centre-of-mass energy $\sqrt{s}=13.6$ TeV for inclusive $Wb \bar b$ production and within a fiducial region relevant for the associated production of a $W$ boson and a Higgs boson decaying into a $b \bar b$ pair, for which $Wb \bar b$ production represents one of the most relevant backgrounds. We find that the NNLO corrections are substantial and that their inclusion is mandatory to obtain reliable predictions., Comment: 8 pages, 2 figures, 2 tables. Final version published in PRD

We review the description of tunnelling phenomena in the semi-classical approximation in ordinary quantum mechanics and in quantum field theory. In particular, we describe in detail the calculation, up to the first quantum corrections, of the decay probability per unit time of a metastable ground state. We apply the relevant formalism to the case of the standard model of electroweak interactions, whose ground state is metastable for sufficiently large values of the top quark mass. Finally, we discuss the impact of gravitational interactions on the calculation of the tunnelling rate., Comment: 106 pages, 12 Figures. Invited review for Journal of Physics G. Minor corrections and references added. Matches published version

The associated production of a Higgs boson with a top-antitop quark pair is a crucial process at the LHC since it allows for a direct measurement of the top-quark Yukawa coupling. We present the computation of the radiative corrections to this process at the next-to-next-to-leading order (NNLO) in QCD perturbation theory. This is the very first computation for a $2 \to 3$ process with massive coloured particles at this perturbative order. We develop a soft Higgs boson approximation for loop amplitudes, which enables us to reliably quantify the impact of the yet unknown two-loop contribution. At the centre-of-mass energy $\sqrt{s}=13$ TeV the NNLO corrections increase the next-to-leading order result for the total cross section by about 4% and lead to a significant reduction of perturbative uncertainties., Comment: 8 pages, 2 tables, 1 figure. Version published on PRL

Dark matter lighter than 10 GeV/c$^2$ encompasses a promising range of candidates. A conceptual design for a new detector, DarkSide-LowMass, is presented, based on the DarkSide-50 detector and progress toward DarkSide-20k, optimized for a low-threshold electron-counting measurement. Sensitivity to light dark matter is explored for various potential energy thresholds and background rates. These studies show that DarkSide-LowMass can achieve sensitivity to light dark matter down to the solar neutrino floor for GeV-scale masses and significant sensitivity down to 10 MeV/c$^2$ considering the Migdal effect or interactions with electrons. Requirements for optimizing the detector's sensitivity are explored, as are potential sensitivity gains from modeling and mitigating spurious electron backgrounds that may dominate the signal at the lowest energies.

We present the mixed QCD-EW two-loop virtual amplitudes for the neutral current Drell-Yan production, one of the bottlenecks for the complete calculation of the NNLO mixed QCD-EW corrections. We present the computational details and the first steps towards their automation. We describe the evaluation of all the relevant two-loop Feynman integrals using analytical and semi-analytical methods, the subtraction of the universal infrared singularities and present the numerical evaluation of the finite remainder., Comment: Contribution to the proceedings of "Loops and Legs in Quantum Field Theory - LL2022, 25-30 April, 2022, Ettal, Germany

Dark matter elastic scattering off nuclei can result in the excitation and ionization of the recoiling atom through the so-called Migdal effect. The energy deposition from the ionization electron adds to the energy deposited by the recoiling nuclear system and allows for the detection of interactions of sub-GeV/c$^2$ mass dark matter. We present new constraints for sub-GeV/c$^2$ dark matter using the dual-phase liquid argon time projection chamber of the DarkSide-50 experiment with an exposure of (12306 $\pm$ 184) kg d. The analysis is based on the ionization signal alone and significantly enhances the sensitivity of DarkSide-50, enabling sensitivity to dark matter with masses down to 40 MeV/c$^2$. Furthermore, it sets the most stringent upper limit on the spin independent dark matter nucleon cross section for masses below $3.6$ GeV/c$^2$., Comment: 7 pages, 3 figures

We present a search for dark matter particles with sub-GeV/$c^2$ masses whose interactions have final state electrons using the DarkSide-50 experiment's (12306 $\pm$ 184) kg d low-radioactivity liquid argon exposure. By analyzing the ionization signals, we exclude new parameter space for the dark matter-electron cross section $\bar{\sigma}_e$, the axioelectric coupling constant $g_{Ae}$, and the dark photon kinetic mixing parameter $\kappa$. We also set the first dark matter direct-detection constraints on the mixing angle $\left|U_{e4}\right|^2$ for keV sterile neutrinos., Comment: 6 pages, 2 figures

We report on the search for dark matter WIMPs in the mass range below 10 GeV/c$^2$, from the analysis of the entire dataset acquired with a low-radioactivity argon target by the DarkSide-50 experiment at LNGS. The new analysis benefits from more accurate calibration of the detector response, improved background model, and better determination of systematic uncertainties, allowing us to accurately model the background rate and spectra down to 0.06 keV$_{er}$. A 90% C.L. exclusion limit for the spin-independent cross section of 3 GeV/c$^2$ mass WIMP on nucleons is set at 6$\times$10$^{-43}$ cm$^2$, about a factor 10 better than the previous DarkSide-50 limit. This analysis extends the exclusion region for spin-independent dark matter interactions below the current experimental constraints in the $[1.2, 3.6]$ GeV/c$^2$ WIMP mass range., Comment: 11 pages, 12 figures

We present the calculation of QCD radiative corrections to heavy-quark production at NNLO. The calculation is implemented in an updated version of \textsc{Matrix}, which allows us to obtain multidifferential predictions for stable top quarks. We also report on the NNLO calculation for $b{\bar b}$ production and comment on the inclusion of top-quark decays and of the extension of our results to $t\bar tH$., Comment: Contribution to the 2022 QCD session of the 56th Rencontres des Moriond

We present an algorithm to evaluate multiloop Feynman integrals with an arbitrary number of internal massive lines, with the masses being in general complex-valued, and its implementation in the \textsc{Mathematica} package \textsc{SeaSyde}. The implementation solves by series expansions the system of differential equations satisfied by the Master Integrals. At variance with respect to other existing codes, the analytical continuation of the solution is performed in the complex plane associated to each kinematical invariant. We present the results of the evaluation of the Master Integrals relevant for the NNLO QCD-EW corrections to the neutral-current Drell-Yan processes., Comment: 23 pages, 6 figures, 2 tables, Version accepted for publication in CPC

We compute mixed QCD-electroweak corrections to the neutral-current Drell-Yan production of a pair of massless leptons in the high invariant mass region. Our computation is fully differential with respect to the final state particles. At relatively low values of the dilepton invariant mass, $m_{\ell \ell} \sim 200$ GeV, we find unexpectedly large mixed QCD-electroweak corrections at the level of -1%. At higher invariant masses, $m_{\ell \ell} \sim 1$ TeV, we observe that these corrections can be well approximated by the product of QCD and electroweak corrections. Hence, thanks to the well-known Sudakov enhancement of the latter, they increase at large invariant mass and reach e.g. -3% at $m_{\ell \ell} = 3$ TeV. Finally, we note that the inclusion of mixed corrections reduces the theoretical uncertainty related to the choice of electroweak input parameters to below the percent level., Comment: 30 pages, 4 figures, 2 appendices