Loss landscapes are a powerful tool for understanding neural network optimization and generalization, yet traditional low-dimensional analyses often miss complex topological features. We present Landscaper, an open-source Python package for arbitrary-dimensional loss landscape analysis. Landscaper combines Hessian-based subspace construction with topological data analysis to reveal geometric structures such as basin hierarchy and connectivity. A key component is the Saddle-Minimum Average Distance (SMAD) for quantifying landscape smoothness. We demonstrate Landscaper's effectiveness across various architectures and tasks, including those involving pre-trained language models, showing that SMAD captures training transitions, such as landscape simplification, that conventional metrics miss. We also illustrate Landscaper's performance in challenging chemical property prediction tasks, where SMAD can serve as a metric for out-of-distribution generalization, offering valuable insights for model diagnostics and architecture design in data-scarce scientific machine learning scenarios.

Loss landscapes are a powerful tool for understanding neural network optimization and generalization, yet traditional low-dimensional analyses often miss complex topological features. We present Landscaper, an open-source Python package for arbitrary-dimensional loss landscape analysis. Landscaper combines Hessian-based subspace construction with topological data analysis to reveal geometric structures such as basin hierarchy and connectivity. A key component is the Saddle-Minimum Average Distance (SMAD) for quantifying landscape smoothness. We demonstrate Landscaper's effectiveness across various architectures and tasks, including those involving pre-trained language models, showing that SMAD captures training transitions, such as landscape simplification, that conventional metrics miss. We also illustrate Landscaper's performance in challenging chemical property prediction tasks, where SMAD can serve as a metric for out-of-distribution generalization, offering valuable insights for model diagnostics and architecture design in data-scarce scientific machine learning scenarios.

There has been witness to remarkable progress in the ability to (a) grow well-controlled dots; (b) characterize dots on a near-atomic level; and (c) theoretically describe their electronic properties using not only continuum, but also atomistic approaches. This book focuses on the broad scientific and technological interest in semiconductor quantum dots. Papers on growth and self-assembly/self-organization, as well as characterization of semiconductor quantum dots based upon various materials systems, e.g., silicon-germanium, III-V materials, and II-VI materials, are included. Discussions of technological applications, ranging from biomedical technology, microelectronics and photonics, to more esoteric applications including quantum computing, are also featured. Topics include: Si and Ge dots; II-VI and other free-standing (colloidal) dots; near-field spectroscopy of quantum dots, wires and metals; organized dots and dot arrays; transport, coulomb blockade and metallic dots; optical spectroscopy and phonons; light-emitting quantum dots; and structural characterization and growth.

The studies eligible for systematic review. (DOC 444 kb)

Molecular rulers based on Foerster Resonance Energy Transfer (FRET) that report conformational changes and intramolecular distances of single biomolecules have helped to understand important biological processes. However, these rulers suffer from low and fluctuating signal intensities from single dyes and limited observation time due to photobleaching. The plasmon resonance in noble metal particles has been suggested as an alternative probe to overcome the limitations of organic fluorophores and the coupling of plasmons in nearby particles has been exploited to detect particle aggregation by a distinct color change in bulk experiments. Here we demonstrate that plasmon coupling can be used to monitor distances between single pairs of gold and silver nanoparticles. We use this effect to follow the directed assembly of gold and silver nanoparticle dimers in real time and to study the time dynamics of single DNA hybridization events. These ''plasmon rulers'' allowed us to continuously monitor separations of up to 70 nm for more than 3000 seconds. Single molecule in vitro studies of biological processes previously inaccessible with fluorescence based molecular rulers are enabled with plasmon rulers with extended time and distance range.

The kinetics and thermodynamics of structural transformations under pressure depend strongly on particle size due to the influence of surface free energy. By suitable design of surface structure, composition, and passivation it is possible, in principle, to prepare nanocrystals in structures inaccessible to bulk materials. However, few realizations of such extreme size-dependent behavior exist. Here, we show with molecular dynamics computer simulation that in a model of CdSe/ZnS core/shell nanocrystals the core high-pressure structure can be made metastable under ambient conditions by tuning the thickness of the shell. In nanocrystals with thick shells, we furthermore observe a wurtzite to NiAs transformation, which does not occur in the pure bulk materials. These phenomena are linked to a fundamental change in the atomistic transformation mechanism from heterogeneous nucleation at the surface to homogeneous nucleation in the crystal core.

Spezifische, unregelmäßige Anordnungen von Gold-Nanokristallen mit Durchmessern von 5 und 10 nm lassen sich gezielt mit Hilfe doppelsträngiger DNA als Templat herstellen (siehe Bild; A′ und B′ bezeichnen die zu den Sequenzen A bzw. B komplementären Oligonucleotidsequenzen). Die hier beschriebenen Methoden sollten sich auf Nanokristalle aus den verschiedensten Materialien anwenden lassen.

It is crucial to reduce natural gas methane emissions, which can potentially offset the climate benefits of replacing coal with gas. Optical gas imaging (OGI) is a widely-used method to detect methane leaks, but is labor-intensive and cannot provide leak detection results without operators' judgment. In this paper, we develop a computer vision approach to OGI-based leak detection using convolutional neural networks (CNN) trained on methane leak images to enable automatic detection. First, we collect ~1 M frames of labeled video of methane leaks from different leaking equipment for building CNN model, covering a wide range of leak sizes (5.3-2051.6 gCH4/h) and imaging distances (4.6-15.6 m). Second, we examine different background subtraction methods to extract the methane plume in the foreground. Third, we then test three CNN model variants, collectively called GasNet, to detect plumes in videos taken at other pieces of leaking equipment. We assess the ability of GasNet to perform leak detection by comparing it to a baseline method that uses optical-flow based change detection algorithm. We explore the sensitivity of results to the CNN structure, with a moderate-complexity variant performing best across distances. We find that the detection accuracy can reach as high as 99%, the overall detection accuracy can exceed 95% for a case across all leak sizes and imaging distances. Binary detection accuracy exceeds 97% for large leaks (~710 gCH4/h) imaged closely (~5-7 m). At closer imaging distances (~5-10 m), CNN-based models have greater than 94% accuracy across all leak sizes. At farthest distances (~13-16 m), performance degrades rapidly, but it can achieve above 95% accuracy to detect large leaks (>950 gCH4/h). The GasNet-based computer vision approach could be deployed in OGI surveys to allow automatic vigilance of methane leak detection with high detection accuracy in the real world.

Specific, designed, nonperiodic arrangements of gold nanocrystals that are 5 and 10 nm in diameter can be prepared with double-stranded DNA serving as a template (see drawing; A' and B' denote oligonucleotide sequences complementary to sequences A and B). The methods described should be applicable to nanocrystals composed of various materials.

We report the observation of size dependent structural disorder by x-ray absorption near-edge spectroscopy (XANES) in InAs and CdSe nanocrystals 17--80 \AA{} in diameter. XANES of the In and Cd ${M}_{4,5}$ edges yields features that are sharp for the bulk solid but broaden considerably as the size of the particle decreases. FEFF7 multiple-scattering simulations reproduce the size dependent broadening of the spectra if a bulklike surface reconstruction of a spherical nanocrystal model is included. This illustrates that XANES is sensitive to the structure of the entire nanocrystal including the surface.

Abstract Background: Newly discovered malignancies at a relatively younger age (<50) are of great scientific interest. Previous studies found that the incidence of these early-onset cancers showed a global increase, although the pattern varied by cancer type. However, analysis of the Hungarian data has not yet been published. Methods: The Hungarian National Cancer Registry (HNCR) is responsible for data collection of the Hungarian cancer patients. Its operation is population-based in accordance with international standards and covers the entire country. Based on the 10th Revision of International Statistical Classification of Diseases and Related Health Problems, newly discovered cancer cases were extracted from the HNCR’s database. The query focused on patients between the ages of 20 and 49 and period from 2001 to 2019. Next to absolute case numbers, age-standardized values were also analyzed (reference: European Standard Population 2013). Spearman’s correlation test was performed to identify gender- and disease-specific trends. Results: During the studied period, the incidence of early-onset cancers showed decrease among males, while it did not change among females. It should be noted that compared to the total number of newly discovered cancer cases, the proportion of early-onset cancers showed decreasing trend in both genders. Categorization by cancer type revealed that among younger age the incidence of female breast and uterine corpus cancer elevated, while cervical cancer showed a decrease - the latter trend exceeded that of the general population. The incidence of colorectal cancer did not change among females, but decreased among males. The incidence of tobacco-related lung cancer and head and neck region decreased. Conclusion: Compared to analyses enrolled global trends, the Hungarian situation of early-onset cancers seemed to be more complex. On one hand, activities in the past few years such as introduction of HPV vaccination and restrictions on smoking resulted a decrease in the number of related cancers. On the other hand, new mechanisms underlying the increasing types of cancers (e.g. breast cancer) need to be identified. Citation Format: István Kenessey, András Wéber, Mária Dobozi, István Szatmári, Petra Parrag, Péter Nagy, Magdolna Dank, . Incidence of early-onset solid cancers in Hungary in the first two decades of the 21st century based on a population-based registry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3569.

ADVERTISEMENT RETURN TO ISSUEPREVCommunicationNEXTKinetics of II-VI and III-V Colloidal Semiconductor Nanocrystal Growth: "Focusing" of Size DistributionsXiaogang Peng, J. Wickham, and A. P. AlivisatosView Author Information Department of Chemistry, University of California Berkeley and Materials Sciences Division Lawrence Berkeley National Lab Berkeley, California 94720-1460 Cite this: J. Am. Chem. Soc. 1998, 120, 21, 5343–5344Publication Date (Web):May 14, 1998Publication History Received18 February 1998Published online14 May 1998Published inissue 1 June 1998https://pubs.acs.org/doi/10.1021/ja9805425https://doi.org/10.1021/ja9805425rapid-communicationACS PublicationsCopyright © 1998 American Chemical SocietyRequest reuse permissionsArticle Views22646Altmetric-Citations1691LEARN ABOUT THESE METRICSArticle Views are the COUNTER-compliant sum of full text article downloads since November 2008 (both PDF and HTML) across all institutions and individuals. These metrics are regularly updated to reflect usage leading up to the last few days.Citations are the number of other articles citing this article, calculated by Crossref and updated daily. Find more information about Crossref citation counts.The Altmetric Attention Score is a quantitative measure of the attention that a research article has received online. Clicking on the donut icon will load a page at altmetric.com with additional details about the score and the social media presence for the given article. Find more information on the Altmetric Attention Score and how the score is calculated. Share Add toView InAdd Full Text with ReferenceAdd Description ExportRISCitationCitation and abstractCitation and referencesMore Options Share onFacebookTwitterWechatLinked InRedditEmail Other access optionsGet e-AlertscloseSupporting Info (1)»Supporting Information Supporting Information SUBJECTS:Cadmium selenide,Crystallization,Monomers,Nanocrystals,Nanoparticles Get e-Alerts

This paper is the first in a pair presented within the session: Findings from the CFS10 Multi-Hazard Test Program. The emphasis within this article is to, in brevity, describe the scope of a landmark full-scale 10-story cold-formed steel (CFS) framed building tested under multi-hazard (earthquake and fire) scenarios at the UC San Diego 6-DOF Large High-Performance Outdoor Shake Table (LHPOST6). Coined CFS10, this unique building specimen is designed beyond current code height limits, adopting advances in cold-formed steel shear wall detailing, varied construction modalities, and enriched with nonstructural components and systems. The landmark CFS10 building specimen was subjected to extreme multi-hazard (earthquake and fire) loading conditions. This paper sets the framework for presentations to be shared at the Congress, while also aiding in ongoing documentation of findings from the program.

The design of new condensation polymers which undergo acid-catalyzed thermolysis is explored with polycarbonates. The polymers are prepared by phase-transfer catalyzed polycondensation using active esters or carbonates and diols. Polycarbonates containing tertiary, allylic or benzylic diol units susceptible to elimination decompose thermally near 200° to volatile materials. Self developing resist materials can be designed by combining the active polycarbonates with photoactive triarylsulfonium salts or other similar compounds which generate strong acids upon irradiation. Exposure of the resist material creates a latent image which can be developed thermally with evolution of volatile carbon dioxide, alkenes, and alcohols.