Search

This document is comprised of a collection of updated preliminary parameters for the key parts of the muon collider. The updated preliminary parameters follow on from the October 2023 Tentative Parameters Report. Particular attention has been given to regions of the facility that are believed to hold greater technical uncertainty in their design and that have a strong impact on the cost and power consumption of the facility. The data is collected from a collaborative spreadsheet and transferred to overleaf.

The accurate modeling of individual movement in cities has significant implications for policy decisions across various sectors. Existing research emphasizes the universality of human mobility, positing that simple models can capture population-level movements. However, population-level accuracy does not guarantee consistent performance across all individuals. By overlooking individual differences, universality laws may accurately describe certain groups while less precisely representing others, resulting in aggregate accuracy from a balance of discrepancies. Using large-scale mobility data, we assess individual-level accuracy of a universal model, the Exploration and Preferential Return (EPR), by examining deviations from expected behavior in two scaling laws - one related to exploration and the other to return patterns. Our findings reveal that, while the model can describe population-wide movement patterns, it displays widespread deviations linked to individuals' behavioral traits, socioeconomic status, and lifestyles, contradicting model assumptions like non-bursty exploration and preferential return. Specifically, individuals poorly represented by the EPR model tend to visit routine locations in sequences, exploring rarely but in a bursty manner when they do. Among socioeconomic factors, income most strongly correlates with significant deviations. Consequently, spatial inhomogeneity emerges in model accuracy, with lower performance concentrated in urbanized, densely populated areas, underscoring policy implications. Our results show that emphasizing population-wide models can propagate socioeconomic inequalities by poorly representing vulnerable population sectors.

(Partial) differential equations (PDEs) are fundamental tools for describing natural phenomena, making their solution crucial in science and engineering. While traditional methods, such as the finite element method, provide reliable solutions, their accuracy is often tied to the use of computationally intensive fine meshes. Moreover, they do not naturally account for measurements or prior solutions, and any change in the problem parameters requires results to be fully recomputed. Neural network-based approaches, such as physics-informed neural networks and neural operators, offer a mesh-free alternative by directly fitting those models to the PDE solution. They can also integrate prior knowledge and tackle entire families of PDEs by simply aggregating additional training losses. Nevertheless, they are highly sensitive to hyperparameters such as collocation points and the weights associated with each loss. This paper addresses these challenges by developing a science-constrained learning (SCL) framework. It demonstrates that finding a (weak) solution of a PDE is equivalent to solving a constrained learning problem with worst-case losses. This explains the limitations of previous methods that minimize the expected value of aggregated losses. SCL also organically integrates structural constraints (e.g., invariances) and (partial) measurements or known solutions. The resulting constrained learning problems can be tackled using a practical algorithm that yields accurate solutions across a variety of PDEs, neural network architectures, and prior knowledge levels without extensive hyperparameter tuning and sometimes even at a lower computational cost.

The solar magnetized corona is responsible for various manifestations with a space weather impact, such as flares, coronal mass ejections (CMEs) and, naturally, the solar wind. Modeling the corona's dynamics and evolution is therefore critical for improving our ability to predict space weather In this work, we demonstrate that generative deep learning methods, such as Denoising Diffusion Probabilistic Models (DDPM), can be successfully applied to simulate future evolutions of the corona as observed in Extreme Ultraviolet (EUV) wavelengths. Our model takes a 12-hour video of an Active Region (AR) as input and simulate the potential evolution of the AR over the subsequent 12 hours, with a time-resolution of two hours. We propose a light UNet backbone architecture adapted to our problem by adding 1D temporal convolutions after each classical 2D spatial ones, and spatio-temporal attention in the bottleneck part. The model not only produce visually realistic outputs but also captures the inherent stochasticity of the system's evolution. Notably, the simulations enable the generation of reliable confidence intervals for key predictive metrics such as the EUV peak flux and fluence of the ARs, paving the way for probabilistic and interpretable space weather forecasting. Future studies will focus on shorter forecasting horizons with increased spatial and temporal resolution, aiming at reducing the uncertainty of the simulations and providing practical applications for space weather forecasting. The code used for this study is available at the following link: https://github.com/gfrancisco20/video_diffusion

In the current era of social media and generative AI, an ability to automatically assess the credibility of online social media content is of tremendous importance. Credibility assessment is fundamentally based on aggregating credibility signals, which refer to small units of information, such as content factuality, bias, or a presence of persuasion techniques, into an overall credibility score. Credibility signals provide a more granular, more easily explainable and widely utilizable information in contrast to currently predominant fake news detection, which utilizes various (mostly latent) features. A growing body of research on automatic credibility assessment and detection of credibility signals can be characterized as highly fragmented and lacking mutual interconnections. This issue is even more prominent due to a lack of an up-to-date overview of research works on automatic credibility assessment. In this survey, we provide such systematic and comprehensive literature review of 175 research papers while focusing on textual credibility signals and Natural Language Processing (NLP), which undergoes a significant advancement due to Large Language Models (LLMs). While positioning the NLP research into the context of other multidisciplinary research works, we tackle with approaches for credibility assessment as well as with 9 categories of credibility signals (we provide a thorough analysis for 3 of them, namely: 1) factuality, subjectivity and bias, 2) persuasion techniques and logical fallacies, and 3) claims and veracity). Following the description of the existing methods, datasets and tools, we identify future challenges and opportunities, while paying a specific attention to recent rapid development of generative AI.

Solar flare forecasting mainly relies on photospheric magnetograms and associated physical features to predict forthcoming flares. However, it is believed that flare initiation mechanisms often originate in the chromosphere and the lower corona. In this study, we employ deep learning as a purely data-driven approach to compare the predictive capabilities of chromospheric and coronal UV and EUV emissions across different wavelengths with those of photospheric line-of-sight magnetograms. Our findings indicate that individual EUV wavelengths can provide discriminatory power comparable or better to that of line-of-sight magnetograms. Moreover, we identify simple multimodal neural network architectures that consistently outperform single-input models, showing complementarity between the flare precursors that can be extracted from the distinct layers of the solar atmosphere. To mitigate potential biases from known misattributions in Active Region flare catalogs, our models are trained and evaluated using full-disk images and a comprehensive flare event catalog at the full-disk level. We introduce a deep-learning architecture suited for extracting temporal features from full-disk videos.

Motor changes are early signs of neurodegenerative diseases (NDs) such as Parkinson's disease (PD) and Alzheimer's disease (AD), but are often difficult to detect, especially in the early stages. In this work, we examine the behavior of a wide array of explainable metrics extracted from the handwriting signals of 113 subjects performing multiple tasks on a digital tablet. The aim is to measure their effectiveness in characterizing and assessing multiple NDs, including AD and PD. To this end, task-agnostic and task-specific metrics are extracted from 14 distinct tasks. Subsequently, through statistical analysis and a series of classification experiments, we investigate which metrics provide greater discriminative power between NDs and healthy controls and among different NDs. Preliminary results indicate that the various tasks at hand can all be effectively leveraged to distinguish between the considered set of NDs, specifically by measuring the stability, the speed of writing, the time spent not writing, and the pressure variations between groups from our handcrafted explainable metrics, which shows p-values lower than 0.0001 for multiple tasks. Using various classification algorithms on the computed metrics, we obtain up to 87% accuracy to discriminate AD and healthy controls (CTL), and up to 69% for PD vs CTL., Comment: 19 pages plus references, to be submitted to IEEE JHBI

Sea ice is a key element of the global Earth system, with a major impact on global climate and regional weather. Unfortunately, accurate sea ice modeling is challenging due to the diversity and complexity of underlying physics happening there, and a relative lack of ground truth observations. This is especially true for the Marginal Ice Zone (MIZ), which is the area where sea ice is affected by incoming ocean waves. Waves contribute to making the area dynamic, and due to the low survival time of the buoys deployed there, the MIZ is challenging to monitor. In 2022-2023, we released 79 OpenMetBuoys (OMBs) around Svalbard, both in the MIZ and the ocean immediately outside of it. OMBs are affordable enough to be deployed in large number, and gather information about drift (GPS position) and waves (1-dimensional elevation spectrum). This provides data focusing on the area around Svalbard with unprecedented spatial and temporal resolution. We expect that this will allow to perform validation and calibration of ice models and remote sensing algorithms.

The advent of quantum computers has justified the development of quantum machine learning algorithms , based on the adaptation of the principles of machine learning to the formalism of qubits. Among such quantum algorithms, anomaly detection represents an important problem crossing several disciplines from cybersecurity, to fraud detection to particle physics. We summarize the key concepts involved in quantum computing, introducing the formal concept of quantum speed up. The survey provides a structured map of anomaly detection based on quantum machine learning. We have grouped existing algorithms according to the different learning methods, namely quantum supervised, quantum unsupervised and quantum reinforcement learning, respectively. We provide an estimate of the hardware resources to provide sufficient computational power in the future. The survey provides a systematic and compact understanding of the techniques belonging to each category. We eventually provide a discussion on the computational complexity of the learning methods in real application domains.

We give an overview of the theoretical description of nuclear reactions involving weakly-bound nuclei. Some of the more widespread reaction formalisms employed in the analysis of these reactions are briefly introduced, including various recent developments. We put special emphasis on the continuum-discretized coupled-channel (CDCC) method and its extensions to incorporate core and target excitations as well as its application to three-body projectiles. The role of the continuum for one-nucleon transfer reactions is also discussed. The problem of the evaluation of inclusive breakup cross sections is addressed within the Ichimura-Austern-Vincent (IAV) model. Other methods, such as those based on a semiclasical description of the scattering process, are also briefly introduced and some of their applications are discussed and a brief discussion on topics of current interest, such as nucleon-nucleon correlations, uncertainty evaluation and non-locality is presented., Comment: review paper; submitted to EPJA

Spiking Neural Networks (SNNs) have the potential for rich spatio-temporal signal processing thanks to exploiting both spatial and temporal parameters. The temporal dynamics such as time constants of the synapses and neurons and delays have been recently shown to have computational benefits that help reduce the overall number of parameters required in the network and increase the accuracy of the SNNs in solving temporal tasks. Optimizing such temporal parameters, for example, through gradient descent, gives rise to a temporal architecture for different problems. As has been shown in machine learning, to reduce the cost of optimization, architectural biases can be applied, in this case in the temporal domain. Such inductive biases in temporal parameters have been found in neuroscience studies, highlighting a hierarchy of temporal structure and input representation in different layers of the cortex. Motivated by this, we propose to impose a hierarchy of temporal representation in the hidden layers of SNNs, highlighting that such an inductive bias improves their performance. We demonstrate the positive effects of temporal hierarchy in the time constants of feed-forward SNNs applied to temporal tasks (Multi-Time-Scale XOR and Keyword Spotting, with a benefit of up to 4.1% in classification accuracy). Moreover, we show that such architectural biases, i.e. hierarchy of time constants, naturally emerge when optimizing the time constants through gradient descent, initialized as homogeneous values. We further pursue this proposal in temporal convolutional SNNs, by introducing the hierarchical bias in the size and dilation of temporal kernels, giving rise to competitive results in popular temporal spike-based datasets., Comment: 16 pages, 9 figures, pre-print

The International Muon Collider Collaboration (IMCC) [1] was established in 2020 following the recommendations of the European Strategy for Particle Physics (ESPP) and the implementation of the European Strategy for Particle Physics-Accelerator R&D Roadmap by the Laboratory Directors Group [2], hereinafter referred to as the the European LDG roadmap. The Muon Collider Study (MuC) covers the accelerator complex, detectors and physics for a future muon collider. In 2023, European Commission support was obtained for a design study of a muon collider (MuCol) [3]. This project started on 1st March 2023, with work-packages aligned with the overall muon collider studies. In preparation of and during the 2021-22 U.S. Snowmass process, the muon collider project parameters, technical studies and physics performance studies were performed and presented in great detail. Recently, the P5 panel [4] in the U.S. recommended a muon collider R&D, proposed to join the IMCC and envisages that the U.S. should prepare to host a muon collider, calling this their "muon shot". In the past, the U.S. Muon Accelerator Programme (MAP) [5] has been instrumental in studies of concepts and technologies for a muon collider., Comment: This document summarises the International Muon Collider Collaboration (IMCC) progress and status of the Muon Collider R&D programme

Modeling observations of the archetypal debris disk around $\beta$ Pic, obtained in 2023 January with the MIRI MRS on board JWST, reveals significant differences compared with that obtained with the IRS on board Spitzer. The bright 5 - 15 $\mu$m continuum excess modeled using a $\sim$600 K black body has disappeared. The previously prominent 18 and 23 $\mu$m crystalline forsterite emission features, arising from cold dust ($\sim$100 K) in the Rayleigh limit, have disappeared and been replaced by very weak features arising from the hotter 500 K dust population. Finally, the shape of the 10 $\mu$m silicate feature has changed, consistent with a shift in the temperature of the warm dust population from $\sim$300 K to $\sim$500 K and an increase in the crystalline fraction of the warm, silicate dust. Stellar radiation pressure may have blown both the hot and the cold crystalline dust particles observed in the Spitzer spectra out of the planetary system during the intervening 20 years between the Spitzer and JWST observations. These results indicate that the $\beta$ Pic system has a dynamic circumstellar environment, and that periods of enhanced collisions can create large clouds of dust that sweep through the planetary system., Comment: 15 pages, 8 figures, ApJ in press

The emergence of Centralized-RAN (C-RAN) has revolutionized mobile network infrastructure, offering streamlined cell-site engineering and enhanced network management capabilities. As C-RAN gains momentum, the focus shifts to optimizing fronthaul links. While fiber fronthaul guarantees performance, wireless alternatives provide cost efficiency and scalability, making them preferable in densely urbanized areas. However, wireless fronthaul often requires expensive over-dimensioning to overcome the challenging atmospheric attenuation typical of high frequencies. We propose a framework designed to continuously align radio access capacity with fronthaul link quality to overcome this rigidity. By gradually adapting radio access capacity to available fronthaul capacity, the framework ensures smooth degradation rather than complete service loss. Various strategies are proposed, considering factors like functional split and beamforming technology and exploring the tradeoff between adaptation strategy complexity and end-to-end system performance. Numerical evaluations using experimental rain attenuation data illustrate the framework's effectiveness in optimizing radio access capacity under realistically variable fronthaul link quality, ultimately proving the importance of adaptive capacity management in maximizing C-RAN efficiency.

Recent LLMs are able to generate high-quality multilingual texts, indistinguishable for humans from authentic human-written ones. Research in machine-generated text detection is however mostly focused on the English language and longer texts, such as news articles, scientific papers or student essays. Social-media texts are usually much shorter and often feature informal language, grammatical errors, or distinct linguistic items (e.g., emoticons, hashtags). There is a gap in studying the ability of existing methods in detection of such texts, reflected also in the lack of existing multilingual benchmark datasets. To fill this gap we propose the first multilingual (22 languages) and multi-platform (5 social media platforms) dataset for benchmarking machine-generated text detection in the social-media domain, called MultiSocial. It contains 472,097 texts, of which about 58k are human-written and approximately the same amount is generated by each of 7 multilingual LLMs. We use this benchmark to compare existing detection methods in zero-shot as well as fine-tuned form. Our results indicate that the fine-tuned detectors have no problem to be trained on social-media texts and that the platform selection for training matters.

Speech separation, the task of isolating multiple speech sources from a mixed audio signal, remains challenging in noisy environments. In this paper, we propose a generative correction method to enhance the output of a discriminative separator. By leveraging a generative corrector based on a diffusion model, we refine the separation process for single-channel mixture speech by removing noises and perceptually unnatural distortions. Furthermore, we optimize the generative model using a predictive loss to streamline the diffusion model's reverse process into a single step and rectify any associated errors by the reverse process. Our method achieves state-of-the-art performance on the in-domain Libri2Mix noisy dataset, and out-of-domain WSJ with a variety of noises, improving SI-SNR by 22-35% relative to SepFormer, demonstrating robustness and strong generalization capabilities., Comment: Accepted at INTERSPEECH 2024

Background: An important aspect of reducing nuclear many-body problems to few-body models is the presence of Pauli forbidden (PF) states, which are excluded in fully antisymmetrized calculations. Insufficient treatments of PF states in deformed halo nuclei underscore the need for model refinement. Purpose: We propose a new method utilizing Nilsson states as PF states in the orthogonality condition model, and investigate the impact of PF states on the properties of resonant states. Method: We investigate the scattering states of ${}^8\mathrm{B}$ within the Particle Rotor Model (PRM) framework based on a deformed ${}^7\mathrm{Be}$ core and $p$ two-body model. We compare several methods for eliminating PF states and test them with the experimental data. Results: Our model successfully reproduces the experimental excitation function for elastic scattering cross section by properly eliminating PF states. The same calculation predicts the presence of a low-energy bump in the inelastic scattering excitation function, although its position is overestimated by about 1 MeV compared to experimental data. Conclusion: This study extends the applicability of the PRM, offering a comprehensive approach for exploring structures and reactions of loosely bound nuclei like ${}^8$B. Future integration with the continuum discretized coupled channels (CDCC) method promises to further advance the research., Comment: 8 pages, 6 figures

Social connections are conduits through which individuals communicate, information propagates, and diseases spread. Identifying individuals who are more likely to adopt ideas and spread them is essential in order to develop effective information campaigns, maximize the reach of resources, and fight epidemics. Influence maximization algorithms are used to identify sets of influencers. Based on extensive computer simulations on synthetic and ten diverse real-world social networks we show that seeding information using these methods creates information gaps. Our results show that these algorithms select influencers who do not disseminate information equitably, threatening to create an increasingly unequal society. To overcome this issue we devise a multi-objective algorithm which maximizes influence and information equity. Our results demonstrate it is possible to reduce vulnerability at a relatively low trade-off with respect to spread. This highlights that in our search for maximizing information we do not need to compromise on information equality.

In this paper, we present four general constructions of $\sigma$ self-orthogonal matrix-product codes associated with Toeplitz matrices. The first one relies on the $\sigma'$ dual of a known $\sigma'$ dual-containing matrix-product code; the second one is founded on quasi-$\widehat{\sigma}$ matrices, where we provide an efficient algorithm for generating them on the basic of Toeplitz matrices; and the last two ones are based on the utilization of certain special Toeplitz matrices. Concrete examples and detailed comparisons are provided. As a byproduct, we also find an application of Toeplitz matrices in $\widetilde{\tau}$-optimal defining matrices.

We approach designing a state-space model for deep learning applications through its dual representation, the transfer function, and uncover a highly efficient sequence parallel inference algorithm that is state-free: unlike other proposed algorithms, state-free inference does not incur any significant memory or computational cost with an increase in state size. We achieve this using properties of the proposed frequency domain transfer function parametrization, which enables direct computation of its corresponding convolutional kernel's spectrum via a single Fast Fourier Transform. Our experimental results across multiple sequence lengths and state sizes illustrates, on average, a 35% training speed improvement over S4 layers -- parametrized in time-domain -- on the Long Range Arena benchmark, while delivering state-of-the-art downstream performances over other attention-free approaches. Moreover, we report improved perplexity in language modeling over a long convolutional Hyena baseline, by simply introducing our transfer function parametrization. Our code is available at https://github.com/ruke1ire/RTF., Comment: Resubmission 02/06/2024: Fixed minor typo of recurrent form RTF

Quantization-aware training comes with a fundamental challenge: the derivative of quantization functions such as rounding are zero almost everywhere and nonexistent elsewhere. Various differentiable approximations of quantization functions have been proposed to address this issue. In this paper, we prove that when the learning rate is sufficiently small, a large class of weight gradient estimators is equivalent with the straight through estimator (STE). Specifically, after swapping in the STE and adjusting both the weight initialization and the learning rate in SGD, the model will train in almost exactly the same way as it did with the original gradient estimator. Moreover, we show that for adaptive learning rate algorithms like Adam, the same result can be seen without any modifications to the weight initialization and learning rate. We experimentally show that these results hold for both a small convolutional model trained on the MNIST dataset and for a ResNet50 model trained on ImageNet.

The measurement of the $^7$Be($d$,$^3$He)$^6$Li* transfer cross section at 5 MeV/u is carried out. The population of the 2.186 MeV excited state of $^6$Li in this reaction channel is observed for the first time. The experimental angular distributions have been analyzed in the finite range DWBA and coupled-channel frameworks. The effect of the $^7$Be($d$,$^3$He)$^6$Li reaction on both the $^6$Li and $^7$Li abundances are investigated at the relevant big-bang nucleosynthesis energies. The excitation function is calculated by TALYS and normalized to the experimental data. The $S$ factor of the ($d$,$^3$He) channel from the present work is about 50$\%$ lower than existing data at nearby energies. At big-bang energies, the $S$ factor is about three orders of magnitude smaller than that of the ($d,p$) channel. The ($d$,$^3$He) reaction rate is found to have a less than 0.1$\%$ effect on the $^{6,7}$Li abundances., Comment: 7 pages, 5 figures

We present the latest generation of MobileNets, known as MobileNetV4 (MNv4), featuring universally efficient architecture designs for mobile devices. At its core, we introduce the Universal Inverted Bottleneck (UIB) search block, a unified and flexible structure that merges Inverted Bottleneck (IB), ConvNext, Feed Forward Network (FFN), and a novel Extra Depthwise (ExtraDW) variant. Alongside UIB, we present Mobile MQA, an attention block tailored for mobile accelerators, delivering a significant 39% speedup. An optimized neural architecture search (NAS) recipe is also introduced which improves MNv4 search effectiveness. The integration of UIB, Mobile MQA and the refined NAS recipe results in a new suite of MNv4 models that are mostly Pareto optimal across mobile CPUs, DSPs, GPUs, as well as specialized accelerators like Apple Neural Engine and Google Pixel EdgeTPU - a characteristic not found in any other models tested. Finally, to further boost accuracy, we introduce a novel distillation technique. Enhanced by this technique, our MNv4-Hybrid-Large model delivers 87% ImageNet-1K accuracy, with a Pixel 8 EdgeTPU runtime of just 3.8ms.

In this paper, we investigate the algebraic structure for polycyclic codes over a specific class of serial rings, defined as $\mathscr R=R[x_1,\ldots, x_s]/\langle t_1(x_1),\ldots, t_s(x_s) \rangle$, where $R$ is a chain ring and each $t_i(x_i)$ in $R[x_i]$ for $i\in\{1,\ldots, s\}$ is a monic square-free polynomial. We define quasi-$s$-dimensional polycyclic codes and establish an $R$-isomorphism between these codes and polycyclic codes over $\mathscr R$. We provide necessary and sufficient conditions for the existence of annihilator self-dual, annihilator self-orthogonal, annihilator linear complementary dual, and annihilator dual-containing polycyclic codes over this class of rings. We also establish the CSS construction for annihilator dual-preserving polycyclic codes over the chain ring $R$ and use this construction to derive quantum codes from polycyclic codes over $\mathscr{R}$.

Anomaly Detection and Segmentation (AD&S) is crucial for industrial quality control. While existing methods excel in generating anomaly scores for each pixel, practical applications require producing a binary segmentation to identify anomalies. Due to the absence of labeled anomalies in many real scenarios, standard practices binarize these maps based on some statistics derived from a validation set containing only nominal samples, resulting in poor segmentation performance. This paper addresses this problem by proposing a test time training strategy to improve the segmentation performance. Indeed, at test time, we can extract rich features directly from anomalous samples to train a classifier that can discriminate defects effectively. Our general approach can work downstream to any AD&S method that provides an anomaly score map as output, even in multimodal settings. We demonstrate the effectiveness of our approach over baselines through extensive experimentation and evaluation on MVTec AD and MVTec 3D-AD., Comment: Accepted at VAND 2.0, CVPRW 2024

Low-precision quantization is recognized for its efficacy in neural network optimization. Our analysis reveals that non-quantized elementwise operations which are prevalent in layers such as parameterized activation functions, batch normalization, and quantization scaling dominate the inference cost of low-precision models. These non-quantized elementwise operations are commonly overlooked in SOTA efficiency metrics such as Arithmetic Computation Effort (ACE). In this paper, we propose ACEv2 - an extended version of ACE which offers a better alignment with the inference cost of quantized models and their energy consumption on ML hardware. Moreover, we introduce PikeLPN, a model that addresses these efficiency issues by applying quantization to both elementwise operations and multiply-accumulate operations. In particular, we present a novel quantization technique for batch normalization layers named QuantNorm which allows for quantizing the batch normalization parameters without compromising the model performance. Additionally, we propose applying Double Quantization where the quantization scaling parameters are quantized. Furthermore, we recognize and resolve the issue of distribution mismatch in Separable Convolution layers by introducing Distribution-Heterogeneous Quantization which enables quantizing them to low-precision. PikeLPN achieves Pareto-optimality in efficiency-accuracy trade-off with up to 3X efficiency improvement compared to SOTA low-precision models., Comment: Accepted in CVPR 2024. 10 Figures, 9 Tables

Medical open-domain question answering demands substantial access to specialized knowledge. Recent efforts have sought to decouple knowledge from model parameters, counteracting architectural scaling and allowing for training on common low-resource hardware. The retrieve-then-read paradigm has become ubiquitous, with model predictions grounded on relevant knowledge pieces from external repositories such as PubMed, textbooks, and UMLS. An alternative path, still under-explored but made possible by the advent of domain-specific large language models, entails constructing artificial contexts through prompting. As a result, "to generate or to retrieve" is the modern equivalent of Hamlet's dilemma. This paper presents MedGENIE, the first generate-then-read framework for multiple-choice question answering in medicine. We conduct extensive experiments on MedQA-USMLE, MedMCQA, and MMLU, incorporating a practical perspective by assuming a maximum of 24GB VRAM. MedGENIE sets a new state-of-the-art in the open-book setting of each testbed, allowing a small-scale reader to outcompete zero-shot closed-book 175B baselines while using up to 706$\times$ fewer parameters. Our findings reveal that generated passages are more effective than retrieved ones in attaining higher accuracy., Comment: ACL 2024 (camera-ready paper)

The spectrum crunch challenge poses a vital threat to the progress of cellular networks and recently prompted the inclusion of millimeter wave (mmWave) and Upper 6GHz (U6G) in the 3GPP standards. These two bands promise to unlock a large portion of untapped spectrum, but the harsh propagation due to the increased carrier frequency might negatively impact the performance of urban Radio Access Network (RAN) deployments. Within the span of a year, two co-located 5G networks operating in these frequency bands were deployed at Politecnico di Milano, Milan, Italy, entirely dedicated to the dense urban performance assessment of the two systems. This paper presents an in-depth analysis of the measurement campaigns conducted on them, with the U6G campaign representing the first of its kind. A benchmark is provided by ray-tracing simulations. The results suggest that networks operating in these frequency bands provide good indoor and outdoor coverage and throughput in urban scenarios, even when deployed in the macro base station setup common to lower frequencies. In addition, a comparative performance analysis of these two key technologies is provided, offering insights on their relative strengths, weaknesses and improvement margins and informing on which bands is better suited for urban macro coverage.