We investigate the gas jet breakup and the resulting microbubble formation in a microfluidic flow-focusing device using ultra high-speed imaging at 1 million frames/s. In recent experiments [Dollet et al., Phys. Rev. Lett. 100, 034504 (2008)] it was found that in the final stage of the collapse the radius of the neck scales with time with a 1/3 power-law exponent, which suggested that gas inertia and the Bernoulli suction effect become important. Here, ultra high-speed imaging was used to capture the complete bubble contour and quantify the gas flow through the neck. It revealed that the resulting decrease in pressure, due to Bernoulli suction, is too low to account for an accelerated pinch-off. The high temporal resolution images enable us to approach the final moment of pinch-off to within 1 {\mu}s. We observe that the final moment of bubble pinch-off is characterized by a scaling exponent of 0.41 +/- 0.01. This exponent is approximately 2/5, which can be derived, based on the observation that during the collapse the neck becomes less slender, due to the exclusive driving through liquid inertia.

Wim van Hoeve Benjamin Dollet Michel Versluis Detlef Lohse

The electronic properties of two-dimensional (2D) materials depend sensitively on the underlying atomic arrangement down to the monolayer level. Here we present a novel strategy for the determination of the band gap and complex dielectric function in 2D materials achieving a spatial resolution down to a few nanometers. This approach is based on machine learning techniques developed in particle physics and makes possible the automated processing and interpretation of spectral images from electron energy-loss spectroscopy (EELS). Individual spectra are classified as a function of the thickness with $K$-means clustering and then used to train a deep-learning model of the zero-loss peak background. As a proof-of-concept we assess the band gap and dielectric function of InSe flakes and polytypic WS$_2$ nanoflowers, and correlate these electrical properties with the local thickness. Our flexible approach is generalizable to other nanostructured materials and to higher-dimensional spectroscopies, and is made available as a new release of the open-source EELSfitter framework.

Abel Brokkelkamp Jaco ter Hoeve Isabel Postmes Sabrya E. van Heijst Louis Maduro Albert V. Davydov Sergiy Krylyuk Juan Rojo Sonia Conesa-Boj

In this paper, the size of bubbles formed through the breakup of a gaseous jet in a co-axial microfluidic device is derived. The gaseous jet surrounded by a co-flowing liquid stream breaks up into monodisperse microbubbles and the size of the bubbles is determined by the radius of the inner gas jet and the bubble formation frequency. We obtain the radius of the gas jet by solving the Navier-Stokes equations for low Reynolds number flows and by minimization of the dissipation energy. The prediction of the bubble size is based on the system's control parameters only, i.e. the inner gas flow rate $Q_i$, the outer liquid flow rate $Q_o$, and the tube radius $R$. For a very low gas-to-liquid flow rate ratio ($Q_i / Q_o \rightarrow 0$) the bubble radius scales as $r_b / R \propto \sqrt{Q_i / Q_o}$, independently of the inner to outer viscosity ratio $\eta_i/\eta_o$ and of the type of the velocity profile in the gas, which can be either flat or parabolic, depending on whether high-molecular-weight surfactants cover the gas-liquid interface or not. However, in the case in which the gas velocity profiles are parabolic and the viscosity ratio is sufficiently low, i.e. $\eta_i/\eta_o \ll 1$, the bubble diameter scales as $r_b \propto (Q_i/Q_o)^\beta$, with $\beta$ smaller than 1/2.

Wim van Hoeve Benjamin Dollet José M. Gordillo Michel Versluis Detlef Lohse

The role of conversational assistants has become more prevalent in helping people increase their productivity. Document-centered assistance, for example to help an individual quickly review a document, has seen less significant progress, even though it has the potential to tremendously increase a user's productivity. This type of document-centered assistance is the focus of this paper. Our contributions are three-fold: (1) We first present a survey to understand the space of document-centered assistance and the capabilities people expect in this scenario. (2) We investigate the types of queries that users will pose while seeking assistance with documents, and show that document-centered questions form the majority of these queries. (3) We present a set of initial machine learned models that show that (a) we can accurately detect document-centered questions, and (b) we can build reasonably accurate models for answering such questions. These positive results are encouraging, and suggest that even greater results may be attained with continued study of this interesting and novel problem space. Our findings have implications for the design of intelligent systems to support task completion via natural interactions with documents.

Maartje ter Hoeve Robert Sim Elnaz Nouri Adam Fourney Maarten de Rijke Ryen W. White

The density of mitotic figures within tumor tissue is known to be highly correlated with tumor proliferation and thus is an important marker in tumor grading. Recognition of mitotic figures by pathologists is known to be subject to a strong inter-rater bias, which limits the prognostic value. State-of-the-art deep learning methods can support the expert in this assessment but are known to strongly deteriorate when applied in a different clinical environment than was used for training. One decisive component in the underlying domain shift has been identified as the variability caused by using different whole slide scanners. The goal of the MICCAI MIDOG 2021 challenge has been to propose and evaluate methods that counter this domain shift and derive scanner-agnostic mitosis detection algorithms. The challenge used a training set of 200 cases, split across four scanning systems. As a test set, an additional 100 cases split across four scanning systems, including two previously unseen scanners, were given. The best approaches performed on an expert level, with the winning algorithm yielding an F_1 score of 0.748 (CI95: 0.704-0.781). In this paper, we evaluate and compare the approaches that were submitted to the challenge and identify methodological factors contributing to better performance.

Marc Aubreville Nikolas Stathonikos Christof A. Bertram Robert Klopleisch Natalie ter Hoeve Francesco Ciompi Frauke Wilm Christian Marzahl Taryn A. Donovan Andreas Maier Jack Breen Nishant Ravikumar Youjin Chung Jinah Park Ramin Nateghi Fattaneh Pourakpour Rutger H. J. Fick Saima Ben Hadj Mostafa Jahanifar Nasir Rajpoot Jakob Dexl Thomas Wittenberg Satoshi Kondo Maxime W. Lafarge Viktor H. Koelzer Jingtang Liang Yubo Wang Xi Long Jingxin Liu Salar Razavi April Khademi Sen Yang Xiyue Wang Mitko Veta Katharina Breininger

We discuss the motion of neutral and charged particles in a plane electro-magnetic wave and its accompanying gravitational field.

M. Faquir J. W. van Holten

We present a covariant and supersymmetric theory of relativistic hydrodynamics in four-dimensional Minkowski space.

T. S. Nyawelo J. W. van Holten S. Groot Nibbelink

Theoretical interpretations of particle physics data, such as the determination of the Wilson coefficients of the Standard Model Effective Field Theory (SMEFT), often involve the inference of multiple parameters from a global dataset. Optimizing such interpretations requires the identification of observables that exhibit the highest possible sensitivity to the underlying theory parameters. In this work we develop a flexible open source framework, ML4EFT, enabling the integration of unbinned multivariate observables into global SMEFT fits. As compared to traditional measurements, such observables enhance the sensitivity to the theory parameters by preventing the information loss incurred when binning in a subset of final-state kinematic variables. Our strategy combines machine learning regression and classification techniques to parameterize high-dimensional likelihood ratios, using the Monte Carlo replica method to estimate and propagate methodological uncertainties. As a proof of concept we construct unbinned multivariate observables for top-quark pair and Higgs+$Z$ production at the LHC, demonstrate their impact on the SMEFT parameter space as compared to binned measurements, and study the improved constraints associated to multivariate inputs. Since the number of neural networks to be trained scales quadratically with the number of parameters and can be fully parallelized, the ML4EFT framework is well-suited to construct unbinned multivariate observables which depend on up to tens of EFT coefficients, as required in global fits.

Raquel Gomez Ambrosio Jaco ter Hoeve Maeve Madigan Juan Rojo Veronica Sanz

We give a description of the individual Ekedahl-Oort strata contained in the supersingular locus in terms of Deligne-Lusztig varieties, refining a result of Harashita.

Maarten Hoeve

Learning algorithms become more powerful, often at the cost of increased complexity. In response, the demand for algorithms to be transparent is growing. In NLP tasks, attention distributions learned by attention-based deep learning models are used to gain insights in the models' behavior. To which extent is this perspective valid for all NLP tasks? We investigate whether distributions calculated by different attention heads in a transformer architecture can be used to improve transparency in the task of abstractive summarization. To this end, we present both a qualitative and quantitative analysis to investigate the behavior of the attention heads. We show that some attention heads indeed specialize towards syntactically and semantically distinct input. We propose an approach to evaluate to which extent the Transformer model relies on specifically learned attention distributions. We also discuss what this implies for using attention distributions as a means of transparency.

Joris Baan Maartje ter Hoeve Marlies van der Wees Anne Schuth Maarten de Rijke