Search

Local Polynomial Regression (LPR) is a widely used nonparametric method for modeling complex relationships due to its flexibility and simplicity. It estimates a regression function by fitting low-degree polynomials to localized subsets of the data, weighted by proximity. However, traditional LPR is sensitive to outliers and high-leverage points, which can significantly affect estimation accuracy. This paper revisits the kernel function used to compute regression weights and proposes a novel framework that incorporates both predictor and response variables in the weighting mechanism. By introducing two positive definite kernels, the proposed method robustly estimates weights, mitigating the influence of outliers through localized density estimation. The method is implemented in Python and is publicly available at https://github.com/yaniv-shulman/rsklpr, demonstrating competitive performance in synthetic benchmark experiments. Compared to standard LPR, the proposed approach consistently improves robustness and accuracy, especially in heteroscedastic and noisy environments, without requiring multiple iterations. This advancement provides a promising extension to traditional LPR, opening new possibilities for robust regression applications.

This paper introduces an adaptive physics-guided neural network (APGNN) framework for predicting quality attributes from image data by integrating physical laws into deep learning models. The APGNN adaptively balances data-driven and physics-informed predictions, enhancing model accuracy and robustness across different environments. Our approach is evaluated on both synthetic and real-world datasets, with comparisons to conventional data-driven models such as ResNet. For the synthetic data, 2D domains were generated using three distinct governing equations: the diffusion equation, the advection-diffusion equation, and the Poisson equation. Non-linear transformations were applied to these domains to emulate complex physical processes in image form. In real-world experiments, the APGNN consistently demonstrated superior performance in the diverse thermal image dataset. On the cucumber dataset, characterized by low material diversity and controlled conditions, APGNN and PGNN showed similar performance, both outperforming the data-driven ResNet. However, in the more complex thermal dataset, particularly for outdoor materials with higher environmental variability, APGNN outperformed both PGNN and ResNet by dynamically adjusting its reliance on physics-based versus data-driven insights. This adaptability allowed APGNN to maintain robust performance across structured, low-variability settings and more heterogeneous scenarios. These findings underscore the potential of adaptive physics-guided learning to integrate physical constraints effectively, even in challenging real-world contexts with diverse environmental conditions.

We consider an alternative to the Monte Carlo method for dust continuous radiative transfer simulations: the Quasi-Monte Carlo method. We briefly discuss what it is, its history, and possible implementations. We compare the Monte Carlo method with four pseudo-random number generators and five Quasi-Monte Carlo implementations using different low-discrepancy sequences and the Hammersley set. For the comparison, we study different test matter geometries and problems. We present comparison results for single scatterings of radiation from a point source, multiple scatterings of radiation from a point source, and single scatterings of radiation from a spherical star. In all cases, Quasi-Monte Carlo shows better convergence than Monte Carlo. In several test cases, the gain in computation time to achieve a fixed error value reached 40 times. We obtained ten times speed up in many of the considered tests., Comment: 57 pages, 10 figures, 3 tables; Submitted to Astronomy and Computing

The (Local) Lifting Property ((L)LP) is introduced by Kirchberg and deals with lifting completely positive maps. We give a characterization of the (L)LP in terms of lifting $\ast$-homomorphisms. We use it to prove that if $A$ and $B$ have the LP and $F$ is their finite-dimensional C*-subalgebra, then $A\ast_F B$ has the LP. This answers a question of Ozawa. We prove that Exel's soft tori have the LP. As a consequence we obtain that $C^*(F_n\times F_n)$ is inductive limit of RFD C*-algebras with the LP. We prove that for a class of C*-algebras including $C^*(F_n\times F_n)$ and all contractible C*-algebras, the LLP is equivalent to Ext being a group. As byproduct of methods developed in the paper we generalize Kirchberg's theorem about extensions with the WEP, give short proofs of several, old and new, facts about soft tori, new unified proofs of Li and Shen's characterization of RFD property of free products amalgamated over a finite-dimensional subalgebra and Blackadar's characterization of semiprojectivity of them.

We present an innovative approach leveraging Physics-Guided Neural Networks (PGNNs) for enhancing agricultural quality assessments. Central to our methodology is the application of physics-guided inverse regression, a technique that significantly improves the model's ability to precisely predict quality metrics of crops. This approach directly addresses the challenges of scalability, speed, and practicality that traditional assessment methods face. By integrating physical principles, notably Fick`s second law of diffusion, into neural network architectures, our developed PGNN model achieves a notable advancement in enhancing both the interpretability and accuracy of assessments. Empirical validation conducted on cucumbers and mushrooms demonstrates the superior capability of our model in outperforming conventional computer vision techniques in postharvest quality evaluation. This underscores our contribution as a scalable and efficient solution to the pressing demands of global food supply challenges.

"Reading for Life" seeks to provide a gateway to that research. Reviewing data on reading and literacy and their impact on the physical, mental, and social-emotional components of health, this paper will: (1) Review the current state of children's literacy and health, particularly in the wake of the COVID-19 pandemic; (2) Explore the importance of early-childhood education and literacy interventions in preschool, and their roles in health outcomes; (3) Summarize existing literature on the relationship between literacy rates and physical, mental, and social-emotional health; and (4) Study successful collaborations between the education and medical fields to implement literacy interventions in pediatric healthcare settings. [This paper was prepared by Scholastic Research & Validation as part of the Yale Child Study Center-Scholastic Collaborative for Child & Family Resilience.]

We introduce Displayed Type Theory (dTT), a multi-modal homotopy type theory with discrete and simplicial modes. In the intended semantics, the discrete mode is interpreted by a model for an arbitrary $\infty$-topos, while the simplicial mode is interpreted by Reedy fibrant augmented semi-simplicial diagrams in that model. This simplicial structure is represented inside the theory by a primitive notion of display or dependency, guarded by modalities, yielding a partially-internal form of unary parametricity. Using the display primitive, we then give a coinductive definition, at the simplicial mode, of a type $\mathsf{SST}$ of semi-simplicial types. Roughly speaking, a semi-simplicial type $X$ consists of a type $X_0$ together with, for each $x:X_0$, a displayed semi-simplicial type over $X$. This mimics how simplices can be generated geometrically through repeated cones, and is made possible by the display primitive at the simplicial mode. The discrete part of $\mathsf{SST}$ then yields the usual infinite indexed definition of semi-simplicial types, both semantically and syntactically. Thus, dTT enables working with semi-simplicial types in full semantic generality., Comment: 106 pages, v2: new section on the correctness of SSTs and various improvements throughout

Parametricity is a property of the syntax of type theory implying, e.g., that there is only one function having the type of the polymorphic identity function. Parametricity is usually proven externally, and does not hold internally. Internalising it is difficult because once there is a term witnessing parametricity, it also has to be parametric itself and this results in the appearance of higher dimensional cubes. In previous theories with internal parametricity, either an explicit syntax for higher cubes is present or the theory is extended with a new sort for the interval. In this paper we present a type theory with internal parametricity which is a simple extension of Martin-L\"of type theory: there are a few new type formers, term formers and equations. Geometry is not explicit in this syntax, but emergent: the new operations and equations only refer to objects up to dimension 3. We show that this theory is modelled by presheaves over the BCH cube category. Fibrancy conditions are not needed because we use span-based rather than relational parametricity. We define a gluing model for this theory implying that external parametricity and canonicity hold. The theory can be seen as a special case of a new kind of modal type theory, and it is the simplest setting in which the computational properties of higher observational type theory can be demonstrated.

By Bekka's theorem the group C*-algebra of an amenable group $G$ is residually finite dimensional (RFD) if and only if $G$ is maximally almost periodic (MAP). We generalize this result in two directions of dynamical flavour. Firstly, we provide a sufficient condition for the RFD property of the C*-algebra of an amenable \'etale groupoid. Secondly, we characterize RFD property for crossed products by amenable actions of discrete groups on C*-algebras. The characterisation can be formulated in various terms, such as primitive ideals, (pure) states and approximations of representations, and can be viewed as a dynamical version of Exel-Loring characterization of RFD C*-algebras. As byproduct of our methods we also characterize the property FD of Lubotzky and Shalom for semidirect products by amenable groups and obtain characterizations of the properties MAP and RF for general semidirect products of groups. The latter descriptions allow us to obtain the properties MAP, RF, RFD and FD for various examples., Comment: 40 pages, v2 contains a modified introduction and several new examples in Section 5

IP prefix hijacks allow adversaries to redirect and intercept traffic, posing a threat to the stability and security of the Internet. To prevent prefix hijacks, networks should deploy RPKI and filter bogus BGP announcements with invalid routes. In this work we evaluate the impact of RPKI deployments on the security and resilience of the Internet. We aim to understand which networks filter invalid routes and how effective that filtering is in blocking prefix hijacks. We extend previous data acquisition and analysis methodologies to obtain more accurate identification of networks that filter invalid routes with RPKI. We find that more than 27% of networks enforce RPKI filtering and show for the first time that deployments follow the business incentives of inter-domain routing: providers have an increased motivation to filter in order to avoid losing customers' traffic. Analyzing the effectiveness of RPKI, we find that the current trend to deploy RPKI on routeservers of Internet Exchange Points (IXPs) only provides a localized protection against hijacks but has negligible impact on preventing their spread globally. In contrast, we show that RPKI filtering in Tier-1 providers greatly benefits the security of the Internet as it limits the spread of hijacks to a localized scope. Based on our observations, we provide recommendations on the future roadmap of RPKI deployment. We make our datasets available for public use [https://sit4.me/rpki]., Comment: Accepted for USENIX Security '23

We show that contrary to appearances, Multimodal Type Theory (MTT) over a 2-category M can be interpreted in any M-shaped diagram of categories having, and functors preserving, M-sized limits, without the need for extra left adjoints. This is achieved by a construction called "co-dextrification" that co-freely adds left adjoints to any such diagram, which can then be used to interpret the "context lock" functors of MTT. Furthermore, if any of the functors in the diagram have right adjoints, these can also be internalized in type theory as negative modalities in the style of FitchTT. We introduce the name Multimodal Adjoint Type Theory (MATT) for the resulting combined general modal type theory. In particular, we can interpret MATT in any finite diagram of toposes and geometric morphisms, with positive modalities for inverse image functors and negative modalities for direct image functors., Comment: 19 pages. v2: Improved notation; extended pre-proceedings version for MFPS 2023. v3: correct some mis-wordings, final. v4: condensed version for conference postproceedings

In this study, we present a novel method for measuring the magnetic susceptibility of liquids using a Mach-Zehnder interferometer. The proposed technique employs a ring magnet to deform the liquid, while a laser beam passes through the liquid to measure the resulting interference pattern. The deformation of the liquid, caused by the known magnetic field of the ring magnet, is used to calculate the magnetic susceptibility. We provide a comprehensive theoretical framework, including the relevant equations and models, for analyzing the data obtained using this method. We compare the Mach-Zehnder interferometer method with other established techniques, highlighting its advantages and disadvantages. Our findings indicate that the Mach-Zehnder interferometer technique offers high accuracy, sensitivity, and potential applications in characterizing magnetic properties of various liquid systems.

Measuring the magnetic field of permanent magnets can be challenging, but recent research has demonstrated the potential of using deformed diamagnetic liquids to estimate the magnetic field. In this paper, we explore two methods for measuring the magnetic field from the response of the diamagnetic liquid. The first method involves measuring the profile of the deformed liquid with a laser and then calculating the square of the magnetic field using an appropriate equation. The second method involves measuring the maximum slope of the liquid and numerically calculating the magnetic field distribution using the model of an ideal solenoid. We present experimental results using these methods and compare them with other established methods for measuring magnetic fields. The results show that the proposed methods are effective and have potential for use in a variety of applications. The proposed methods can help address the challenge of measuring magnetic fields in situations where other methods are not suitable or practical.

In recent years, deep learning has achieved remarkable success in various fields such as image recognition, natural language processing, and speech recognition. The effectiveness of deep learning largely depends on the optimization methods used to train deep neural networks. In this paper, we provide an overview of first-order optimization methods such as Stochastic Gradient Descent, Adagrad, Adadelta, and RMSprop, as well as recent momentum-based and adaptive gradient methods such as Nesterov accelerated gradient, Adam, Nadam, AdaMax, and AMSGrad. We also discuss the challenges associated with optimization in deep learning and explore techniques for addressing these challenges, including weight initialization, batch normalization, and layer normalization. Finally, we provide recommendations for selecting optimization methods for different deep learning tasks and datasets. This paper serves as a comprehensive guide to optimization methods in deep learning and can be used as a reference for researchers and practitioners in the field., Comment: We have decided to withdraw our manuscript 'Optimization Methods in Deep Learning: A Comprehensive Overview' from arXiv as we are now pursuing different research directions. We apologize for any inconvenience and appreciate your understanding

Measuring instruments are vital for obtaining accurate data in various fields, including scientific research, engineering, and manufacturing. However, data acquisition can be challenging, particularly when direct cable connections are not feasible or difficult to establish. In this paper, we propose a novel method for measuring instrument data acquisition utilizing a camera to capture the instrument display and image processing techniques to extract the measured values. Our method combines computer vision and machine learning techniques to recognize and extract numerical values from the instrument display. We demonstrate the effectiveness and accuracy of this approach by applying it to capture the magnetic field of a permanent magnet using a gauss meter and webcam. Our results indicate that the proposed method offers a practical and accurate solution for data acquisition in cases where direct cable connections are not possible. This method has potential applications in scientific research, engineering, and manufacturing industries.

We consider weakly closed transitive algebras of operators containing non-zero compact operators in real Banach spaces (Lomonosov algebras). It is shown that they are naturally divided in three classes: the algebras of real, complex and quaternion classes. The properties and characterizations of algebras in each class as well as some useful examples are presented. It is shown that in separable real Hilbert spaces there is a continuum of pairwise non-similar Lomonosov algebras of complex type and of quaternion type.

It is proved that a commutative algebra $A$ of operators on a reflexive real Banach space has an invariant subspace if each operator $T\in A$ satisfies the condition $$\|1- \varepsilon T^2\|_e \le 1 + o(\varepsilon) \text{ when } \varepsilon\searrow 0,$$ where $\|\cdot\|_e$ is the essential norm. This implies the existence of an invariant subspace for every commutative family of essentially selfadjoint operators on a real Hilbert space.

We explore the security of residential routers and find a range of critical vulnerabilities. Our evaluations show that 10 out of 36 popular routers are vulnerable to injections of fake records via misinterpretation of special characters. We also find that in 15 of the 36 routers the mechanisms, that are meant to prevent cache poisoning attacks, can be circumvented. In our Internet-wide study with an advertisement network, we identified and analyzed 976 residential routers used by web clients, out of which more than 95% were found vulnerable to our attacks. Overall, vulnerable routers are prevalent and are distributed among 177 countries and 4830 networks. To understand the core factors causing the vulnerabilities we perform black- and white-box analyses of the routers. We find that many problems can be attributed to incorrect assumptions on the protocols' behaviour and the Internet, misunderstanding of the standard recommendations, bugs, and simplified DNS software implementations. We provide recommendations to mitigate our attacks. We also set up a tool to enable everyone to evaluate the security of their routers at https://xdi-attack.net/., Comment: 31th USENIX Security Symposium (USENIX Security 22), 2022

This study presents a novel and user-friendly technique for detecting the lateral capillary force on a floating spherical particle. The technique leverages the interplay between the capillary attracting forces, hydrostatic pressure forces, and magnetic repulsion forces. A magnetic field is applied to induce a surface curvature in the liquid, resulting in a non-uniform distribution of capillary and hydrostatic pressure forces across the particle's surface. This leads to a stable equilibrium position of the particle at a specific distance from the magnet. The study analyzes the equilibrium position and other relevant parameters in comparison with the developed theory. Classical mechanics and intermolecular forces are applied to establish the theoretical basis for the method, modeling the behavior of the particle in response to the magnetic field, surface curvature, and hydrostatic pressure. The equilibrium position of the particle is determined by numerically solving the balance of forces equation.

For magnetic field calculations, cylindrical permanent magnets are often approximated as ideal, azimuthally symmetric solenoids. Despite the frequent usage of this approximation, research papers demonstrating the validity and limitations of this approach are scarcely available. In this paper, the experimentally derived magnetic field of a cylindrical permanent magnet is compared with the analytically calculated magnetic field of an ideal solenoid. An experimental setup for measuring the magnetic field distribution is demonstrated and employed for gathering the data. The proposed setup allows to measure the distributions of the axial and radial components of the magnetic field surrounding the magnet. The experimental data is in a very good agreement with the theoretical predictions, confirming the validity of using the model of an ideal solenoid for predicting a magnetic field distribution of a permanent magnet.

We study the formation of images in a reflective sphere in three configurations using caustics of the field of light rays. The optical wavefront emerging from a source point reaching a subject following passage through the optical system is, in general, a Gaussian surface with partial focus along the two principal directions of the Gaussian surface; i.e. there are two images of the source point, each with partial focus. As the source point moves, the images move on two surfaces, referred to as \emph{viewable surfaces}. In our systems, one viewable surface consists of points with radial focus and the other consists of points with azimuthal focus. The problems we study are (1) imaging of a parallel beam of light, (2) imaging of the infinite viewed from a location outside the sphere, and (3) imaging of a planar object viewed through the point of its intersection with the radial line normal to the plane. We verify the existence of two images experimentally and show that the distance between them agrees with the computations.

Information sharing on social networks is ubiquitous, intuitive, and occasionally accidental. However, people may be unaware of the potential negative consequences of disclosures, such as reputational damages. Yet, people use social networks to disclose information about themselves or others, advised only by their own experiences and the context-invariant informed consent mechanism. In two online experiments (N=515 and N=765), we investigated how to aid informed sharing decisions and associate them with the potential outcomes via notifications. Based on the measurements of sharing attitudes, our results showed that the effectiveness of informing the users via notifications may depend on the timing, content, and layout of the notifications, as well as on the users' curiosity and rational cognitive style, motivating information processing. Furthermore, positive emotions may result in disregard of important information. We discuss the implications for user privacy and self-presentation. We provide recommendations on privacy-supporting system design and suggest directions for further research., Comment: The Version of Record of this manuscript has been published and is available in the International Journal of Human-Computer Interaction on 27.06.2022, https://www.tandfonline.com/doi/full/10.1080/10447318.2022.2086592

We examine Hilbert-Schmidt stability (HS-stability) of discrete amenable groups from several angles. We give a short, elementary proof that finitely generated nilpotent groups are HS-stable. We investigate the permanence of HS-stability under central extensions by showing HS-stability is preserved by finite central quotients, but is not preserved in general. We give a characterization of HS-stability for semidirect products $G\rtimes_\gamma \mathbb{Z}$ with $G$ abelian. We use it to construct the first example of a finitely generated amenable HS-stable group which is not permutation stable. Finally, it is proved that for amenable groups flexible HS-stability is equivalent to HS-stability, and very flexible HS stability is equivalent to maximal almost periodicity. There is some overlap of our work with the very recent and very nice preprint of Levit and Vigdorovich. We detail this overlap in the introduction. Where our work overlaps it appears that we take different approaches to the proofs and we feel the two works compliment each other., Comment: 28 pages

A novel comparative technique enabling measurement of the magnetic susceptibility of diamagnetic liquids with the "Moses effect" is presented. The technique is based on the experimental establishment of the deformation of the liquid/vapor interface by a steady magnetic field. The deformation of the liquid surface by a modest magnetic field $(B\sim 0.60T)$ is measured with an optical technique. The surface tension of the liquid is taken into account. The magnetic susceptibilities of the investigated liquids (Ethanol and Glycerol) were calculated from the maximal slope of the liquid/air interface. The suggested approach yields an accuracy on the order of about $0.4-0.6\%$ and hence is superior to previous methods which used the \textquotedblleft Moses effect\textquotedblright\ for the same purpose.

Cryptographic algorithm agility is an important property for DNSSEC: it allows easy deployment of new algorithms if the existing ones are no longer secure. In this work we show that the cryptographic agility in DNSSEC, although critical for provisioning DNS with strong cryptography, also introduces a vulnerability. We find that under certain conditions, when new algorithms are listed in signed DNS responses, the resolvers do not validate DNSSEC. As a result, domains that deploy new ciphers may in fact cause the resolvers not to validate DNSSEC. We exploit this to develop DNSSEC-downgrade attacks and experimentally and ethically evaluate them against popular DNS resolver implementations, public DNS providers, and DNS services used by web clients worldwide. We find that major DNS providers as well as 45% of DNS resolvers used by web clients are vulnerable to our attacks., Comment: This work has been accepted for publication at the ACM SIGSAC Conference on Computer and Communications Security (CCS 22)

We perform the first analysis of methodologies for launching DNS cache poisoning: manipulation at the IP layer, hijack of the inter-domain routing and probing open ports via side channels. We evaluate these methodologies against DNS resolvers in the Internet and compare them with respect to effectiveness, applicability and stealth. Our study shows that DNS cache poisoning is a practical and pervasive threat. We then demonstrate cross-layer attacks that leverage DNS cache poisoning for attacking popular systems, ranging from security mechanisms, such as RPKI, to applications, such as VoIP. In addition to more traditional adversarial goals, most notably impersonation and Denial of Service, we show for the first time that DNS cache poisoning can even enable adversaries to bypass cryptographic defences: we demonstrate how DNS cache poisoning can facilitate BGP prefix hijacking of networks protected with RPKI even when all the other networks apply route origin validation to filter invalid BGP announcements. Our study shows that DNS plays a much more central role in the Internet security than previously assumed. We recommend mitigations for securing the applications and for preventing cache poisoning.

We demonstrate the first downgrade attacks against RPKI. The key design property in RPKI that allows our attacks is the tradeoff between connectivity and security: when networks cannot retrieve RPKI information from publication points, they make routing decisions in BGP without validating RPKI. We exploit this tradeoff to develop attacks that prevent the retrieval of the RPKI objects from the public repositories, thereby disabling RPKI validation and exposing the RPKI-protected networks to prefix hijack attacks. We demonstrate experimentally that at least 47% of the public repositories are vulnerable against a specific version of our attacks, a rate-limiting off-path downgrade attack. We also show that all the current RPKI relying party implementations are vulnerable to attacks by a malicious publication point. This translates to 20.4% of the IPv4 address space. We provide recommendations for preventing our downgrade attacks. However, resolving the fundamental problem is not straightforward: if the relying parties prefer security over connectivity and insist on RPKI validation when ROAs cannot be retrieved, the victim AS may become disconnected from many more networks than just the one that the adversary wishes to hijack. Our work shows that the publication points are a critical infrastructure for Internet connectivity and security. Our main recommendation is therefore that the publication points should be hosted on robust platforms guaranteeing a high degree of connectivity.

Internet resources form the basic fabric of the digital society. They provide the fundamental platform for digital services and assets, e.g., for critical infrastructures, financial services, government. Whoever controls that fabric effectively controls the digital society. In this work we demonstrate that the current practices of Internet resources management, of IP addresses, domains, certificates and virtual platforms are insecure. Over long periods of time adversaries can maintain control over Internet resources which they do not own and perform stealthy manipulations, leading to devastating attacks. We show that network adversaries can take over and manipulate at least 68% of the assigned IPv4 address space as well as 31% of the top Alexa domains. We demonstrate such attacks by hijacking the accounts associated with the digital resources. For hijacking the accounts we launch off-path DNS cache poisoning attacks, to redirect the password recovery link to the adversarial hosts. We then demonstrate that the adversaries can manipulate the resources associated with these accounts. We find all the tested providers vulnerable to our attacks. We recommend mitigations for blocking the attacks that we present in this work. Nevertheless, the countermeasures cannot solve the fundamental problem - the management of the Internet resources should be revised to ensure that applying transactions cannot be done so easily and stealthily as is currently possible.

The traditional design principle for Internet protocols indicates: "Be strict when sending and tolerant when receiving" [RFC1958], and DNS is no exception to this. The transparency of DNS in handling the DNS records, also standardised specifically for DNS [RFC3597], is one of the key features that made it such a popular platform facilitating a constantly increasing number of new applications. An application simply creates a new DNS record and can instantly start distributing it over DNS without requiring any changes to the DNS servers and platforms. Our Internet wide study confirms that more than 1.3M (96% of tested) open DNS resolvers are standard compliant and treat DNS records transparently. In this work we show that this `transparency' introduces a severe vulnerability in the Internet: we demonstrate a new method to launch string injection attacks by encoding malicious payloads into DNS records. We show how to weaponise such DNS records to attack popular applications. For instance, we apply string injection to launch a new type of DNS cache poisoning attack, which we evaluated against a population of open resolvers and found 105K to be vulnerable. Such cache poisoning cannot be prevented with common setups of DNSSEC. Our attacks apply to internal as well as to public services, for instance, we reveal that all eduroam services are vulnerable to our injection attacks, allowing us to launch exploits ranging from unauthorised access to eduroam networks to resource starvation. Depending on the application, our attacks cause system crashes, data corruption and leakage, degradation of security, and can introduce remote code execution and arbitrary errors. In our evaluation of the attacks in the Internet we find that all the standard compliant open DNS resolvers we tested allow our injection attacks against applications and users on their networks.

Hofmann and Streicher famously showed how to lift Grothendieck universes into presheaf topoi, and Streicher has extended their result to the case of sheaf topoi by sheafification. In parallel, van den Berg and Moerdijk have shown in the context of algebraic set theory that similar constructions continue to apply even in weaker metatheories. Unfortunately, sheafification seems not to preserve an important realignment property enjoyed by the presheaf universes that plays a critical role in models of univalent type theory as well as synthetic Tait computability, a recent technique to establish syntactic properties of type theories and programming languages. In the context of multiple universes, the realignment property also implies a coherent choice of codes for connectives at each universe level, thereby interpreting the cumulativity laws present in popular formulations of Martin-L\"of type theory. We observe that a slight adjustment to an argument of Shulman constructs a cumulative universe hierarchy satisfying the realignment property at every level in any Grothendieck topos. Hence one has direct-style interpretations of Martin-L\"of type theory with cumulative universes into all Grothendieck topoi. A further implication is to extend the reach of recent synthetic methods in the semantics of cubical type theory and the syntactic metatheory of type theory and programming languages to all Grothendieck topoi., Comment: Integrated feedback from reviewers, fixed typographic errors

Based on Taylor's hereditarily directed plump ordinals, we define the directed plump ordering on W-types in Martin-L\"of type theory. This ordering is similar to the plump ordering but comes equipped with non-empty finite joins in addition to the usual properties of the plump ordering., Comment: Added funding acknowledgment

Scene understanding is an active research area. Commercial depth sensors, such as Kinect, have enabled the release of several RGB-D datasets over the past few years which spawned novel methods in 3D scene understanding. More recently with the launch of the LiDAR sensor in Apple's iPads and iPhones, high quality RGB-D data is accessible to millions of people on a device they commonly use. This opens a whole new era in scene understanding for the Computer Vision community as well as app developers. The fundamental research in scene understanding together with the advances in machine learning can now impact people's everyday experiences. However, transforming these scene understanding methods to real-world experiences requires additional innovation and development. In this paper we introduce ARKitScenes. It is not only the first RGB-D dataset that is captured with a now widely available depth sensor, but to our best knowledge, it also is the largest indoor scene understanding data released. In addition to the raw and processed data from the mobile device, ARKitScenes includes high resolution depth maps captured using a stationary laser scanner, as well as manually labeled 3D oriented bounding boxes for a large taxonomy of furniture. We further analyze the usefulness of the data for two downstream tasks: 3D object detection and color-guided depth upsampling. We demonstrate that our dataset can help push the boundaries of existing state-of-the-art methods and it introduces new challenges that better represent real-world scenarios.

Unconventional superconductors have Cooper pairs with lower symmetries than in conventional superconductors. In most unconventional superconductors, the additional symmetry breaking occurs in relation to typical ingredients such as strongly correlated Fermi liquid phases, magnetic fluctuations, or strong spin-orbit coupling in noncentrosymmetric structures. In this article, we show that the time-reversal symmetry breaking in the superconductor LaNiGa$_2$ is enabled by its previously unknown topological electronic band structure. Our single crystal diffraction experiments indicate a nonsymmorphic crystal structure, in contrast to the previously reported symmorphic structure. The nonsymmorphic symmetries transform the $k_z=\pi/c$ plane of the Brillouin zone boundary into a node-surface. Band-structure calculations reveal that distinct Fermi surfaces become degenerate on the node-surface and form Dirac lines and a Dirac loop at the Fermi level. Two symmetry related Dirac points remain degenerate under spin-orbit coupling. ARPES measurements confirm the calculations and provide evidence for the Fermi surface degeneracies on the node-surface. These unique topological features enable an unconventional superconducting gap in which time-reversal symmetry can be broken in the absence of other typical ingredients. LaNiGa$_2$ is therefore a topological crystalline superconductor that breaks time-reversal symmetry without any overlapping magnetic ordering or fluctuations. Our findings will enable future discoveries of additional topological superconductors.

We explore a new type of malicious script attacks: the persistent parasite attack. Persistent parasites are stealthy scripts, which persist for a long time in the browser's cache. We show to infect the caches of victims with parasite scripts via TCP injection. Once the cache is infected, we implement methodologies for propagation of the parasites to other popular domains on the victim client as well as to other caches on the network. We show how to design the parasites so that they stay long time in the victim's cache not restricted to the duration of the user's visit to the web site. We develop covert channels for communication between the attacker and the parasites, which allows the attacker to control which scripts are executed and when, and to exfiltrate private information to the attacker, such as cookies and passwords. We then demonstrate how to leverage the parasites to perform sophisticated attacks, and evaluate the attacks against a range of applications and security mechanisms on popular browsers. Finally we provide recommendations for countermeasures., Comment: The paper has been accepted for publication at the 51st Annual IEEE/IFIP International Conference on Dependable Systems and Networks (DSN 2021)

Quantization based model compression serves as high performing and fast approach for inference that yields models which are highly compressed when compared to their full-precision floating point counterparts. The most extreme quantization is a 1-bit representation of parameters such that they have only two possible values, typically -1(0) or +1, enabling efficient implementation of the ubiquitous dot product using only additions. The main contribution of this work is the introduction of a method to smooth the combinatorial problem of determining a binary vector of weights to minimize the expected loss for a given objective by means of empirical risk minimization with backpropagation. This is achieved by approximating a multivariate binary state over the weights utilizing a deterministic and differentiable transformation of real-valued, continuous parameters. The proposed method adds little overhead in training, can be readily applied without any substantial modifications to the original architecture, does not introduce additional saturating nonlinearities or auxiliary losses, and does not prohibit applying other methods for binarizing the activations. Contrary to common assertions made in the literature, it is demonstrated that binary weighted networks can train well with the same standard optimization techniques and similar hyperparameter settings as their full-precision counterparts, specifically momentum SGD with large learning rates and $L_2$ regularization. To conclude experiments demonstrate the method performs remarkably well across a number of inductive image classification tasks with various architectures compared to their full-precision counterparts. The source code is publicly available at https://bitbucket.org/YanivShu/binary_weighted_networks_public.

We define and study LNL polycategories, which abstract the judgmental structure of classical linear logic with exponentials. Many existing structures can be represented as LNL polycategories, including LNL adjunctions, linear exponential comonads, LNL multicategories, IL-indexed categories, linearly distributive categories with storage, commutative and strong monads, CBPV-structures, models of polarized calculi, Freyd-categories, and skew multicategories, as well as ordinary cartesian, symmetric, and planar multicategories and monoidal categories, symmetric polycategories, and linearly distributive and *-autonomous categories. To study such classes of structures uniformly, we define a notion of LNL doctrine, such that each of these classes of structures can be identified with the algebras for some such doctrine. We show that free algebras for LNL doctrines can be presented by a sequent calculus, and that every morphism of doctrines induces an adjunction between their 2-categories of algebras.

The question of which separable C*-algebras have abelian central sequence algebras was raised and studied by Phillips ([Ph88]) and Ando-Kirchberg ([AK14]). In this paper we give a complete answer to their question: A separable C*-algebra $A$ has abelian central sequence algebra if and only if A satisfies Fell's condition. Moreover, we introduce a higher-dimensional analogue of Fell's condition and show that it completely characterizes subhomogeneity of central sequence algebras. In contrast, we show that any non-trivial extension by compact operators has not only non-abelian but not even residually type I central sequence algebra. In particular its central sequence algebra is not type I and not residually finite-dimensional (RFD). Our techniques extensively use properties of nilpotent elements in C*-algebras.

We simulate the polarized radiative transfer in vicinities of the UX Ori type stars during their minima. Our model of an eclipse by an extended disc perturbation generalizes the compact gas-dust cloud eclipse model. We apply the radiative transfer method based on enumeration using the directions grid to model the influence of the perturbation extensions along azimuth and radius on the eclipse depth and parameters of the linear polarization. We investigate eclipses both for the flared disc and for the disc with a puffing-up in the dust sublimation zone. The puffing-up is obtained by adding a disc wind to the model. Comparison with a compact cloud eclipse model reveals that the eclipse by a large-scale azimuthally extended perturbation may be significantly deeper and show a greater linear polarization degree. We also demonstrate that the perturbation extension together with the disc puffing-up can strongly affect the degree of polarization and colour index of the star during the eclipse. The position angle of the linear polarization may also change markedly during and after an eclipse by a large scale perturbation for the model with a puffed-up inner rim. Also, in this model, the maximum degree of the linear polarization can be achieved not at the brightness minimum but closer to the end of the eclipse. We discuss the modelling results in the context of the photopolarimetric observations of UX Ori stars., Comment: 15 pages, 15 figures, 2 tables; Accepted for publication in MNRAS

Without the axiom of choice, the free exact completion of the category of sets (i.e. the category of setoids) may not be complete or cocomplete. We will show that nevertheless, it can be enhanced to a derivator: the formal structure of categories of diagrams related by Kan extension functors. Moreover, this derivator is the free cocompletion of a point in a class of "1-truncated derivators" (which behave like a 1-category rather than a higher category). In classical mathematics, the free cocompletion of a point relative to all derivators is the homotopy theory of spaces. Thus, if there is a homotopy theory that can be shown to have this universal property constructively, its 1-truncation must contain not only sets, but also setoids. This suggests that either setoids are an unavoidable aspect of constructive homotopy theory, or more radical modifications to the notion of homotopy theory are needed., Comment: 42 pages. v2: Clarified discussion of anafunctors and anaderivators

In this paper we pursue three aims. The first one is to apply Amitsur's relations and radicals theory to the study of the lattices Id_{A} of closed two-sided ideals of C*-algebras A. We show that many new and many well-known results about C*-algebras follow naturally from this approach. To use "relation-radical" approach, we consider various subclasses of the class A of all C*-algebras, which we call C*-properties, as they often linked to some properties of C*-algebras. We consider C*-properties P consisting of CCR- and of GCR-algebras; of C*-algebras with continuous trace; of real rank zero, AF, nuclear C*-algebras, etc. Each P defines reflexive relations in all lattices Id_A. Our second aim is to determine the hierarchy and interconnection between properties in A. Our third aim is to study the link between the radicals of relations in the lattices Id_A and the topological radicals on A.

In the English speech-to-text (STT) machine learning task, acoustic models are conventionally trained on uncased Latin characters, and any necessary orthography (such as capitalization, punctuation, and denormalization of non-standard words) is imputed by separate post-processing models. This adds complexity and limits performance, as many formatting tasks benefit from semantic information present in the acoustic signal but absent in transcription. Here we propose a new STT task: end-to-end neural transcription with fully formatted text for target labels. We present baseline Conformer-based models trained on a corpus of 5,000 hours of professionally transcribed earnings calls, achieving a CER of 1.7. As a contribution to the STT research community, we release the corpus free for non-commercial use at https://datasets.kensho.com/datasets/scribe., Comment: 5 pages, 1 figure. Submitted to INTERSPEECH 2021

The Univalence Principle is the statement that equivalent mathematical structures are indistinguishable. We prove a general version of this principle that applies to all set-based, categorical, and higher-categorical structures defined in a non-algebraic and space-based style, as well as models of higher-order theories such as topological spaces. In particular, we formulate a general definition of indiscernibility for objects of any such structure, and a corresponding univalence condition that generalizes Rezk's completeness condition for Segal spaces and ensures that all equivalences of structures are levelwise equivalences. Our work builds on Makkai's First-Order Logic with Dependent Sorts, but is expressed in Voevodsky's Univalent Foundations (UF), extending previous work on the Structure Identity Principle and univalent categories in UF. This enables indistinguishability to be expressed simply as identification, and yields a formal theory that is interpretable in classical homotopy theory, but also in other higher topos models. It follows that Univalent Foundations is a fully equivalence-invariant foundation for higher-categorical mathematics, as intended by Voevodsky., Comment: A short version of this book is available as arXiv:2004.06572. v2: added references and some details on morphisms of premonoidal categories; v3: added proof that exo-nat is sharp if cofibrant, improved presentation following advice from referee

Resource Public Key Infrastructure (RPKI) is vital to the security of inter-domain routing. However, RPKI enables Regional Internet Registries (RIRs) to unilaterally takedown IP prefixes - indeed, such attacks have been launched by nation-state adversaries. The threat of IP prefix takedowns is one of the factors hindering RPKI adoption. In this work, we propose the first distributed RPKI system, based on threshold signatures, that requires the coordination of a number of RIRs to make changes to RPKI objects; hence, preventing unilateral prefix takedown. We perform extensive evaluations using our implementation demonstrating the practicality of our solution. Furthermore, we show that our system is scalable and remains efficient even when RPKI is widely deployed.

We present a unified framework for Petri nets and various variants, such as pre-nets and Kock's whole-grain Petri nets. Our framework is based on a less well-studied notion that we call $\Sigma$-nets, which allow finer control over whether tokens are treated using the collective or individual token philosophy. We describe three forms of execution semantics in which pre-nets generate strict monoidal categories, $\Sigma$-nets (including whole-grain Petri nets) generate symmetric strict monoidal categories, and Petri nets generate commutative monoidal categories, all by left adjoint functors. We also construct adjunctions relating these categories of nets to each other, in particular showing that all kinds of net can be embedded in the unifying category of $\Sigma$-nets, in a way that commutes coherently with their execution semantics., Comment: 29 pages

Modern neural network architectures typically have many millions of parameters and can be pruned significantly without substantial loss in effectiveness which demonstrates they are over-parameterized. The contribution of this work is two-fold. The first is a method for approximating a multivariate Bernoulli random variable by means of a deterministic and differentiable transformation of any real-valued multivariate random variable. The second is a method for model selection by element-wise multiplication of parameters with approximate binary gates that may be computed deterministically or stochastically and take on exact zero values. Sparsity is encouraged by the inclusion of a surrogate regularization to the $L_0$ loss. Since the method is differentiable it enables straightforward and efficient learning of model architectures by an empirical risk minimization procedure with stochastic gradient descent and theoretically enables conditional computation during training. The method also supports any arbitrary group sparsity over parameters or activations and therefore offers a framework for unstructured or flexible structured model pruning. To conclude experiments are performed to demonstrate the effectiveness of the proposed approach.

We study the algebraic structure of the eigenvalues of a Hamiltonian that corresponds to a many-body fermionic system. As the Hamiltonian is quadratic in fermion creation and/or annihilation operators, the system is exactly integrable and the complete single fermion excitation energy spectrum is constructed using the non-interacting fermions that are eigenstates of the quadratic matrix related to the system Hamiltonian. Connection to the Riemann Hypothesis is discussed.

Many applications and protocols depend on the ability to generate a pool of servers to conduct majority-based consensus mechanisms and often this is done by doing plain DNS queries. A recent off-path attack [1] against NTP and security enhanced NTP with Chronos [2] showed that relying on DNS for generating the pool of NTP servers introduces a weak link. In this work, we propose a secure, backward-compatible address pool generation method using distributed DNS-over-HTTPS (DoH) resolvers which is aimed to prevent such attacks against server pool generation.

We demonstrate the first practical off-path time shifting attacks against NTP as well as against Man-in-the-Middle (MitM) secure Chronos-enhanced NTP. Our attacks exploit the insecurity of DNS allowing us to redirect the NTP clients to attacker controlled servers. We perform large scale measurements of the attack surface in NTP clients and demonstrate the threats to NTP due to vulnerable DNS.

The critical role that Network Time Protocol (NTP) plays in the Internet led to multiple efforts to secure it against time-shifting attacks. A recent proposal for enhancing the security of NTP with Chronos against on-path attackers seems the most promising one and is on a standardisation track of the IETF. In this work we demonstrate off-path attacks against Chronos enhanced NTP clients. The weak link is a central security feature of Chronos: The server pool generation mechanism using DNS. We show that the insecurity of DNS allows to subvert the security of Chronos making the time-shifting attacks against Chronos-NTP even easier than attacks against plain NTP.