Search

The last decade has seen widespread adoption of Machine Learning (ML) components in software systems. This has occurred in nearly every domain, from natural language processing to computer vision. These ML components range from relatively simple neural networks to complex and resource-intensive large language models. However, despite this widespread adoption, little is known about the supply chain relationships that produce these models, which can have implications for compliance and security. In this work, we conduct an extensive analysis of 760,460 models and 175,000 datasets mined from the popular model-sharing site Hugging Face. First, we evaluate the current state of documentation in the Hugging Face supply chain, report real-world examples of shortcomings, and offer actionable suggestions for improvement. Next, we analyze the underlying structure of the extant supply chain. Finally, we explore the current licensing landscape against what was reported in prior work and discuss the unique challenges posed in this domain. Our results motivate multiple research avenues, including the need for better license management for ML models/datasets, better support for model documentation, and automated inconsistency checking and validation. We make our research infrastructure and dataset available to facilitate future research.

Generative AI (GenAI) tools have already started to transform software development practices. Despite their utility in tasks such as writing code, the use of these tools raises important legal questions and potential risks, particularly those associated with copyright law. In the midst of this uncertainty, this paper presents a study jointly conducted by software engineering and legal researchers that surveyed 574 GitHub developers who use GenAI tools for development activities. The survey and follow-up interviews probed the developers' opinions on emerging legal issues as well as their perception of copyrightability, ownership of generated code, and related considerations. We also investigate potential developer misconceptions, the impact of GenAI on developers' work, and developers' awareness of licensing/copyright risks. Qualitative and quantitative analysis showed that developers' opinions on copyright issues vary broadly and that many developers are aware of the nuances these legal questions involve. We provide: (1) a survey of 574 developers on the licensing and copyright aspects of GenAI for coding, (2) a snapshot of practitioners' views at a time when GenAI and perceptions of it are rapidly evolving, and (3) an analysis of developers' views, yielding insights and recommendations that can inform future regulatory decisions in this evolving field.

When learning to play an imperfect information game, it is often easier to first start with the basic mechanics of the game rules. For example, one can play several example rounds with private cards revealed to all players to better understand the basic actions and their effects. Building on this intuition, this paper introduces {\it progressive hiding}, an algorithm that learns to play imperfect information games by first learning the basic mechanics and then progressively adding information constraints over time. Progressive hiding is inspired by methods from stochastic multistage optimization such as scenario decomposition and progressive hedging. We prove that it enables the adaptation of counterfactual regret minimization to games where perfect recall is not satisfied. Numerical experiments illustrate that progressive hiding can achieve optimal payoff in a benchmark of emergent communication trading game.

Recently, a framework for computing the symmetry-resolved entanglement entropy for non-invertible symmetries in $1{+}1$d conformal field theories has been proposed by Saura-Bastida, Das, Sierra and Molina-Vilaplana [Phys. Rev. D109, 105026]. We revisit their theoretical setup, paying particular attention to possible contributions from the conformal boundary conditions imposed at the entangling surface -- a potential subtlety that was not addressed in the original proposal. We find that the presence of boundaries modifies the construction of projectors onto irreducible sectors, compared to what can be expected from a pure bulk approach. This is a direct consequence of the fusion algebra of non-invertible symmetries being different in the presence or absence of boundaries on which defects can end. We apply our formalism to the case of the Fibonacci category symmetry in the three-state Potts and tricritical Ising model and the Rep($S_3$) fusion category symmetry in the $SU(2)_4$ Wess-Zumino-Witten conformal field theory. We numerically corroborate our findings by simulating critical anyonic chains with these symmetries as a finite lattice substitute for the expected entanglement Hamiltonian. Our predictions for the symmetry-resolved entanglement for non-invertible symmetries seem to disagree with the recent work by Saura-Bastida et al., Comment: 32 pages, 14 figures, v2; references added

Online advertising banners are sold in real-time through auctions.Typically, the more banners a user is shown, the smaller the marginalvalue of the next banner for this user is. This fact can be detected bybasic ML models, that can be used to predict how previously won auctionsdecrease the current opportunity value. However, learning is not enough toproduce a bid that correctly accounts for how winning the current auctionimpacts the future values. Indeed, a policy that uses this prediction tomaximize the expected payoff of the current auction could be dubbedimpatient because such policy does not fully account for the repeatednature of the auctions. Under this perspective, it seems that most biddersin the literature are impatient. Unsurprisingly, impatience induces a cost.We provide two empirical arguments for the importance of this cost ofimpatience. First, an offline counterfactual analysis and, second, a notablebusiness metrics improvement by mitigating the cost of impatience withpolicy learning

In future nuclear arms control treaties, it will be necessary to authenticate the hardware and software components of verification measurement systems, i.e., to ensure these systems are functioning as intended and have not been tampered with by malicious actors. While methods such as source code hashing and static analysis can help verify the integrity of software components, they may not be capable of detecting tampering with environment variables, external libraries, or the firmware and hardware of radiation measurement systems. In this article, we introduce the concept of physical differential fuzz testing as a challenge-response-style tamper indicator that can holistically and simultaneously test all the above components in a cyber-physical system. In essence, we randomly sample (or "fuzz") the untampered system's parameter space, including both normal and off-normal parameter values, and consider the time series of outputs as the baseline signature of the system. Re-running the same input sequence on a untampered system will produce an output sequence consistent with this baseline, while running the same input sequence on a tampered system will produce a modified output sequence and raise an alarm. We then apply this concept to authenticating the radiation measurement equipment in nuclear weapon verification systems and conduct demonstration fuzz testing measurements with a sodium iodide (NaI) gamma ray spectrometer. Because there is Poisson noise in the measured output spectra, we also use a mechanism for comparing inherently noisy or stochastic fuzzing sequences. We show that physical differential fuzz testing can detect two types of tamper attempts, and conclude that it is a promising framework for authenticating future cyber-physical systems in nuclear arms control, safeguards, and beyond., Comment: 15 pages, 7 figures, 1 table

Most modern software products incorporate open source components, which requires compliance with each component's licenses. As noncompliance can lead to significant repercussions, organizations often seek advice from legal practitioners to maintain license compliance, address licensing issues, and manage the risks of noncompliance. While legal practitioners play a critical role in the process, little is known in the software engineering community about their experiences within the open source license compliance ecosystem. To fill this knowledge gap, a joint team of software engineering and legal researchers designed and conducted a survey with 30 legal practitioners and related occupations and then held 16 follow-up interviews. We identified different aspects of OSS license compliance from the perspective of legal practitioners, resulting in 14 key findings in three main areas of interest: the general ecosystem of compliance, the specific compliance practices of legal practitioners, and the challenges that legal practitioners face. We discuss the implications of our findings., Comment: 24 pages, 2 figures, FSE 2024

The demand-supply balance of electricity systems is fundamentally linked to climate conditions. In light of this, the present study aims to model the effect of climate change on the European electricity system, specifically on its long-term reliability. A resource adequate power system -- a system where electricity supply covers demand -- is sensitive to generation capacity, demand patterns, and the network structure and capacity. Climate change is foreseen to affect each of these components. In this analysis, we focused on two drivers of power system adequacy: the impact of temperature variations on electricity demand, and of water inflows changes on hydro generation. Using a post-processing approach, based on results found in the literature, the inputs of a large-scale electricity market model covering the European region were modified. The results show that climate change may decrease total LOLE (Loss of Load Expectation) hours in Europe by more than 50%, as demand will largely decrease because of a higher temperatures during winter. We found that the climate change impact on demand tends to decrease LOLE values, while the climate change effects on hydrological conditions tend to increase LOLE values. The study is built on a limited amount of open-source data and can flexibly incorporate various sets of assumptions. Outcomes also show the current difficulties to reliably model the effects of climate change on power system adequacy. Overall, our presented method displays the relevance of climate change effects in electricity network studies.

The effectiveness of advertising in e-commerce largely depends on the ability of merchants to bid on and win impressions for their targeted users. The bidding procedure is highly complex due to various factors such as market competition, user behavior, and the diverse objectives of advertisers. In this paper we consider the problem at the level of user timelines instead of individual bid requests, manipulating full policies (i.e. pre-defined bidding strategies) and not bid values. In order to optimally allocate policies to users, typical multiple treatments allocation methods solve knapsack-like problems which aim at maximizing an expected value under constraints. In the industrial contexts such as online advertising, we argue that optimizing for the probability of success is a more suited objective than expected value maximization, and we introduce the SuccessProbaMax algorithm that aims at finding the policy allocation which is the most likely to outperform a fixed reference policy. Finally, we conduct comprehensive experiments both on synthetic and real-world data to evaluate its performance. The results demonstrate that our proposed algorithm outperforms conventional expected-value maximization algorithms in terms of success rate., Comment: accepted to AAAI 2024, 7 pages main text, 9 pages references and appendix, 22 figures

In January 2020, a workshop was held at EMBL-EBI (Hinxton, UK) to discuss data requirements for deposition and validation of cryoEM structures, with a focus on single-particle analysis. The meeting was attended by 45 experts in data processing, model building and refinement, validation, and archiving of such structures. This report describes the workshop's motivation and history, the topics discussed, and consensus recommendations resulting from the workshop. Some challenges for future methods-development efforts in this area are also highlighted, as is the implementation to date of some of the recommendations., Comment: Outcomes of a wwPDB/EMDB workshop on cryoEM data management, deposition and validation

The distance to the stars is a fundamental parameter, which is determined via two primary methods - parallax and luminosity. While the parallax is a direct trigonometric method, the luminosity distance is usually influenced by interstellar extinction. As long as the optical properties of dust grains are wavelength-dependent this contamination can be corrected. With increasing grain size, however, the extinction properties become grey, i.e. they contribute at wavelengths but can no longer be detected by photometry. However, as the grain size increases, the extinction properties become grey, meaning these particles contribute by a constant at wavelengths below 1 mic, making them undetectable by photometry in the optical. In this study, we compare the parallactic and luminosity distances of a pristine sample of 33 well-known early-type stars with non-peculiar reddening curves and find that the luminosity distance overestimates the parallactic distance in 80% of the cases. This discrepancy can be removed when incorporating a population of large, submicrometer-sized dust grains in a dust model that provides grey extinction which diminishes the luminosity distance accordingly., Comment: accepted by ApJL 25/12/2024; custom software availabe at https://github.com/rsiebenm/dark_dust

We introduce a class of Bayesian bidding games for which we prove that the set of pure Nash equilibria is a (non-empty) sublattice and we give a sufficient condition for uniqueness that is often verified in the context of markets with inelastic demand. We propose a dynamic that converges to the extrema of the equilibrium set and derive a scheme to compute the extreme Nash equilibria.

Imaging the structure and observing the dynamics of isolated proteins using single-particle X-ray diffractive imaging (SPI) is one of the potential applications of X-ray free-electron lasers (XFELs). Currently, SPI experiments on isolated proteins are limited by three factors: low signal strength, limited data and high background from gas scattering. The last two factors are largely due to the shortcomings of the aerosol sample delivery methods in use. Here we present our modified electrospray ionization (ESI) source, which we dubbed Helium-ESI (He-ESI). With it, we increased particle delivery into the interaction region by a factor of 10, for 26 nm-sized biological particles, and decreased the gas load in the interaction chamber corresponding to an 80% reduction in gas scattering when compared to the original ESI. These improvements will lead to a significant increase in the quality and quantity of SPI diffraction patterns in future experiments using He-ESI, resulting in higher-resolution structures.

The prime-counting function $\pi(x)$ which returns the number of primes smaller or equal to a given number is a topic of interest in number theory. An algorithm based on a cyclic group isomorphic to $Z/nZ$, the so-called $Z$-functions, was proposed in view to outperform its pieers. The approach suggests a time complexity $\mathcal{O}(x^{1/2})$ in agreement with optimality of a 2-D squared adaptive-recursive algorithm. The present work is a presentation of various approaches as ascending factorization, the main sequence of odd numbers and partial sequences, T-series, counting function of prime composites, $Z$-modular forms and combinatorial aspects., Comment: 10 pages

We consider a repeated auction where the buyer's utility for an item depends on the time that elapsed since his last purchase. We present an algorithm to build the optimal bidding policy, and then, because optimal might be impractical, we discuss the cost for the buyer of limiting himself to shading policies.

Convolutional frontends are a typical choice for Transformer-based automatic speech recognition to preprocess the spectrogram, reduce its sequence length, and combine local information in time and frequency similarly. However, the width and height of an audio spectrogram denote different information, e.g., due to reverberation as well as the articulatory system, the time axis has a clear left-to-right dependency. On the contrary, vowels and consonants demonstrate very different patterns and occupy almost disjoint frequency ranges. Therefore, we hypothesize, global attention over frequencies is beneficial over local convolution. We obtain 2.4 % relative word error rate reduction (rWERR) on a production scale Conformer transducer replacing its convolutional neural network frontend by the proposed F-Attention module on Alexa traffic. To demonstrate generalizability, we validate this on public LibriSpeech data with a long short term memory-based listen attend and spell architecture obtaining 4.6 % rWERR and demonstrate robustness to (simulated) noisy conditions.

We consider the dataset valuation problem, that is, the problem of quantifying the incremental gain, to some relevant pre-defined utility of a machine learning task, of aggregating an individual dataset to others. The Shapley value is a natural tool to perform dataset valuation due to its formal axiomatic justification, which can be combined with Monte Carlo integration to overcome the computational tractability challenges. Such generic approximation methods, however, remain expensive in some cases. In this paper, we exploit the knowledge about the structure of the dataset valuation problem to devise more efficient Shapley value estimators. We propose a novel approximation, referred to as discrete uniform Shapley, which is expressed as an expectation under a discrete uniform distribution with support of reasonable size. We justify the relevancy of the proposed framework via asymptotic and non-asymptotic theoretical guarantees and illustrate its benefits via an extensive set of numerical experiments., Comment: 22 pages