Search

The softer a material is, the more its mechanics are sensitive to interfaces. In soft gels, an elastic polymeric network is filled with free-flowing molecules. In theory, either of these components could dominate the material interfacial properties. In practice, current measurements cannot distinguish between the two, nor can they rule out material inhomogeneities, which could modulate the apparent properties of the interfaces. Here, we introduce an experimental approach that elucidates the interfacial mechanics of soft solids. Coupling quantum dots, controlled deformations, and precise confocal measurements, we fully separate the material inhomogeneities of a silicone gel from its true interfacial properties. We quantify a gradient in bulk elastic properties near the surface, with a characteristic length scale of about 20 {\mu}m. In addition, we observe a surface excess elasticity, whose associated gradient is unresolvable with light microscopy. The composition of the external medium has a strong affect on the observed value of the surface elasticity. Thus, we conclude that the surface elasticity of this silicone network is an interfacial property., Comment: 8 pages, 3 figures

Our objective is to generate Audio Descriptions (ADs) for both movies and TV series in a training-free manner. We use the power of off-the-shelf Visual-Language Models (VLMs) and Large Language Models (LLMs), and develop visual and text prompting strategies for this task. Our contributions are three-fold: (i) We demonstrate that a VLM can successfully name and refer to characters if directly prompted with character information through visual indications without requiring any fine-tuning; (ii) A two-stage process is developed to generate ADs, with the first stage asking the VLM to comprehensively describe the video, followed by a second stage utilising a LLM to summarise dense textual information into one succinct AD sentence; (iii) A new dataset for TV audio description is formulated. Our approach, named AutoAD-Zero, demonstrates outstanding performance (even competitive with some models fine-tuned on ground truth ADs) in AD generation for both movies and TV series, achieving state-of-the-art CRITIC scores., Comment: Project Page: https://www.robots.ox.ac.uk/~vgg/research/autoad-zero/

This paper tackles the task of emotion-cause pair extraction in the unsupervised domain adaptation setting. The problem is challenging as the distributions of the events causing emotions in target domains are dramatically different than those in source domains, despite the distributions of emotional expressions between domains are overlapped. Inspired by causal discovery, we propose a novel deep latent model in the variational autoencoder (VAE) framework, which not only captures the underlying latent structures of data but also utilizes the easily transferable knowledge of emotions as the bridge to link the distributions of events in different domains. To facilitate knowledge transfer across domains, we also propose a novel variational posterior regularization technique to disentangle the latent representations of emotions from those of events in order to mitigate the damage caused by the spurious correlations related to the events in source domains. Through extensive experiments, we demonstrate that our model outperforms the strongest baseline by approximately 11.05% on a Chinese benchmark and 2.45% on a English benchmark in terms of weighted-average F1 score. The source code will be publicly available upon acceptance., Comment: 12 pages, 6 figures, 4 tables; Under Review in EMNLP 2024

We introduce Vibe-Eval: a new open benchmark and framework for evaluating multimodal chat models. Vibe-Eval consists of 269 visual understanding prompts, including 100 of hard difficulty, complete with gold-standard responses authored by experts. Vibe-Eval is open-ended and challenging with dual objectives: (i) vibe checking multimodal chat models for day-to-day tasks and (ii) rigorously testing and probing the capabilities of present frontier models. Notably, our hard set contains >50% questions that all frontier models answer incorrectly. We explore the nuances of designing, evaluating, and ranking models on ultra challenging prompts. We also discuss trade-offs between human and automatic evaluation, and show that automatic model evaluation using Reka Core roughly correlates to human judgment. We offer free API access for the purpose of lightweight evaluation and plan to conduct formal human evaluations for public models that perform well on the Vibe-Eval's automatic scores. We release the evaluation code and data, see https://github.com/reka-ai/reka-vibe-eval

Mapping traversal costs in an environment and planning paths based on this map are important for autonomous navigation. We present a neurobotic navigation system that utilizes a Spiking Neural Network Wavefront Planner and E-prop learning to concurrently map and plan paths in a large and complex environment. We incorporate a novel method for mapping which, when combined with the Spiking Wavefront Planner, allows for adaptive planning by selectively considering any combination of costs. The system is tested on a mobile robot platform in an outdoor environment with obstacles and varying terrain. Results indicate that the system is capable of discerning features in the environment using three measures of cost, (1) energy expenditure by the wheels, (2) time spent in the presence of obstacles, and (3) terrain slope. In just twelve hours of online training, E-prop learns and incorporates traversal costs into the path planning maps by updating the delays in the Spiking Wavefront Planner. On simulated paths, the Spiking Wavefront Planner plans significantly shorter and lower cost paths than A* and RRT*. The spiking wavefront planner is compatible with neuromorphic hardware and could be used for applications requiring low size, weight, and power., Comment: 8 pages, 7 figures, 2 tables

Generating Audio Description (AD) for movies is a challenging task that requires fine-grained visual understanding and an awareness of the characters and their names. Currently, visual language models for AD generation are limited by a lack of suitable training data, and also their evaluation is hampered by using performance measures not specialized to the AD domain. In this paper, we make three contributions: (i) We propose two approaches for constructing AD datasets with aligned video data, and build training and evaluation datasets using these. These datasets will be publicly released; (ii) We develop a Q-former-based architecture which ingests raw video and generates AD, using frozen pre-trained visual encoders and large language models; and (iii) We provide new evaluation metrics to benchmark AD quality that are well-matched to human performance. Taken together, we improve the state of the art on AD generation., Comment: CVPR2024. Project page: https://www.robots.ox.ac.uk/~vgg/research/autoad/

We introduce Reka Core, Flash, and Edge, a series of powerful multimodal language models trained from scratch by Reka. Reka models are able to process and reason with text, images, video, and audio inputs. This technical report discusses details of training some of these models and provides comprehensive evaluation results. We show that Reka Edge and Reka Flash are not only state-of-the-art but also outperform many much larger models, delivering outsized values for their respective compute class. Meanwhile, our most capable and largest model, Reka Core, approaches the best frontier models on both automatic evaluations and blind human evaluations. On image question answering benchmarks (e.g. MMMU, VQAv2), Core performs competitively to GPT4-V. Meanwhile, on multimodal chat, Core ranks as the second most preferred model under a blind third-party human evaluation setup, outperforming other models such as Claude 3 Opus. On text benchmarks, Core not only performs competitively to other frontier models on a set of well-established benchmarks (e.g. MMLU, GSM8K) but also outperforms GPT4-0613 on human evaluation. On video question answering (Perception-Test), Core outperforms Gemini Ultra. Models are shipped in production at http://chat.reka.ai . A showcase of non cherry picked qualitative examples can also be found at http://showcase.reka.ai .

Humans see low and high spatial frequency components at the same time, and combine the information from both to form a visual scene. Drawing on this neuroscientific inspiration, we propose an altered Vision Transformer architecture where patches from scaled down versions of the input image are added to the input of the first Transformer Encoder layer. We name this model Retina Vision Transformer (RetinaViT) due to its inspiration from the human visual system. Our experiments show that when trained on the ImageNet-1K dataset with a moderate configuration, RetinaViT achieves a 3.3% performance improvement over the original ViT. We hypothesize that this improvement can be attributed to the inclusion of low spatial frequency components in the input, which improves the ability to capture structural features, and to select and forward important features to deeper layers. RetinaViT thereby opens doors to further investigations into vertical pathways and attention patterns.

It is widely known that freezing breaks soft, wet materials. However, the mechanism underlying this damage is still not clear. To understand this process, we freeze model, brittle hydrogel samples, while observing the growth of ice-filled cracks that break these apart. We show that damage is not caused by the expansion of water upon freezing, or the growth of ice-filled cavities in the hydrogel. Instead, local ice growth dehydrates the surrounding hydrogel, leading to drying-induced fracture. This dehydration is driven by the process of cryosuction, whereby undercooled ice sucks nearby water towards itself, feeding its growth. Our results highlight the strong analogy between freezing damage and desiccation cracking, which we anticipate being useful for developing an understanding of both topics. Our results should also give useful insights into a wide range of freezing processes, including cryopreservation, food science and frost heave.

The Shuttleworth equation: a linear stress-strain relation ubiquitously used in modeling the behavior of soft surfaces. Its validity in the realm of materials subject to large deformation is a topic of current debate. Here, we allow for large deformation by deriving the constitutive behavior of the surface from the general framework of finite kinematics. We distinguish cases of finite and infinitesimal surface relaxation preceding an infinitesimal applied deformation. The Shuttleworth equation identifies as the Cauchy stress measure in the fully linearized setting. We show that both in finite and linearized cases, measured elastic constants depend on the utilized stress measure. In addition, we discuss the physical implications of our results and analyze the impact of surface relaxation on the estimation of surface elastic moduli in the light of two different test cases.

What are the barriers to employment for young people who are not in education, employment or training (NEET) or who have experience of care? What works best to support them? In partnership with the Employment Related Services Association (ERSA), our latest report draws conclusions from a survey of practitioners and young people to understand the barriers they face, what services are offered, and what is considered to work best in helping NEET young people--and especially care leavers--to progress into work. More than one in ten young people in the UK between the ages of 16 and 24 are classified as not in employment, education or training (NEET), and the rate of young people being NEET has not fallen below 10% over the last twenty years. We know that being NEET can have immediate negative consequences for a young person, such as decreasing levels of self-esteem, increasing the likelihood of engaging in risky behaviours such as substance misuse and criminal activities, and increasing their chances of living in poverty. While there has been much research into the factors that lead to becoming NEET, how these can be mitigated, and the impacts of being NEET, there is not the same body of evidence to inform the support provided to young people to move into and stay in education, employment and training (EET). This study looks at what the existing research has to say and surveys young people and the practitioners who work with them on four key questions: (1) What are the key barriers to EET for young people and care leavers who are NEET?; (2) What is currently offered to young people and care leavers in terms of entry to EET support?; (3) What is currently offered in terms of in-work support for young people and care leavers?; and (4) What types of provision are most effective at dealing with the barriers young people face to EET? Drawing out lessons for practitioners and policymakers, this study contributes to our understanding of what young people see as the issues, the support they need and what service providers are offering to some of the most vulnerable people in society. [This report was produced with the Employment Related Services Association (ERSA).]

We introduce a new multivariate data set that utilizes multiple spacecraft collecting in-situ and remote sensing heliospheric measurements shown to be linked to physical processes responsible for generating solar energetic particles (SEPs). Using the Geostationary Operational Environmental Satellites (GOES) flare event list from Solar Cycle (SC) 23 and part of SC 24 (1998-2013), we identify 252 solar events (flares) that produce SEPs and 17,542 events that do not. For each identified event, we acquire the local plasma properties at 1 au, such as energetic proton and electron data, upstream solar wind conditions, and the interplanetary magnetic field vector quantities using various instruments onboard GOES and the Advanced Composition Explorer (ACE) spacecraft. We also collect remote sensing data from instruments onboard the Solar Dynamic Observatory (SDO), Solar and Heliospheric Observatory (SoHO), and the Wind solar radio instrument WAVES. The data set is designed to allow for variations of the inputs and feature sets for machine learning (ML) in heliophysics and has a specific purpose for forecasting the occurrence of SEP events and their subsequent properties. This paper describes a dataset created from multiple publicly available observation sources that is validated, cleaned, and carefully curated for our machine-learning pipeline. The dataset has been used to drive the newly-developed Multivariate Ensemble of Models for Probabilistic Forecast of Solar Energetic Particles (MEMPSEP; see MEMPSEP I (Chatterjee et al., 2023) and MEMPSEP II (Dayeh et al., 2023) for associated papers).

Audio Description (AD) is the task of generating descriptions of visual content, at suitable time intervals, for the benefit of visually impaired audiences. For movies, this presents notable challenges -- AD must occur only during existing pauses in dialogue, should refer to characters by name, and ought to aid understanding of the storyline as a whole. To this end, we develop a new model for automatically generating movie AD, given CLIP visual features of the frames, the cast list, and the temporal locations of the speech; addressing all three of the 'who', 'when', and 'what' questions: (i) who -- we introduce a character bank consisting of the character's name, the actor that played the part, and a CLIP feature of their face, for the principal cast of each movie, and demonstrate how this can be used to improve naming in the generated AD; (ii) when -- we investigate several models for determining whether an AD should be generated for a time interval or not, based on the visual content of the interval and its neighbours; and (iii) what -- we implement a new vision-language model for this task, that can ingest the proposals from the character bank, whilst conditioning on the visual features using cross-attention, and demonstrate how this improves over previous architectures for AD text generation in an apples-to-apples comparison., Comment: ICCV2023. Project page: https://www.robots.ox.ac.uk/vgg/research/autoad/

The Sun continuously affects the interplanetary environment through a host of interconnected and dynamic physical processes. Solar flares, Coronal Mass Ejections (CMEs), and Solar Energetic Particles (SEPs) are among the key drivers of space weather in the near-Earth environment and beyond. While some CMEs and flares are associated with intense SEPs, some show little to no SEP association. To date, robust long-term (hours-days) forecasting of SEP occurrence and associated properties (e.g., onset, peak intensities) does not effectively exist and the search for such development continues. Through an Operations-2-Research support, we developed a self-contained model that utilizes a comprehensive dataset and provides a probabilistic forecast for SEP event occurrence and its properties. The model is named Multivariate Ensemble of Models for Probabilistic Forecast of Solar Energetic Particles (MEMPSEP). MEMPSEP workhorse is an ensemble of Convolutional Neural Networks that ingests a comprehensive dataset (MEMPSEP III - (Moreland et al., 2023)) of full-disc magnetogram-sequences and in-situ data from different sources to forecast the occurrence (MEMPSEP I - this work) and properties (MEMPSEP II - Dayeh et al. (2023)) of a SEP event. This work focuses on estimating true SEP occurrence probabilities achieving a 2.5% improvement in reliability and a Brier score of 0.14. The outcome provides flexibility for the end-users to determine their own acceptable level of risk, rather than imposing a detection threshold that optimizes an arbitrary binary classification metric. Furthermore, the model-ensemble, trained to utilize the large class-imbalance between events and non-events, provides a clear measure of uncertainty in our forecast, Comment: 17 pages, 8 figures, 1 table, accepted for publication in Space Weather journal

Solar Energetic Particles (SEPs) form a critical component of Space Weather. The complex, intertwined dynamics of SEP sources, acceleration, and transport make their forecasting very challenging. Yet, information about SEP arrival and their properties (e.g., peak flux) is crucial for space exploration on many fronts. We have recently introduced a novel probabilistic ensemble model called the Multivariate Ensemble of Models for Probabilistic Forecast of Solar Energetic Particles (MEMPSEP). Its primary aim is to forecast the occurrence and physical properties of SEPs. The occurrence forecasting, thoroughly discussed in a preceding paper (Chatterjee et al., 2023), is complemented by the work presented here, which focuses on forecasting the physical properties of SEPs. The MEMPSEP model relies on an ensemble of Convolutional Neural Networks, which leverage a multi-variate dataset comprising full-disc magnetogram sequences and numerous derived and in-situ data from various sources. Skill scores demonstrate that MEMPSEP exhibits improved predictions on SEP properties for the test set data with SEP occurrence probability above 50%, compared to those with a probability below 50%. Results present a promising approach to address the challenging task of forecasting SEP physical properties, thus improving our forecasting capabilities and advancing our understanding of the dominant parameters and processes that govern SEP production.

This comprehensive study conducts an in-depth analysis of existing COVID-19 ontologies, scrutinizing their objectives, classifications, design methodologies, and domain focal points. The study is conducted through a dual-stage approach, commencing with a systematic review of relevant literature and followed by an ontological assessment utilizing a parametric methodology. Through this meticulous process, twenty-four COVID-19 Ontologies (CovOs) are selected and examined. The findings highlight the scope, intended purpose, granularity of ontology, modularity, formalism, vocabulary reuse, and extent of domain coverage. The analysis reveals varying levels of formality in ontology development, a prevalent preference for utilizing OWL as the representational language, and diverse approaches to constructing class hierarchies within the models. Noteworthy is the recurrent reuse of ontologies like OBO models (CIDO, GO, etc.) alongside CODO. The METHONTOLOGY approach emerges as a favored design methodology, often coupled with application-based or data-centric evaluation methods. Our study provides valuable insights for the scientific community and COVID-19 ontology developers, supplemented by comprehensive ontology metrics. By meticulously evaluating and documenting COVID-19 information-driven ontological models, this research offers a comparative cross-domain perspective, shedding light on knowledge representation variations. The present study significantly enhances understanding of CovOs, serving as a consolidated resource for comparative analysis and future development, while also pinpointing research gaps and domain emphases, thereby guiding the trajectory of future ontological advancements., Comment: 16 pages, accepted for publication in 17th International Conference on Metadata and Semantics Research (MTSR2023), University of Milano-Bicocca, Milan, Italy, October 23-27, 2023

Space weather is a multidisciplinary research area connecting scientists from across heliophysics domains seeking a coherent understanding of our space environment that can also serve modern life and society's needs. COSPAR's ISWAT (International Space Weather Action Teams) 'clusters' focus attention on different areas of space weather study while ensuring the coupled system is broadly addressed via regular communications and interactions. The ISWAT cluster "H3: Radiation Environment in the Heliosphere" (https://www.iswat-cospar.org/h3) has been working to provide a scientific platform to understand, characterize and predict the energetic particle radiation in the heliosphere with the practical goal of mitigating radiation risks associated with areospace activities, satellite industry and human space explorations. In particular, present approaches help us understand the physical phenomena at large, optimizing the output of multi-viewpoint observations and pushing current models to their limits. In this paper, we review the scientific aspects of the radiation environment in the heliosphere covering four different radiation types: Solar Energetic Particles (SEPs), Ground Level Enhancement (GLE, a type of SEP events with energies high enough to trigger the enhancement of ground-level detectors), Galactic Cosmic Rays (GCRs) and Anomalous Cosmic Rays (ACRs). We focus on related advances in the research community in the past 10-20 years and what we still lack in terms of understanding and predictive capabilities. Finally we also consider some recommendations related to the improvement of both observational and modeling capabilities in the field of space radiation environment.

Extreme Ultraviolet images of the Sun are becoming an integral part of space weather prediction tasks. However, having different surveys requires the development of instrument-specific prediction algorithms. As an alternative, it is possible to combine multiple surveys to create a homogeneous dataset. In this study, we utilize the temporal overlap of SoHO/EIT and SDO/AIA 171~\AA ~surveys to train an ensemble of deep learning models for creating a single homogeneous survey of EUV images for 2 solar cycles. Prior applications of deep learning have focused on validating the homogeneity of the output while overlooking the systematic estimation of uncertainty. We use an approach called `Approximate Bayesian Ensembling' to generate an ensemble of models whose uncertainty mimics that of a fully Bayesian neural network at a fraction of the cost. We find that ensemble uncertainty goes down as the training set size increases. Additionally, we show that the model ensemble adds immense value to the prediction by showing higher uncertainty in test data that are not well represented in the training data., Comment: 20 pages, 8 figures, accepted for publication in ApJS