Search

This perspective outlines several linear optical effects featured by molecular polaritons arising in the collective strong light-matter coupling regime, focusing on the limit when the number of molecules per photon mode is large. We show that, under these circumstances, molecular absorption within a cavity can be understood as the overlap between the polariton transmission and bare molecular absorption spectra, suggesting that polaritons act in part as optical filters. This framework demystifies and provides a straightforward explanation for a large class of theoretical models of polaritonic phenomena, highlighting that similar effects might be achievable outside a cavity with shaped laser pulses. With a few modifications, this simple conceptual picture can also be adapted to understand the incoherent nonlinear response of polaritonic systems. However, we note that there are experimental observations in the collective regime that exhibit phenomena that go beyond this treatment. Our analysis underscores the importance of the notion that the field still needs to establish a clear distinction between polaritonic phenomena that can be fully explained through classical optics and those that require a more advanced theoretical framework. The linear optics approach presented here is exact when the number of molecules tends to infinity and is quite accurate for a large, but finite, number of molecules. We highlight the limitations of this treatment when the rates of the single-molecule processes that facilitate dark-state-to-polariton relaxation cannot be neglected and in systems under strong coupling with few molecules. Further exploration in these areas is needed to uncover novel polaritonic phenomena.

Radiology reporting is a complex task requiring detailed medical image understanding and precise language generation, for which generative multimodal models offer a promising solution. However, to impact clinical practice, models must achieve a high level of both verifiable performance and utility. We augment the utility of automated report generation by incorporating localisation of individual findings on the image - a task we call grounded report generation - and enhance performance by incorporating realistic reporting context as inputs. We design a novel evaluation framework (RadFact) leveraging the logical inference capabilities of large language models (LLMs) to quantify report correctness and completeness at the level of individual sentences, while supporting the new task of grounded reporting. We develop MAIRA-2, a large radiology-specific multimodal model designed to generate chest X-ray reports with and without grounding. MAIRA-2 achieves state of the art on existing report generation benchmarks and establishes the novel task of grounded report generation., Comment: 72 pages, 21 figures. v2 updates the model and adds results on the PadChest-GR dataset

Surface ion traps with two-dimensional layouts of trapping regions are natural architectures for storing large numbers of ions and supporting the connectivity needed to implement quantum algorithms. Many of the components and operations needed to fully exploit this architecture have already been demonstrated, including operation at cryogenic temperatures with low heating, low excitation transport, and ion control and detection with integrated photonics. Here we demonstrate a trap that addresses the scaling challenge of increasing power dissipation as the RF electrode increases in size. By raising the RF electrode and removing most of the insulating dielectric layer below it we reduce both ohmic and dielectric power dissipation. We also measure heating rates across a range of motional frequencies and for different voltage sources in a trap with a raised RF electrode but solid dielectric., Comment: 7 pages, 4 figures

The vacuum Rabi splitting (VRS) in molecular polaritons stands as a fundamental measure of collective light-matter coupling. Despite its significance, the impact of molecular disorder on VRS is not fully understood yet. This study delves into the complexities of VRS amidst various distributions and degrees of disorder. Our analysis provides precise analytical expressions for linear absorption, transmission, and reflection spectra, along with a "sum" rule, offering a straightforward protocol for extracting accurate collective light-matter coupling values from experimental data. Importantly, our study cautions against directly translating large VRS to the onset of ultrastrong coupling regime. Furthermore, for rectangular disorder, we witness the emergence of narrow side bands alongside a broad central peak, indicating an extended coherence lifetime even in the presence of substantial disorder. These findings not only enhance our understanding of VRS in disordered molecular systems but also open avenues for achieving prolonged coherence lifetimes between the cavity and molecules via the interplay of collective coupling and disorder.

The increasing use of medical imaging in healthcare settings presents a significant challenge due to the increasing workload for radiologists, yet it also offers opportunity for enhancing healthcare outcomes if effectively leveraged. 3D image retrieval holds potential to reduce radiologist workloads by enabling clinicians to efficiently search through diagnostically similar or otherwise relevant cases, resulting in faster and more precise diagnoses. However, the field of 3D medical image retrieval is still emerging, lacking established evaluation benchmarks, comprehensive datasets, and thorough studies. This paper attempts to bridge this gap by introducing a novel benchmark for 3D Medical Image Retrieval (3D-MIR) that encompasses four different anatomies imaged with computed tomography. Using this benchmark, we explore a diverse set of search strategies that use aggregated 2D slices, 3D volumes, and multi-modal embeddings from popular multi-modal foundation models as queries. Quantitative and qualitative assessments of each approach are provided alongside an in-depth discussion that offers insight for future research. To promote the advancement of this field, our benchmark, dataset, and code are made publicly available.

We present a powerful formalism, disordered collective dynamics using truncated equations (d-CUT-E), to simulate the ultrafast quantum dynamics of molecular polaritons in the collective strong coupling regime, where a disordered ensemble of $N\gg10^{6}$ molecules couples to a cavity mode. Notably, we can capture this dynamics with a cavity hosting a single \textit{effective} molecule with $\sim N_{bins}$ electronic states, where $N_{bins}\ll N$ is the number of bins discretizing the disorder distribution. Using d-CUT-E we conclude that strong coupling, as evaluated from linear optical spectra, can be a poor proxy for polariton chemistry. For highly disordered ensembles, total reaction yield upon broadband excitation is identical to that outside of the cavity, while narrowband excitation produces distinct reaction yields solely due to differences in the initial states prepared prior to the reaction., Comment: 16 pages, 12 figures

Graphene flakes were produced by nanosecond plasma discharge at atmospheric pressure between an electrode and the surface of distilled water, in which were placed graphite flakes. The discharge ionizes the gas and forms free radicals on the surface of the water, functionalizing the graphite flakes in solution. The plasma also gives enough energy to break the Van der Waals bonds between the graphene layers but not enough to break the covalent C-C bonds within the layers. Transmission electron microscopy confirmed the hexagonal structure of graphene sheets, and showed that they were monocrystalline. No contamination was found in the obtained nanomaterial. An unknown phenomenon has been found in the activated distilled water, making its electrical conductivity decrease with an increasing temperature. An acidification of the water is observed. The gas in which the discharge takes place plays a major role on the process, no exfoliation is observed if plasmogen argon gas is used., Comment: 13 pages, 22 figures

Highly Oriented Pyrolytic Graphite was exfoliated via pulsed discharge plasma in liquid nitrogen. The potential mechanisms involved were investigated by observing the treated surface of the graphitic material and the obtained particles. Non-exfoliating defects from the plasma treatment were observed and experimental parameter were modified to counteract those. One experiment was performed without exposing the HOPG directly to the discharges so as to better understand the plasma role. The exfoliated particles were observed via TEM and SEM to evaluate the defects, the size, the purity and the crystallinity but no quantitative characterization of their thickness was possible so the actual number of layer of each particle is unknown. Nonetheless, few layers graphene (FLG) was successfully exfoliated through this process. The proposed mechanisms were extrapolated from the observation of the damaged HOPG surface and the obtained particles but the correlation found does not prove causation., Comment: 14 pages, 21 figures

We introduce an analytical model for ideal organic laser diodes based on the argument that their intrinsic active layers necessitate operation in the bipolar space charge-limited current regime. Expressions for the threshold current and voltage agree well with drift-diffusion modeling of complete p-i-n devices and an analytical bound is established for laser operation in the presence of annihilation and excited-state absorption losses. These results establish a foundation for the development of organic laser diode technology.

Individual trapped atomic qubits represent one of the most promising technologies to scale quantum computers, owing to their negligible idle errors and the ability to implement a full set of reconfigurable gate operations via focused optical fields. However, the fidelity of quantum gate operations can be limited by weak confinement of the atoms transverse to the laser. We present measurements of this effect by performing individually-addressed entangling gates in chains of up to 25 trapped atomic ions that are weakly confined along the chain axis. We present a model that accurately describes the observed decoherence from the residual heating of the ions caused by noisy electric fields. We propose to suppress these effects through the use of ancilla ions interspersed in the chain to sympathetically cool the qubit ions throughout a quantum circuit.

Optical microcavities and metallic nanostructures have been shown to significantly modulate the dynamics and spectroscopic response of molecular systems. We present a study of the nonlinear optics of a model consisting of $N$ anharmonic multilevel systems (e.g., Morse oscillators) undergoing collective strong coupling with a resonant infrared microcavity. We find that, under experimentally accessible conditions, molecular systems in microcavities may have nonlinear phenomena significantly intensified due to the high quality of polariton resonances and the enhanced microcavity electromagnetic energy density relative to free space. Particularly large enhancement of multiphoton absorption happens when multipolariton states are resonant with bare molecule multiphoton transitions. In particular, our model predicts two-photon absorption cross section enhancements by several orders of magnitude relative to free space when the Rabi splitting $\Omega_R$ is approximately equal to the molecular anharmonic shift $2\Delta$. Our results provide rough upper bounds to resonant nonlinear response enhancement factors as relaxation to dark states is treated phenomenologically. Notably, ensembles of two-level systems undergoing strong coupling with a cavity (described by the Tavis-Cummings model) show no such optical nonlinearity enhancements, highlighting the rich phenomenology afforded by multilevel anharmonic systems. Similar conclusions are expected to hold for excitonic systems that share features with our model (e.g., molecular dyes with accessible S_0 -> S_1 -> S_2 transitions) and strongly interact with a UV-visible cavity.

Recent progress on few-shot learning largely relies on annotated data for meta-learning: base classes sampled from the same domain as the novel classes. However, in many applications, collecting data for meta-learning is infeasible or impossible. This leads to the cross-domain few-shot learning problem, where there is a large shift between base and novel class domains. While investigations of the cross-domain few-shot scenario exist, these works are limited to natural images that still contain a high degree of visual similarity. No work yet exists that examines few-shot learning across different imaging methods seen in real world scenarios, such as aerial and medical imaging. In this paper, we propose the Broader Study of Cross-Domain Few-Shot Learning (BSCD-FSL) benchmark, consisting of image data from a diverse assortment of image acquisition methods. This includes natural images, such as crop disease images, but additionally those that present with an increasing dissimilarity to natural images, such as satellite images, dermatology images, and radiology images. Extensive experiments on the proposed benchmark are performed to evaluate state-of-art meta-learning approaches, transfer learning approaches, and newer methods for cross-domain few-shot learning. The results demonstrate that state-of-art meta-learning methods are surprisingly outperformed by earlier meta-learning approaches, and all meta-learning methods underperform in relation to simple fine-tuning by 12.8% average accuracy. Performance gains previously observed with methods specialized for cross-domain few-shot learning vanish in this more challenging benchmark. Finally, accuracy of all methods tend to correlate with dataset similarity to natural images, verifying the value of the benchmark to better represent the diversity of data seen in practice and guiding future research., Comment: ECCV 2020. Website: https://www.learning-with-limited-labels.com/

Recent advances in computer vision and deep learning have led to breakthroughs in the development of automated skin image analysis. In particular, skin cancer classification models have achieved performance higher than trained expert dermatologists. However, no attempt has been made to evaluate the consistency in performance of machine learning models across populations with varying skin tones. In this paper, we present an approach to estimate skin tone in benchmark skin disease datasets, and investigate whether model performance is dependent on this measure. Specifically, we use individual typology angle (ITA) to approximate skin tone in dermatology datasets. We look at the distribution of ITA values to better understand skin color representation in two benchmark datasets: 1) the ISIC 2018 Challenge dataset, a collection of dermoscopic images of skin lesions for the detection of skin cancer, and 2) the SD-198 dataset, a collection of clinical images capturing a wide variety of skin diseases. To estimate ITA, we first develop segmentation models to isolate non-diseased areas of skin. We find that the majority of the data in the the two datasets have ITA values between 34.5{\deg} and 48{\deg}, which are associated with lighter skin, and is consistent with under-representation of darker skinned populations in these datasets. We also find no measurable correlation between performance of machine learning model and ITA values, though more comprehensive data is needed for further validation., Comment: NeurIPS 2019 Workshop on Fair ML for Health

This article summarizes the BCN20000 dataset, composed of 19424 dermoscopic images of skin lesions captured from 2010 to 2016 in the facilities of the Hospital Cl\'inic in Barcelona. With this dataset, we aim to study the problem of unconstrained classification of dermoscopic images of skin cancer, including lesions found in hard-to-diagnose locations (nails and mucosa), large lesions which do not fit in the aperture of the dermoscopy device, and hypo-pigmented lesions. The BCN20000 will be provided to the participants of the ISIC Challenge 2019, where they will be asked to train algorithms to classify dermoscopic images of skin cancer automatically., Comment: Abstract for BCN20000

Using machine learning in high-stakes applications often requires predictions to be accompanied by explanations comprehensible to the domain user, who has ultimate responsibility for decisions and outcomes. Recently, a new framework for providing explanations, called TED, has been proposed to provide meaningful explanations for predictions. This framework augments training data to include explanations elicited from domain users, in addition to features and labels. This approach ensures that explanations for predictions are tailored to the complexity expectations and domain knowledge of the consumer. In this paper, we build on this foundational work, by exploring more sophisticated instantiations of the TED framework and empirically evaluate their effectiveness in two diverse domains, chemical odor and skin cancer prediction. Results demonstrate that meaningful explanations can be reliably taught to machine learning algorithms, and in some cases, improving modeling accuracy., Comment: presented at 2019 ICML Workshop on Human in the Loop Learning (HILL 2019), Long Beach, USA. arXiv admin note: substantial text overlap with arXiv:1805.11648

Plasma electrolytic oxidation (PEO) processing of EV 31 magnesium alloy has been carried out in fluoride containing electrolyte under bipolar pulse current regime. Unusual PEO cathodic micro-discharges have been observed and investigated. It is shown that the cathodic micro-discharges exhibit a collective intermittent behavior which is discussed in terms of charge accumulations at the layer/electrolyte and layer/metal interfaces. Optical emission spectroscopy is used to determine the electron density (typ. 10 15 cm-3) and the electron temperature (typ. 7500 K) while the role of F-anions on the appearance of cathodic micro-discharges is pointed out. Plasma Electrolytic Oxidation (PEO) is a promising plasma-assisted surface treatment of light metallic alloys (e.g. Al, Mg, Ti). Although the PEO process makes it possible to grow oxide coatings with interesting corrosion and wear resistant properties, the physical mechanisms of coating growth are not yet completely understood. Typically, the process consists in applying a high voltage difference between a metallic piece and a counter-electrode which are both immersed in an electrolyte bath. Compare to anodizing, the main differences concern the electrolyte composition and the current and voltage ranges which are at least one order of magnitude higher in PEO 1. These significant differences in current and voltage imply the dielectric breakdown and consequently the appearance of micro-discharges on the surface of the sample under processing. Those micro-discharges are recognized as being the main contributors to the formation of a dielectric porous crystalline oxide coating. 2 Nevertheless, the breakdown mechanism that governs the appearance of those micro-discharges is still under investigation. Hussein et al. 3 proposed a mechanism with three different plasma formation processes based on differences in plasma chemical composition. The results of Jovovi{\'c} et al. 4,5 concerning physical properties of the plasma seem to corroborate this mechanism, and also point out the importance of the substrate material in the plasma composition. 6 Compared with DC conducted PEO process, using a bipolar pulsed DC or AC current supply gives supplementary control latitude through the current waveform parameters. The effect of these parameter on the micro-discharges behavior has been investigated in several previous works. 2,3,7,8 One of the main results of these studies is the absence of micro-discharge during the cathodic current half-period. 9-11 Even if the cathodic half-period has an obvious effect on the efficiency of PEO as well as on the coating growth and composition, the micro-plasmas appear only in anodic half-period. Sah et al. 8 have observed the cathodic breakdown of an oxide layer but at very high current density (10 kA.dm-${}^2$), and after several steps of sample preparation. Several models of micro-discharges appearance in AC current have already been proposed. 1,2,8,12,13 Though cathodic micro-discharges have never been observed within usual process conditions, the present study aims at defining suitable conditions to promote cathodic micro-discharges and at studying the main characteristics of these micro-plasmas.

Artificial intelligence systems are being increasingly deployed due to their potential to increase the efficiency, scale, consistency, fairness, and accuracy of decisions. However, as many of these systems are opaque in their operation, there is a growing demand for such systems to provide explanations for their decisions. Conventional approaches to this problem attempt to expose or discover the inner workings of a machine learning model with the hope that the resulting explanations will be meaningful to the consumer. In contrast, this paper suggests a new approach to this problem. It introduces a simple, practical framework, called Teaching Explanations for Decisions (TED), that provides meaningful explanations that match the mental model of the consumer. We illustrate the generality and effectiveness of this approach with two different examples, resulting in highly accurate explanations with no loss of prediction accuracy for these two examples., Comment: This article leverages some content from arXiv:1805.11648; presented at ACM/AAAI AIES'19