Hasil untuk "Modern"

Krzysztof Adamkiewicz, Brian Moser, Stanislav Frolov et al.

Recent text-to-image (T2I) diffusion models produce visually stunning images and demonstrate excellent prompt following. But do they perform well as synthetic vision data generators? In this work, we revisit the promise of synthetic data as a scalable substitute for real training sets and uncover a surprising performance regression. We generate large-scale synthetic datasets using state-of-the-art T2I models released between 2022 and 2025, train standard classifiers solely on this synthetic data, and evaluate them on real test data. Despite observable advances in visual fidelity and prompt adherence, classification accuracy on real test data consistently declines with newer T2I models as training data generators. Our analysis reveals a hidden trend: These models collapse to a narrow, aesthetic-centric distribution that undermines diversity and real data distribution coverage. Overall, our findings challenge a growing assumption in vision research, namely that progress in generative realism implies progress in data realism. We thus highlight an urgent need to rethink the capabilities of modern T2I models as reliable training data generators.

Anish Lakkapragada

Classical statistical inference and learning theory often fail to explain the success of modern neural networks. A key reason is that these models are non-identifiable (singular), violating core assumptions behind PAC bounds and asymptotic normality. Singular learning theory (SLT), a physics-inspired framework grounded in algebraic geometry, has gained popularity for its ability to close this theory-practice gap. In this paper, we empirically study SLT in toy settings relevant to interpretability and phase transitions. First, we understand the SLT free energy $\mathcal{F}_n$ by testing an Arrhenius-style rate hypothesis using both a grokking modulo-arithmetic model and Anthropic's Toy Models of Superposition. Second, we understand the local learning coefficient $λ_α$ by measuring how it scales with problem difficulty across several controlled network families (polynomial regressors, low-rank linear networks, and low-rank autoencoders). Our experiments recover known scaling laws while others yield meaningful deviations from theoretical expectations. Overall, our paper illustrates the many merits of SLT for understanding neural network phase transitions, and poses open research questions for the field.

Christodoulos Peltekis, Chrysostomos Nicopoulos, Giorgos Dimitrakopoulos

Modern Machine Learning (ML) applications often benefit from structured sparsity, a technique that efficiently reduces model complexity and simplifies handling of sparse data in hardware. Sparse systolic tensor arrays - specifically designed to accelerate these structured-sparse ML models - play a pivotal role in enabling efficient computations. As ML is increasingly integrated into safety-critical systems, it is of paramount importance to ensure the reliability of these systems. This paper introduces an online error-checking technique capable of detecting and locating permanent faults within sparse systolic tensor arrays before computation begins. The new technique relies on merely four test vectors and exploits the weight values already loaded within the systolic array to comprehensively test the system. Fault-injection campaigns within the gate-level netlist, while executing three well-established Convolutional Neural Networks (CNN), validate the efficiency of the proposed approach, which is shown to achieve very high fault coverage, while incurring minimal performance and area overheads.

Arash Dehghani, Hossein Saberi

This paper reviews the state-of-the-art in deepfake generation and detection, focusing on modern deep learning technologies and tools based on the latest scientific advancements. The rise of deepfakes, leveraging techniques like Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), Diffusion models and other generative models, presents significant threats to privacy, security, and democracy. This fake media can deceive individuals, discredit real people and organizations, facilitate blackmail, and even threaten the integrity of legal, political, and social systems. Therefore, finding appropriate solutions to counter the potential threats posed by this technology is essential. We explore various deepfake methods, including face swapping, voice conversion, reenactment and lip synchronization, highlighting their applications in both benign and malicious contexts. The review critically examines the ongoing "arms race" between deepfake generation and detection, analyzing the challenges in identifying manipulated contents. By examining current methods and highlighting future research directions, this paper contributes to a crucial understanding of this rapidly evolving field and the urgent need for robust detection strategies to counter the misuse of this powerful technology. While focusing primarily on audio, image, and video domains, this study allows the reader to easily grasp the latest advancements in deepfake generation and detection.

Ora Nova Fandina, Gal Amram, Eitan Farchi et al.

Application modernization in legacy languages such as COBOL, PL/I, and REXX faces an acute shortage of resources, both in expert availability and in high-quality human evaluation data. While Large Language Models as a Judge (LaaJ) offer a scalable alternative to expert review, their reliability must be validated before being trusted in high-stakes workflows. Without principled validation, organizations risk a circular evaluation loop, where unverified LaaJs are used to assess model outputs, potentially reinforcing unreliable judgments and compromising downstream deployment decisions. Although various automated approaches to validating LaaJs have been proposed, alignment with human judgment remains a widely used and conceptually grounded validation strategy. In many real-world domains, the availability of human-labeled evaluation data is severely limited, making it difficult to assess how well a LaaJ aligns with human judgment. We introduce SparseAlign, a formal framework for assessing LaaJ alignment with sparse human-labeled data. SparseAlign combines a novel pairwise-confidence concept with a score-sensitive alignment metric that jointly capture ranking consistency and score proximity, enabling reliable evaluator selection even when traditional statistical methods are ineffective due to limited annotated examples. SparseAlign was applied internally to select LaaJs for COBOL code explanation. The top-aligned evaluators were integrated into assessment workflows, guiding model release decisions. We present a case study of four LaaJs to demonstrate SparseAlign's utility in real-world evaluation scenarios.

K. Mursula, A. A. Pevtsov, T. Asikainen et al.

The Sun experienced a period of unprecedented activity during the 20th century, now called the Modern Maximum (MM). The decay of the MM after cycle 19 has changed the Sun, the heliosphere, and the planetary environments in many ways. However, studies disagree on whether this decay has proceeded synchronously in different solar parameters or not. One key issue is if the relation between two long parameters of solar activity, the sunspot number and the solar 10.7cm radio flux, has remained the same during this decay. A recent study argues that there is an inhomogeneity in the 10.7cm radio flux in 1980, which leads to a step-like jump ("1980 jump") in this relation. Here we show that the relation between sunspot number and 10.7cm radio flux varies in time, not due to an inhomogeneous radio flux but due to physical changes in the solar atmosphere. We used radio fluxes at four different wavelengths measured in Japan, and studied their long-term relation with the sunspot number and the 10.7cm radio flux. We also used two other solar parameters, the MgII index and the number of active regions. We find that the 1980 jump is only the first of a series of 1-2-year "humps" that mainly occur during solar maxima. All radio fluxes increase with respect to the sunspot number from the 1970s to 2010s. These results reestablish the 10.7cm flux as a homogeneous measure of solar activity. The fluxes of the longer radio waves are found to increase with respect to the shorter waves, which suggests a long-term change in the solar radio spectrum. We also find that the MgII index and the number of active regions also increased with respect to the sunspot number, further verifying the difference in the long-term evolution in chromospheric and photospheric parameters. Our results provide evidence for important structural changes in solar magnetic fields and the solar atmosphere during the decay of the MM.

Yufan Li, Pragya Sur

Debiasing is a fundamental concept in high-dimensional statistics. While degrees-of-freedom adjustment is the state-of-the-art technique in high-dimensional linear regression, it is limited to i.i.d. samples and sub-Gaussian covariates. These constraints hinder its broader practical use. Here, we introduce Spectrum-Aware Debiasing--a novel method for high-dimensional regression. Our approach applies to problems with structured dependencies, heavy tails, and low-rank structures. Our method achieves debiasing through a rescaled gradient descent step, deriving the rescaling factor using spectral information of the sample covariance matrix. The spectrum-based approach enables accurate debiasing in much broader contexts. We study the common modern regime where the number of features and samples scale proportionally. We establish asymptotic normality of our proposed estimator (suitably centered and scaled) under various convergence notions when the covariates are right-rotationally invariant. Such designs have garnered recent attention due to their crucial role in compressed sensing. Furthermore, we devise a consistent estimator for its asymptotic variance. Our work has two notable by-products: first, we use Spectrum-Aware Debiasing to correct bias in principal components regression (PCR), providing the first debiased PCR estimator in high dimensions. Second, we introduce a principled test for checking alignment between the signal and the eigenvectors of the sample covariance matrix. This test is independently valuable for statistical methods developed using approximate message passing, leave-one-out, or convex Gaussian min-max theorems. We demonstrate our method through simulated and real data experiments. Technically, we connect approximate message passing algorithms with debiasing and provide the first proof of the Cauchy property of vector approximate message passing (V-AMP).

Da Xu, Chuanwei Ruan

In the relatively short history of machine learning, the subtle balance between engineering and theoretical progress has been proved critical at various stages. The most recent wave of AI has brought to the IR community powerful techniques, particularly for pattern recognition. While many benefits from the burst of ideas as numerous tasks become algorithmically feasible, the balance is tilting toward the application side. The existing theoretical tools in IR can no longer explain, guide, and justify the newly-established methodologies. The consequences can be suffering: in stark contrast to how the IR industry has envisioned modern AI making life easier, many are experiencing increased confusion and costs in data manipulation, model selection, monitoring, censoring, and decision making. This reality is not surprising: without handy theoretical tools, we often lack principled knowledge of the pattern recognition model's expressivity, optimization property, generalization guarantee, and our decision-making process has to rely on over-simplified assumptions and human judgments from time to time. Time is now to bring the community a systematic tutorial on how we successfully adapt those tools and make significant progress in understanding, designing, and eventually productionize impactful IR systems. We emphasize systematicity because IR is a comprehensive discipline that touches upon particular aspects of learning, causal inference analysis, interactive (online) decision-making, etc. It thus requires systematic calibrations to render the actual usefulness of the imported theoretical tools to serve IR problems, as they usually exhibit unique structures and definitions. Therefore, we plan this tutorial to systematically demonstrate our learning and successful experience of using advanced theoretical tools for understanding and designing IR systems.

Batel Oved, Orian Leitersdorf, Ronny Ronen et al.

Recent research has sought to accelerate cryptographic hash functions as they are at the core of modern cryptography. Traditional designs, however, suffer from the von Neumann bottleneck that originates from the separation of processing and memory units. An emerging solution to overcome this bottleneck is processing-in-memory (PIM): performing logic within the same devices responsible for memory to eliminate data-transfer and simultaneously provide massive computational parallelism. In this paper, we seek to vastly accelerate the state-of-the-art SHA-3 cryptographic function using the memristive memory processing unit (mMPU), a general-purpose memristive PIM architecture. To that end, we propose a novel in-memory algorithm for variable rotation, and utilize an efficient mapping of the SHA-3 state vector for memristive crossbar arrays to efficiently exploit PIM parallelism. We demonstrate a massive energy efficiency of 1,422 Gbps/W, improving a state-of-the-art memristive SHA-3 accelerator (SHINE-2) by 4.6x.

Satwik Patnaik, Mohammed Ashraf, Ozgur Sinanoglu et al.

Split manufacturing (SM) and layout camouflaging (LC) are two promising techniques to obscure integrated circuits (ICs) from malicious entities during and after manufacturing. While both techniques enable protecting the intellectual property (IP) of ICs, SM can further mitigate the insertion of hardware Trojans (HTs). In this paper, we strive for the "best of both worlds," that is we seek to combine the individual strengths of SM and LC. By jointly extending SM and LC techniques toward 3D integration, an up-and-coming paradigm based on stacking and interconnecting of multiple chips, we establish a modern approach to hardware security. Toward that end, we develop a security-driven CAD and manufacturing flow for 3D ICs in two variations, one for IP protection and one for HT prevention. Essential concepts of that flow are (i) "3D splitting" of the netlist to protect, (ii) obfuscation of the vertical interconnects (i.e., the wiring between stacked chips), and (iii) for HT prevention, a security-driven synthesis stage. We conduct comprehensive experiments on DRC-clean layouts of multi-million-gate DARPA and OpenCores designs (and others). Strengthened by extensive security analysis for both IP protection and HT prevention, we argue that entering the third dimension is eminent for effective and efficient hardware security.

Md. Khairul Islam, Toufique Ahmed, Rifat Shahriyar et al.

The modern code review process is an integral part of the current software development practice. Considerable effort is given here to inspect code changes, find defects, suggest an improvement, and address the suggestions of the reviewers. In a code review process, usually, several iterations take place where an author submits code changes and a reviewer gives feedback until is happy to accept the change. In around 12% cases, the changes are abandoned, eventually wasting all the efforts. In this research, our objective is to design a tool that can predict whether a code change would be merged or abandoned at an early stage to reduce the waste of efforts of all stakeholders (e.g., program author, reviewer, project management, etc.) involved. The real-world demand for such a tool was formally identified by a study by Fan et al. [1]. We have mined 146,612 code changes from the code reviews of three large and popular open-source software and trained and tested a suite of supervised machine learning classifiers, both shallow and deep learning based. We consider a total of 25 features in each code change during the training and testing of the models. The best performing model named PredCR (Predicting Code Review), a LightGBM-based classifier achieves around 85% AUC score on average and relatively improves the state-of-the-art [1] by 14-23%. In our empirical study on the 146,612 code changes from the three software projects, we find that (1) The new features like reviewer dimensions that are introduced in PredCR are the most informative. (2) Compared to the baseline, PredCR is more effective towards reducing bias against new developers. (3) PredCR uses historical data in the code review repository and as such the performance of PredCR improves as a software system evolves with new and more data.

Frederik De Ceuster, Ward Homan, Jeremy Yates et al.

Radiative transfer is a key component in almost all astrophysical and cosmological simulations. We present Magritte: a modern open-source software library for 3D radiative transfer. It uses a deterministic ray-tracer and formal solver, i.e. it computes the radiation field by tracing rays through the model and solving the radiative transfer equation in its second-order form along a fixed set of rays originating from each point. Magritte can handle structured and unstructured input meshes, as well as smoothed-particle hydrodynamics (SPH) particle data. In this first paper, we describe the numerical implementation, semi-analytic tests and cross-code benchmarks for the non-LTE line radiative transfer module of Magritte. This module uses the radiative transfer solver to self-consistently determine the populations of the quantised energy levels of atoms and molecules using an accelerated Lambda iteration (ALI) scheme. We compare Magritte with the established radiative transfer solvers Ratran (1D) and Lime (3D) on the van Zadelhoff benchmark and present a first application to a simple Keplerian disc model. Comparing with Lime, we conclude that Magritte produces more accurate and more precise results, especially at high optical depth, and that it is faster.

Norm M. Tubman, C. Daniel Freeman, Daniel S. Levine et al.

Recent advances in selected CI, including the adaptive sampling configuration interaction (ASCI) algorithm and its heat bath extension, have made the ASCI approach competitive with the most accurate techniques available, and hence an increasingly powerful tool in solving quantum Hamiltonians. In this work, we show that a useful paradigm for generating efficient selected CI/exact diagonalization algorithms is driven by fast sorting algorithms, much in the same way iterative diagonalization is based on the paradigm of matrix vector multiplication. We present several new algorithms for all parts of performing a selected CI, which includes new ASCI search, dynamic bit masking, fast orbital rotations, fast diagonal matrix elements, and residue arrays. The algorithms presented here are fast and scalable, and we find that because they are built on fast sorting algorithms they are more efficient than all other approaches we considered. After introducing these techniques we present ASCI results applied to a large range of systems and basis sets in order to demonstrate the types of simulations that can be practically treated at the full-CI level with modern methods and hardware, presenting double- and triple-zeta benchmark data for the G1 dataset. The largest of these calculations is Si$_{2}$H$_{6}$ which is a simulation of 34 electrons in 152 orbitals. We also present some preliminary results for fast deterministic perturbation theory simulations that use hash functions to maintain high efficiency for treating large basis sets.

Min Xu, Yongxian Wang, Anthony Theodore Chronopoulos et al.

As one of open-source codes widely used in computational ocean acoustics, FOR3D can provide a very good estimate for underwater acoustic propagation. In this paper, we propose a performance optimization and parallelization to speed up the running of FOR3D. We utilized a variety of methods to enhance the entire performance, such as using a multi-threaded programming model to exploit the potential capability of the many-core node of high-performance computing (HPC) system, tuning compile options, using efficient tuned mathematical library and utilizing vectorization optimization instruction. In addition, we extended the application from single-frequency calculation to multi-frequency calculation successfully by using OpenMP+MPI hybrid programming techniques on the mainstream HPC platform. A detailed performance evaluation was performed and the results showed that the proposed parallelization obtained good accelerated effect of 25.77X when testing a typical three-dimensional medium-sized case on Tianhe-2 supercomputer. It also showed that the tuned parallel version has a weak-scalability. The speed of calculation of underwater sound field can be greatly improved by the strategy mentioned in this paper. The method used in this paper is not only applicable to other similar computing models in computational ocean acoustics but also a guideline of performance enhancement for scientific and engineering application running on modern many-core-computing platform.

Alexey Golovnev

This is a contribution to the Proceedings of the 7th Mathematical Physics Meeting: Summer School and Conference on Modern Mathematical Physics, held in Belgrade 09 -- 19 September 2012. We give an easily accessible introduction to the ADM decomposition of the curvature components. After that we review the basic problems associated with attempts of constructing a viable massive gravity theory. And finally, we present the metric formulations of ghost-free massive gravity models, and comment on existence problem of the matrix square root.

Yu. L. Bolotin, A. Kostenko, O. A. Lemets et al.

In this review we consider in detail different theoretical topics associated with interaction in the dark sector. We study linear and nonlinear interactions which depend on the dark matter and dark energy densities. We consider a number of different models (including the holographic dark energy and dark energy in a fractal universe) with interacting dark energy (DE) and dark matter (DM), have done a thorough analysis of these models. The main task of this review was not only to give an idea about the modern set of different models of dark energy, but to show how much can be diverse dynamics of the universe in these models. We find that the dynamics of a Universe that contains interaction in the dark sector can differ significantly from the Standard Cosmological Model (SCM).

Rosemary A. Mardling

Modern applications of celestial mechanics include the study of closely packed systems of exoplanets, circumbinary planetary systems, binary-binary interactions in star clusters, and the dynamics of stars near the galactic centre. While developments have historically been guided by the architecture of the Solar System, the need for more general formulations with as few restrictions on the parameters as possible is obvious. Here we present clear and concise generalisations of two classic expansions of the three-body disturbing function, simplifying considerably their original form and making them accessible to the non-specialist. Governing the interaction between the inner and outer orbits of a hierarchical triple, the disturbing function in its general form is the conduit for energy and angular momentum exchange and as such, governs the secular and resonant evolution of the system and its stability characteristics. Focusing here on coplanar systems, the first expansion is one in the ratio of inner to outer semimajor axes and is valid for all eccentricities, while the second is an expansion in eccentricity and is valid for all semimajor axis ratios [...]. Our generalizations make both formulations valid for arbitrary mass ratios. [...]. We demonstrate the equivalence of the new expansions, identifying the role of the spherical harmonic order m in both and its physical significance in the three-body problem, and introducing the concept of principal resonances. Several examples of the accessibility of both expansions are given including resonance widths and the secular rates of change of the elements. Results in their final form are gathered together at the end of the paper for the reader mainly interested in their application, including a guide for the choice of expansion.

Marcelo Gleiser, Sara Imari Walker

The search for life elsewhere in the universe is a pivotal question in modern science. However, to address whether life is common in the universe we must first understand the likelihood of abiogenesis by studying the origin of life on Earth. A key missing piece is the origin of biomolecular homochirality: permeating almost every life-form on Earth is the presence of exclusively levorotary amino acids and dextrorotary sugars. In this work we discuss recent results suggesting that life's homochirality resulted from sequential chiral symmetry breaking triggered by environmental events in a mechanism referred to as punctuated chirality. Applying these arguments to other potentially life-bearing platforms has significant implications for the search for extraterrestrial life: we predict that a statistically representative sampling of extraterrestrial stereochemistry will be racemic on average.

P. Shabajee, K. Postlethwaite

Relativity, Quantum Mechanics and Chaos theory are three of the most significant scientific advances of the 20th Century - each fundamentally changing our understanding of the physical universe. The authors ask why the UK National Curriculum in science almost entirely ignores them. Children and young people regularly come into contact with the language, concepts and implications of these theories through the media and through new technologies, and they are the basis of many contemporary scientific and technological developments. There is surely, therefore, an urgent need to include the concepts of '20th Century physics' within the curriculum.

S. E. Emelin, A. L. Pirozerski

The brief characteristic of a place and state of modern experimental investigations of a ball lightning on the basis of metastable substance is given.