Search

The World Health Organization, in response to the growing burden of fungal disease, established a process to develop a fungal priority pathogens list. This systematic review aimed to evaluate the epidemiology and impact of eumycetoma. PubMed and Web of Science were searched to identify studies published between 1 January 2011 and 19 February 2021. Studies reporting on mortality, inpatient care, complications and sequelae, antifungal susceptibility, risk factors, preventability, annual incidence, global distribution, and emergence during the study time frames were selected. Overall, 14 studies were eligible for inclusion. Morbidity was frequent with moderate to severe impairment of quality of life in 60.3%, amputation in up to 38.5%, and recurrent or long-term disease in 31.8%-73.5% of patients. Potential risk factors included male gender (56.6%-79.6%), younger age (11-30 years; 64%), and farming occupation (62.1%-69.7%). Mycetoma was predominantly reported in Sudan, particularly in central Sudan (37%-76.6% of cases). An annual incidence of 0.1/100 000 persons and 0.32/100 000 persons/decade was reported in the Philippines and Uganda, respectively. In Uganda, a decline in incidence from 3.37 to 0.32/100 000 persons between two consecutive 10-year periods (2000-2009 and 2010-2019) was detected. A community-based, multi-pronged prevention programme was associated with a reduction in amputation rates from 62.8% to 11.9%. With the pre-specified criteria, no studies of antifungal drug susceptibility, mortality, and hospital lengths of stay were identified. Future research should include larger cohort studies, greater drug susceptibility testing, and global surveillance to develop evidence-based treatment guidelines and to determine more accurately the incidence and trends over time.

While previous reviews found a positive association between pre-existing cancer diagnosis and COVID-19-related death, most early studies did not distinguish long-term cancer survivors from those recently diagnosed/treated, nor adjust for important confounders including age. We aimed to consolidate higher-quality evidence on risk of COVID-19-related death for people with recent/active cancer (compared to people without) in the pre-COVID-19-vaccination period. We searched the WHO COVID-19 Global Research Database (20 December 2021), and Medline and Embase (10 May 2023). We included studies adjusting for age and sex, and providing details of cancer status. Risk-of-bias assessment was based on the Newcastle-Ottawa Scale. Pooled adjusted odds or risk ratios (aORs, aRRs) or hazard ratios (aHRs) and 95% confidence intervals (95% CIs) were calculated using generic inverse-variance random-effects models. Random-effects meta-regressions were used to assess associations between effect estimates and time since cancer diagnosis/treatment. Of 23 773 unique title/abstract records, 39 studies were eligible for inclusion (2 low, 17 moderate, 20 high risk of bias). Risk of COVID-19-related death was higher for people with active or recently diagnosed/treated cancer (general population: aOR = 1.48, 95% CI: 1.36-1.61, I2 = 0; people with COVID-19: aOR = 1.58, 95% CI: 1.41-1.77, I2 = 0.58; inpatients with COVID-19: aOR = 1.66, 95% CI: 1.34-2.06, I2 = 0.98). Risks were more elevated for lung (general population: aOR = 3.4, 95% CI: 2.4-4.7) and hematological cancers (general population: aOR = 2.13, 95% CI: 1.68-2.68, I2 = 0.43), and for metastatic cancers. Meta-regression suggested risk of COVID-19-related death decreased with time since diagnosis/treatment, for example, for any/solid cancers, fitted aOR = 1.55 (95% CI: 1.37-1.75) at 1 year and aOR = 0.98 (95% CI: 0.80-1.20) at 5 years post-cancer diagnosis/treatment. In conclusion, before CO

Understanding human actions from body poses is critical for assistive robots sharing space with humans in order to make informed and safe decisions about the next interaction. However, precise temporal localization and annotation of activity sequences is time-consuming and the resulting labels are often noisy. If not effectively addressed, label noise negatively affects the model's training, resulting in lower recognition quality. Despite its importance, addressing label noise for skeleton-based action recognition has been overlooked so far. In this study, we bridge this gap by implementing a framework that augments well-established skeleton-based human action recognition methods with label-denoising strategies from various research areas to serve as the initial benchmark. Observations reveal that these baselines yield only marginal performance when dealing with sparse skeleton data. Consequently, we introduce a novel methodology, NoiseEraSAR, which integrates global sample selection, co-teaching, and Cross-Modal Mixture-of-Experts (CM-MOE) strategies, aimed at mitigating the adverse impacts of label noise. Our proposed approach demonstrates better performance on the established benchmark, setting new state-of-the-art standards. The source code for this study will be made accessible at https://github.com/xuyizdby/NoiseEraSAR., Comment: The source code will be made accessible at https://github.com/xuyizdby/NoiseEraSAR

In a world driven by data visualization, ensuring the inclusive accessibility of charts for Blind and Visually Impaired (BVI) individuals remains a significant challenge. Charts are usually presented as raster graphics without textual and visual metadata needed for an equivalent exploration experience for BVI people. Additionally, converting these charts into accessible formats requires considerable effort from sighted individuals. Digitizing charts with metadata extraction is just one aspect of the issue; transforming it into accessible modalities, such as tactile graphics, presents another difficulty. To address these disparities, we propose Chart4Blind, an intelligent user interface that converts bitmap image representations of line charts into universally accessible formats. Chart4Blind achieves this transformation by generating Scalable Vector Graphics (SVG), Comma-Separated Values (CSV), and alternative text exports, all comply with established accessibility standards. Through interviews and a formal user study, we demonstrate that even inexperienced sighted users can make charts accessible in an average of 4 minutes using Chart4Blind, achieving a System Usability Scale rating of 90%. In comparison to existing approaches, Chart4Blind provides a comprehensive solution, generating end-to-end accessible SVGs suitable for assistive technologies such as embossed prints (papers and laser cut), 2D tactile displays, and screen readers. For additional information, including open-source codes and demos, please visit our project page https://moured.github.io/chart4blind/., Comment: Accepted to IUI 2024. 19 pages, 7 figures, 2 table. For a demo video, see this https://moured.github.io/chart4blind/ . The source code is available at https://github.com/moured/chart4blind_code

Generative AI has the potential to revolutionize drug discovery. Yet, despite recent advances in machine learning, existing models cannot generate molecules that satisfy all desired physicochemical properties. Herein, we describe IDOLpro, a novel generative chemistry AI combining deep diffusion with multi-objective optimization for structure-based drug design. The latent variables of the diffusion model are guided by differentiable scoring functions to explore uncharted chemical space and generate novel ligands in silico, optimizing a plurality of target physicochemical properties. We demonstrate its effectiveness by generating ligands with optimized binding affinity and synthetic accessibility on two benchmark sets. IDOLpro produces ligands with binding affinities over 10% higher than the next best state-of-the-art on each test set. On a test set of experimental complexes, IDOLpro is the first to surpass the performance of experimentally observed ligands. IDOLpro can accommodate other scoring functions (e.g. ADME-Tox) to accelerate hit-finding, hit-to-lead, and lead optimization for drug discovery.

In questa tesi ci si propone di analizzare le implicazioni fisiche della geometria dello spazio ambiente nel contesto dei condensati di Bose-Einstein (BEC) e le possibili applicazioni nell’ambito dei modelli analogici della cosmologia dei buchi neri. A tal fine si deriva la formulazione idrodinamica dell'equazione di Gross-Pitaevskii (GPE) nel caso di una generica varietà Riemanniana e si osserva la comparsa di una nuova forza, che dipende essenzialmente da due parametri: la geometria della varietà e le derivate prime del profilo di densità. Si studiano le condizioni stazionarie in relazione alla presenza di varietà a curvatura scalare negativa. Analizzando tali varietà si stabilisce una relazione esplicita tra le superfici a curvatura negativa e l'equazione di seno-Gordon, che risulta un'approssimazione della GPE nel caso di accoppiamento di fasi. Assumendo condizioni stazionarie, si ottiene un nuovo tipo di equazioni di Einstein e si è spinti a ricercare altri legami tra le equazioni che governano i condensati e la cosmologia. A tal fine si considerano i BEC relativistici, che vengono utilizzati nello studio del comportamento dell'universo primordiale e della sua espansione. Facendo uso delle conoscenze ottenute nel caso di varietà Riemanniane generiche, otteniamo nuove equazioni di Einstein nel caso multi-dimensionale. Successivamente, si considerano i modelli analogici utilizzati per lo studio della formazione di buchi neri e per il calcolo della radiazione di Hawking. Attraverso un processo di linearizzazione si nota come sia possibile far emergere una metrica acustica Lorentziana che governi il moto delle fluttuazioni della fase; a questo scopo si considera il caso di un vortice dritto che presenta un profilo di densità in cui le derivate prime assumono un valore massimo all’interno del tubo vorticoso e la geometria dello spazio ambiente diventa rilevante. In questa situazione si scopre che è effettivamente possibile far emergere una metrica Lorentzia, In this thesis we analyze the physical implications of the geometry of the ambient space in the context of Bose-Einstein condensates (BECs) and possible applications to the field of analogue models in the cosmology of black holes. To this end we derive the hydrodynamic formulation of the Gross-Pitaevskii equation (GPE) in the case of a generic Riemannian manifold. We observe the appearance of a new force, which essentially depends on two parameters: the geometry of the manifold and the first derivatives of the density profile. The stationary conditions are studied in relation to the presence of manifolds with negative scalar curvature. By analyzing these manifolds, an explicit relationship is established between the negatively curved surfaces and the sine-Gordon equation, which results in an approximation of the GPE in the presence of phase coupling. By assuming stationary conditions, we obtain a new type of Einstein field equations and we look for other possible connections between the equations governing condensates and cosmology. For this purpose, we consider relativistic BECs, that are used in the study of the early universe and its expansion, and we obtain Einstein equation in the multi-dimensional case. Then we consider the analogue models used for the study of the formation of black holes and for the calculation of Hawking radiation. Through a linearization process it is possible to derive a Lorentzian acoustic metric for the phase fluctuations; for this purpose, we consider the case of a straight vortex defect with a density profile where the first derivatives have maximum value inside the vortex tube and the geometry of the ambient space becomes relevant. In this situation it turns out that it is possible to determine a Lorentzian metric, and some useful approximations are proposed for its explicit computation. Finally, some concluding remarks are presented on possible future research directions, given by the study of the evolution of isophase surfaces in r

AmberTools is a free and open-source collection of programs used to set up, run, and analyze molecular simulations. The newer features contained within AmberTools23 are briefly described in this Application note.

In this paper, we tackle the new task of video-based Activated Muscle Group Estimation (AMGE) aiming at identifying active muscle regions during physical activity in the wild. To this intent, we provide the MuscleMap dataset featuring >15K video clips with 135 different activities and 20 labeled muscle groups. This dataset opens the vistas to multiple video-based applications in sports and rehabilitation medicine under flexible environment constraints. The proposed MuscleMap dataset is constructed with YouTube videos, specifically targeting High-Intensity Interval Training (HIIT) physical exercise in the wild. To make the AMGE model applicable in real-life situations, it is crucial to ensure that the model can generalize well to numerous types of physical activities not present during training and involving new combinations of activated muscles. To achieve this, our benchmark also covers an evaluation setting where the model is exposed to activity types excluded from the training set. Our experiments reveal that the generalizability of existing architectures adapted for the AMGE task remains a challenge. Therefore, we also propose a new approach, TransM3E, which employs a multi-modality feature fusion mechanism between both the video transformer model and the skeleton-based graph convolution model with novel cross-modal knowledge distillation executed on multi-classification tokens. The proposed method surpasses all popular video classification models when dealing with both, previously seen and new types of physical activities. The contributed dataset and code will be publicly available at https://github.com/KPeng9510/MuscleMap., Comment: The contributed dataset and code will be publicly available at https://github.com/KPeng9510/MuscleMap

Domain adaptation is essential for activity recognition to ensure accurate and robust performance across diverse environments, sensor types, and data sources. Unsupervised domain adaptation methods have been extensively studied, yet, they require large-scale unlabeled data from the target domain. In this work, we focus on Few-Shot Domain Adaptation for Activity Recognition (FSDA-AR), which leverages a very small amount of labeled target videos to achieve effective adaptation. This approach is appealing for applications because it only needs a few or even one labeled example per class in the target domain, ideal for recognizing rare but critical activities. However, the existing FSDA-AR works mostly focus on the domain adaptation on sports videos, where the domain diversity is limited. We propose a new FSDA-AR benchmark using five established datasets considering the adaptation on more diverse and challenging domains. Our results demonstrate that FSDA-AR performs comparably to unsupervised domain adaptation with significantly fewer labeled target domain samples. We further propose a novel approach, RelaMiX, to better leverage the few labeled target domain samples as knowledge guidance. RelaMiX encompasses a temporal relational attention network with relation dropout, alongside a cross-domain information alignment mechanism. Furthermore, it integrates a mechanism for mixing features within a latent space by using the few-shot target domain samples. The proposed RelaMiX solution achieves state-of-the-art performance on all datasets within the FSDA-AR benchmark. To encourage future research of few-shot domain adaptation for activity recognition, our code will be publicly available at https://github.com/KPeng9510/RelaMiX., Comment: The benchmark and source code will be publicly available at https://github.com/KPeng9510/RelaMiX

Interactive Image Segmentation (IIS) has emerged as a promising technique for decreasing annotation time. Substantial progress has been made in pre- and post-processing for IIS, but the critical issue of interaction ambiguity, notably hindering segmentation quality, has been under-researched. To address this, we introduce AdaptiveClick -- a click-aware transformer incorporating an adaptive focal loss that tackles annotation inconsistencies with tools for mask- and pixel-level ambiguity resolution. To the best of our knowledge, AdaptiveClick is the first transformer-based, mask-adaptive segmentation framework for IIS. The key ingredient of our method is the Click-Aware Mask-adaptive transformer Decoder (CAMD), which enhances the interaction between click and image features. Additionally, AdaptiveClick enables pixel-adaptive differentiation of hard and easy samples in the decision space, independent of their varying distributions. This is primarily achieved by optimizing a generalized Adaptive Focal Loss (AFL) with a theoretical guarantee, where two adaptive coefficients control the ratio of gradient values for hard and easy pixels. Our analysis reveals that the commonly used Focal and BCE losses can be considered special cases of the proposed AFL. With a plain ViT backbone, extensive experimental results on nine datasets demonstrate the superiority of AdaptiveClick compared to state-of-the-art methods. The source code is publicly available at https://github.com/lab206/AdaptiveClick., Comment: Accepted to IEEE Transactions on Neural Networks and Learning Systems (TNNLS). The source code is publicly available at https://github.com/lab206/AdaptiveClick

This paper raises the new task of Fisheye Semantic Completion (FSC), where dense texture, structure, and semantics of a fisheye image are inferred even beyond the sensor field-of-view (FoV). Fisheye cameras have larger FoV than ordinary pinhole cameras, yet its unique special imaging model naturally leads to a blind area at the edge of the image plane. This is suboptimal for safety-critical applications since important perception tasks, such as semantic segmentation, become very challenging within the blind zone. Previous works considered the out-FoV outpainting and in-FoV segmentation separately. However, we observe that these two tasks are actually closely coupled. To jointly estimate the tightly intertwined complete fisheye image and scene semantics, we introduce the new FishDreamer which relies on successful ViTs enhanced with a novel Polar-aware Cross Attention module (PCA) to leverage dense context and guide semantically-consistent content generation while considering different polar distributions. In addition to the contribution of the novel task and architecture, we also derive Cityscapes-BF and KITTI360-BF datasets to facilitate training and evaluation of this new track. Our experiments demonstrate that the proposed FishDreamer outperforms methods solving each task in isolation and surpasses alternative approaches on the Fisheye Semantic Completion. Code and datasets are publicly available at https://github.com/MasterHow/FishDreamer., Comment: Accepted to CVPR OmniCV 2023. Code and datasets will be available at https://github.com/MasterHow/FishDreamer

Background: The European Society for Medical Oncology-Magnitude of Clinical Benefit Scale (ESMO-MCBS) has been developed to grade clinical benefit of cancer therapies. Improvement in quality of life (QoL) is considered relevant, especially in the non-curative setting. This is reflected by an upgrade of the preliminary ESMO-MCBS score if QoL is improved compared to the control arm or a downgrade if an improvement in progression-free survival is not paralleled by an improvement in QoL or overall survival. Given the importance of QoL for the final score, a need to ensure the robustness of QoL data was recognised. Design: A checklist was created based on existing guidelines for QoL research. Field testing was carried out using clinical trials that either received an adjustment of the preliminary ESMO-MCBS score based on QoL or had QoL as the primary endpoint. Several rounds of revision and re-testing of the checklist were undertaken until a final consensus was reached. Results: The final checklist consists of four items and can be applied if three prerequisites are met: (i) QoL is at least a secondary endpoint, (ii) evidence of reliability and validity of the instrument is provided, and (iii) a statistically and clinically significant improvement in QoL is observed. The four items on the checklist pertain to the (i) hypothesis, (ii) compliance and missing data, (iii) presentation of the results, and (iv) statistical and clinical relevance. Field testing revealed that a clear QoL hypothesis and correction for multiple testing were mostly lacking, while the main statistical method was always described. Conclusions: Implementation of the ESMO-MCBS QoL checklist will facilitate objective and transparent decision making on QoL data within the ESMO-MCBS scoring process. Trials published until 1 January 2025 will have to meet the prerequisites and at least two items for crediting QoL benefit in the final ESMO-MCBS score. Trials published thereafter will have to meet all four i, SCOPUS: ar.j, info:eu-repo/semantics/published

In real-world scenarios, human actions often fall outside the distribution of training data, making it crucial for models to recognize known actions and reject unknown ones. However, using pure skeleton data in such open-set conditions poses challenges due to the lack of visual background cues and the distinct sparse structure of body pose sequences. In this paper, we tackle the unexplored Open-Set Skeleton-based Action Recognition (OS-SAR) task and formalize the benchmark on three skeleton-based datasets. We assess the performance of seven established open-set approaches on our task and identify their limits and critical generalization issues when dealing with skeleton information. To address these challenges, we propose a distance-based cross-modality ensemble method that leverages the cross-modal alignment of skeleton joints, bones, and velocities to achieve superior open-set recognition performance. We refer to the key idea as CrossMax - an approach that utilizes a novel cross-modality mean max discrepancy suppression mechanism to align latent spaces during training and a cross-modality distance-based logits refinement method during testing. CrossMax outperforms existing approaches and consistently yields state-of-the-art results across all datasets and backbones. The benchmark, code, and models will be released at https://github.com/KPeng9510/OS-SAR., Comment: Accepted to AAAI 2024. The benchmark, code, and models will be released at https://github.com/KPeng9510/OS-SAR

Deep learning-based models are at the forefront of most driver observation benchmarks due to their remarkable accuracies but are also associated with high computational costs. This is challenging, as resources are often limited in real-world driving scenarios. This paper introduces a lightweight framework for resource-efficient driver activity recognition. The framework enhances 3D MobileNet, a neural architecture optimized for speed in video classification, by incorporating knowledge distillation and model quantization to balance model accuracy and computational efficiency. Knowledge distillation helps maintain accuracy while reducing the model size by leveraging soft labels from a larger teacher model (I3D), instead of relying solely on original ground truth data. Model quantization significantly lowers memory and computation demands by using lower precision integers for model weights and activations. Extensive testing on a public dataset for in-vehicle monitoring during autonomous driving demonstrates that this new framework achieves a threefold reduction in model size and a 1.4-fold improvement in inference time, compared to an already optimized architecture. The code for this study is available at https://github.com/calvintanama/qd-driver-activity-reco., Comment: Accepted at IROS 2023

To integrate action recognition methods into autonomous robotic systems, it is crucial to consider adverse situations involving target occlusions. Such a scenario, despite its practical relevance, is rarely addressed in existing self-supervised skeleton-based action recognition methods. To empower robots with the capacity to address occlusion, we propose a simple and effective method. We first pre-train using occluded skeleton sequences, then use k-means clustering (KMeans) on sequence embeddings to group semantically similar samples. Next, we employ K-nearest-neighbor (KNN) to fill in missing skeleton data based on the closest sample neighbors. Imputing incomplete skeleton sequences to create relatively complete sequences as input provides significant benefits to existing skeleton-based self-supervised models. Meanwhile, building on the state-of-the-art Partial Spatio-Temporal Learning (PSTL), we introduce an Occluded Partial Spatio-Temporal Learning (OPSTL) framework. This enhancement utilizes Adaptive Spatial Masking (ASM) for better use of high-quality, intact skeletons. The effectiveness of our imputation methods is verified on the challenging occluded versions of the NTURGB+D 60 and NTURGB+D 120. The source code will be made publicly available at https://github.com/cyfml/OPSTL., Comment: The source code will be made publicly available at https://github.com/cyfml/OPSTL

Self-supervised representation learning for human action recognition has developed rapidly in recent years. Most of the existing works are based on skeleton data while using a multi-modality setup. These works overlooked the differences in performance among modalities, which led to the propagation of erroneous knowledge between modalities while only three fundamental modalities, i.e., joints, bones, and motions are used, hence no additional modalities are explored. In this work, we first propose an Implicit Knowledge Exchange Module (IKEM) which alleviates the propagation of erroneous knowledge between low-performance modalities. Then, we further propose three new modalities to enrich the complementary information between modalities. Finally, to maintain efficiency when introducing new modalities, we propose a novel teacher-student framework to distill the knowledge from the secondary modalities into the mandatory modalities considering the relationship constrained by anchors, positives, and negatives, named relational cross-modality knowledge distillation. The experimental results demonstrate the effectiveness of our approach, unlocking the efficient use of skeleton-based multi-modality data. Source code will be made publicly available at https://github.com/desehuileng0o0/IKEM., Comment: Accepted to ICASSP 2024. The source code will be made publicly available at https://github.com/desehuileng0o0/IKEM

For driver observation frameworks, clean datasets collected in controlled simulated environments often serve as the initial training ground. Yet, when deployed under real driving conditions, such simulator-trained models quickly face the problem of distributional shifts brought about by changing illumination, car model, variations in subject appearances, sensor discrepancies, and other environmental alterations. This paper investigates the viability of transferring video-based driver observation models from simulation to real-world scenarios in autonomous vehicles, given the frequent use of simulation data in this domain due to safety issues. To achieve this, we record a dataset featuring actual autonomous driving conditions and involving seven participants engaged in highly distracting secondary activities. To enable direct SIM to REAL transfer, our dataset was designed in accordance with an existing large-scale simulator dataset used as the training source. We utilize the Inflated 3D ConvNet (I3D) model, a popular choice for driver observation, with Gradient-weighted Class Activation Mapping (Grad-CAM) for detailed analysis of model decision-making. Though the simulator-based model clearly surpasses the random baseline, its recognition quality diminishes, with average accuracy dropping from 85.7% to 46.6%. We also observe strong variations across different behavior classes. This underscores the challenges of model transferability, facilitating our research of more robust driver observation systems capable of dealing with real driving conditions.

The digitization of documents allows for wider accessibility and reproducibility. While automatic digitization of document layout and text content has been a long-standing focus of research, this problem in regard to graphical elements, such as statistical plots, has been under-explored. In this paper, we introduce the task of fine-grained visual understanding of mathematical graphics and present the Line Graphics (LG) dataset, which includes pixel-wise annotations of 5 coarse and 10 fine-grained categories. Our dataset covers 520 images of mathematical graphics collected from 450 documents from different disciplines. Our proposed dataset can support two different computer vision tasks, i.e., semantic segmentation and object detection. To benchmark our LG dataset, we explore 7 state-of-the-art models. To foster further research on the digitization of statistical graphs, we will make the dataset, code, and models publicly available to the community., Comment: Accepted at The 17th International Conference on Document Analysis and Recognition (ICDAR 2023)

Here we show that the Gross-Pitaevskii equation (GPE) for Bose-Einstein condensates (BECs) admits hydrodynamic interpretation in a general Riemannian metric, and show that in this metric the momentum equation has a new term that is associated with local curvature and density distribution profile. In particular conditions of steady state a new Einstein's field equation is determined in presence of negative curvature. Since GPE governs BECs defects that are useful, analogue models in cosmology, a relativistic form of GPE is also considered to show connection with models of analogue gravity, thus providing further grounds for future investigations of black hole dynamics in cosmology.

In this work we consider the Gross-Pitaevskii equation for Bose-Einstein condensates (BECs) in a general Riemannian metric. Given initial conditions dictated by an external potential, we consider the free expansion of the condensate when the external potential is turned off. Focusing on the forces associated with the geometry of the initial configuration, we show how these are related to the Ricci curvature tensor and the Ricci scalar and we find an Einstein field equation governing the steady flow. Some important correlations between the study of defects in BECs and the appearance of cosmological singularities will be addressed, in particular the emergence of an effective Lorentzian spacetime geometry, which is what is needed to obtain Hawking radiation effects.

Modality selection is an important step when designing multimodal systems, especially in the case of cross-domain activity recognition as certain modalities are more robust to domain shift than others. However, selecting only the modalities which have a positive contribution requires a systematic approach. We tackle this problem by proposing an unsupervised modality selection method (ModSelect), which does not require any ground-truth labels. We determine the correlation between the predictions of multiple unimodal classifiers and the domain discrepancy between their embeddings. Then, we systematically compute modality selection thresholds, which select only modalities with a high correlation and low domain discrepancy. We show in our experiments that our method ModSelect chooses only modalities with positive contributions and consistently improves the performance on a Synthetic-to-Real domain adaptation benchmark, narrowing the domain gap., Comment: 14 pages, 6 figures, Accepted at ECCV 2022 OOD workshop

Domain shifts, such as appearance changes, are a key challenge in real-world applications of activity recognition models, which range from assistive robotics and smart homes to driver observation in intelligent vehicles. For example, while simulations are an excellent way of economical data collection, a Synthetic-to-Real domain shift leads to a > 60% drop in accuracy when recognizing activities of Daily Living (ADLs). We tackle this challenge and introduce an activity domain generation framework which creates novel ADL appearances (novel domains) from different existing activity modalities (source domains) inferred from video training data. Our framework computes human poses, heatmaps of body joints, and optical flow maps and uses them alongside the original RGB videos to learn the essence of source domains in order to generate completely new ADL domains. The model is optimized by maximizing the distance between the existing source appearances and the generated novel appearances while ensuring that the semantics of an activity is preserved through an additional classification loss. While source data multimodality is an important concept in this design, our setup does not rely on multi-sensor setups, (i.e., all source modalities are inferred from a single video only.) The newly created activity domains are then integrated in the training of the ADL classification networks, resulting in models far less susceptible to changes in data distributions. Extensive experiments on the Synthetic-to-Real benchmark Sims4Action demonstrate the potential of the domain generation paradigm for cross-domain ADL recognition, setting new state-of-the-art results. Our code is publicly available at https://github.com/Zrrr1997/syn2real_DG, Comment: 8 pages, 7 figures, to be published in IROS 2022

Failure to timely diagnose and effectively treat depression leads to over 280 million people suffering from this psychological disorder worldwide. The information cues of depression can be harvested from diverse heterogeneous resources, e.g., audio, visual, and textual data, raising demand for new effective multi-modal fusion approaches for automatic estimation. In this work, we tackle the task of automatically identifying depression from multi-modal data and introduce a sub-attention mechanism for linking heterogeneous information while leveraging Convolutional Bidirectional LSTM as our backbone. To validate this idea, we conduct extensive experiments on the public DAIC-WOZ benchmark for depression assessment featuring different evaluation modes and taking gender-specific biases into account. The proposed model yields effective results with 0.89 precision and 0.70 F1-score in detecting major depression and 4.92 MAE in estimating the severity. Our attention-based fusion module consistently outperforms conventional late fusion approaches and achieves competitive performance compared to the previously published depression estimation frameworks, while learning to diagnose the disorder end-to-end and relying on far fewer preprocessing steps., Comment: Accepted to ECCV 2022 ACVR Workshop. Code is publicly available at https://github.com/PingCheng-Wei/DepressionEstimation

Messenger RNA isoform differences are predominantly driven by alternative first, internal, and last exons. Despite the importance of classifying exons to understand isoform structure, few tools examine isoform-specific exon usage. We recently observed that alternative transcription start sites often arise near internal exons, often creating “hybrid” first/internal exons. To systematically detect hybrid exons, we built the hybrid-internal-terminal (HIT) pipeline to classify exons depending on their isoform-specific usage. On the basis of splice junction reads in RNA sequencing data and probabilistic modeling, the HIT index identified thousands of previously misclassified hybrid first-internal and internal-last exons. Hybrid exons are enriched in long genes and genes involved in RNA splicing and have longer flanking introns and strong splice sites. Their usage varies considerably across human tissues. By developing the first method to classify exons according to isoform contexts, our findings document the occurrence of hybrid exons, a common quirk of the human transcriptome.

Driver observation models are rarely deployed under perfect conditions. In practice, illumination, camera placement and type differ from the ones present during training and unforeseen behaviours may occur at any time. While observing the human behind the steering wheel leads to more intuitive human-vehicle-interaction and safer driving, it requires recognition algorithms which do not only predict the correct driver state, but also determine their prediction quality through realistic and interpretable confidence measures. Reliable uncertainty estimates are crucial for building trust and are a serious obstacle for deploying activity recognition networks in real driving systems. In this work, we for the first time examine how well the confidence values of modern driver observation models indeed match the probability of the correct outcome and show that raw neural network-based approaches tend to significantly overestimate their prediction quality. To correct this misalignment between the confidence values and the actual uncertainty, we consider two strategies. First, we enhance two activity recognition models often used for driver observation with temperature scaling-an off-the-shelf method for confidence calibration in image classification. Then, we introduce Calibrated Action Recognition with Input Guidance (CARING)-a novel approach leveraging an additional neural network to learn scaling the confidences depending on the video representation. Extensive experiments on the Drive&Act dataset demonstrate that both strategies drastically improve the quality of model confidences, while our CARING model out-performs both, the original architectures and their temperature scaling enhancement, leading to best uncertainty estimates.

Visual recognition inside the vehicle cabin leads to safer driving and more intuitive human-vehicle interaction but such systems face substantial obstacles as they need to capture different granularities of driver behaviour while dealing with highly limited body visibility and changing illumination. Multimodal recognition mitigates a number of such issues: prediction outcomes of different sensors complement each other due to different modality-specific strengths and weaknesses. While several late fusion methods have been considered in previously published frameworks, they constantly feature different architecture backbones and building blocks making it very hard to isolate the role of the chosen late fusion strategy itself. This paper presents an empirical evaluation of different paradigms for decision-level late fusion in video-based driver observation. We compare seven different mechanisms for joining the results of single-modal classifiers which have been both popular, (e.g. score averaging) and not yet considered (e.g. rank-level fusion) in the context of driver observation evaluating them based on different criteria and benchmark settings. This is the first systematic study of strategies for fusing outcomes of multimodal predictors inside the vehicles, conducted with the goal to provide guidance for fusion scheme selection., Comment: Accepted at Intelligent Vehicles Symposium 2022, IEEE

Autonomous vehicles utilize urban scene segmentation to understand the real world like a human and react accordingly. Semantic segmentation of normal scenes has experienced a remarkable rise in accuracy on conventional benchmarks. However, a significant portion of real-life accidents features abnormal scenes, such as those with object deformations, overturns, and unexpected traffic behaviors. Since even small mis-segmentation of driving scenes can lead to serious threats to human lives, the robustness of such models in accident scenarios is an extremely important factor in ensuring safety of intelligent transportation systems. In this paper, we propose a Multi-source Meta-learning Unsupervised Domain Adaptation (MMUDA) framework, to improve the generalization of segmentation transformers to extreme accident scenes. In MMUDA, we make use of Multi-Domain Mixed Sampling to augment the images of multiple-source domains (normal scenes) with the target data appearances (abnormal scenes). To train our model, we intertwine and study a meta-learning strategy in the multi-source setting for robustifying the segmentation results. We further enhance the segmentation backbone (SegFormer) with a HybridASPP decoder design, featuring large window attention spatial pyramid pooling and strip pooling, to efficiently aggregate long-range contextual dependencies. Our approach achieves a mIoU score of 46.97% on the DADA-seg benchmark, surpassing the previous state-of-the-art model by more than 7.50%. Code will be made publicly available at https://github.com/xinyu-laura/MMUDA., Comment: Code will be made publicly available at https://github.com/xinyu-laura/MMUDA

Traditional video-based human activity recognition has experienced remarkable progress linked to the rise of deep learning, but this effect was slower as it comes to the downstream task of driver behavior understanding. Understanding the situation inside the vehicle cabin is essential for Advanced Driving Assistant System (ADAS) as it enables identifying distraction, predicting driver's intent and leads to more convenient human-vehicle interaction. At the same time, driver observation systems face substantial obstacles as they need to capture different granularities of driver states, while the complexity of such secondary activities grows with the rising automation and increased driver freedom. Furthermore, a model is rarely deployed under conditions identical to the ones in the training set, as sensor placements and types vary from vehicle to vehicle, constituting a substantial obstacle for real-life deployment of data-driven models. In this work, we present a novel vision-based framework for recognizing secondary driver behaviours based on visual transformers and an additional augmented feature distribution calibration module. This module operates in the latent feature-space enriching and diversifying the training set at feature-level in order to improve generalization to novel data appearances, (e.g., sensor changes) and general feature quality. Our framework consistently leads to better recognition rates, surpassing previous state-of-the-art results of the public Drive&Act benchmark on all granularity levels. Our code is publicly available at https://github.com/KPeng9510/TransDARC., Comment: Accepted to IROS 2022. Code is publicly available at https://github.com/KPeng9510/TransDARC

Semantic segmentation benchmarks in the realm of autonomous driving are dominated by large pre-trained transformers, yet their widespread adoption is impeded by substantial computational costs and prolonged training durations. To lift this constraint, we look at efficient semantic segmentation from a perspective of comprehensive knowledge distillation and consider to bridge the gap between multi-source knowledge extractions and transformer-specific patch embeddings. We put forward the Transformer-based Knowledge Distillation (TransKD) framework which learns compact student transformers by distilling both feature maps and patch embeddings of large teacher transformers, bypassing the long pre-training process and reducing the FLOPs by >85.0%. Specifically, we propose two fundamental and two optimization modules: (1) Cross Selective Fusion (CSF) enables knowledge transfer between cross-stage features via channel attention and feature map distillation within hierarchical transformers; (2) Patch Embedding Alignment (PEA) performs dimensional transformation within the patchifying process to facilitate the patch embedding distillation; (3) Global-Local Context Mixer (GL-Mixer) extracts both global and local information of a representative embedding; (4) Embedding Assistant (EA) acts as an embedding method to seamlessly bridge teacher and student models with the teacher's number of channels. Experiments on Cityscapes, ACDC, NYUv2, and Pascal VOC2012 datasets show that TransKD outperforms state-of-the-art distillation frameworks and rivals the time-consuming pre-training method. The source code is publicly available at https://github.com/RuipingL/TransKD., Comment: The source code is publicly available at https://github.com/RuipingL/TransKD

Occlusions are universal disruptions constantly present in the real world. Especially for sparse representations, such as human skeletons, a few occluded points might destroy the geometrical and temporal continuity critically affecting the results. Yet, the research of data-scarce recognition from skeleton sequences, such as one-shot action recognition, does not explicitly consider occlusions despite their everyday pervasiveness. In this work, we explicitly tackle body occlusions for Skeleton-based One-shot Action Recognition (SOAR). We mainly consider two occlusion variants: 1) random occlusions and 2) more realistic occlusions caused by diverse everyday objects, which we generate by projecting the existing IKEA 3D furniture models into the camera coordinate system of the 3D skeletons with different geometric parameters. We leverage the proposed pipeline to blend out portions of skeleton sequences of the three popular action recognition datasets and formalize the first benchmark for SOAR from partially occluded body poses. Another key property of our benchmark are the more realistic occlusions generated by everyday objects, as even in standard recognition from 3D skeletons, only randomly missing joints were considered. We re-evaluate existing state-of-the-art frameworks for SOAR in the light of this new task and further introduce Trans4SOAR - a new transformer-based model which leverages three data streams and mixed attention fusion mechanism to alleviate the adverse effects caused by occlusions. While our experiments demonstrate a clear decline in accuracy with missing skeleton portions, this effect is smaller with Trans4SOAR, which outperforms other architectures on all datasets. Although we specifically focus on occlusions, Trans4SOAR additionally yields state-of-the-art in the standard SOAR without occlusion, surpassing the best published approach by 2.85% on NTU-120., Comment: Accepted to IEEE Transactions on Multimedia (TMM). Code is publicly available at https://github.com/KPeng9510/Trans4SOAR

We explore the problem of automatically inferring the amount of kilocalories used by human during physical activity from his/her video observation. To study this underresearched task, we introduce Vid2Burn -- an omni-source benchmark for estimating caloric expenditure from video data featuring both, high- and low-intensity activities for which we derive energy expenditure annotations based on models established in medical literature. In practice, a training set would only cover a certain amount of activity types, and it is important to validate, if the model indeed captures the essence of energy expenditure, (e.g., how many and which muscles are involved and how intense they work) instead of memorizing fixed values of specific activity categories seen during training. Ideally, the models should look beyond such category-specific biases and regress the caloric cost in videos depicting activity categories not explicitly present during training. With this property in mind, Vid2Burn is accompanied with a cross-category benchmark, where the task is to regress caloric expenditure for types of physical activities not present during training. An extensive evaluation of state-of-the-art approaches for video recognition modified for the energy expenditure estimation task demonstrates the difficulty of this problem, especially for new activity types at test-time, marking a new research direction. Dataset and code are available at https://github.com/KPeng9510/Vid2Burn., Comment: Accepted to CVPR 2022 CVPM Workshop. Dataset and code are available at https://github.com/KPeng9510/Vid2Burn

La ciencia de la mejora racial o eugenesia fue creada a fines del siglo XIX por Francis Galton. Basado en darwinismo social, esta ciencia se hizo popular entre la influyente élite intelectual latinoamericana, lo que condujo a la creación de organizaciones comprometidas con el mejoramiento racial de la población a través de la medicina sanitaria y el combate a las "amenazas degenerativas" representada por negros, mestizos, extranjeros, y discapacitados. Apoyado por la perspectiva histórica, este artículo investiga el proyecto utópico-racial del movimiento eugenésico en América Latina en los años 1920-1930. Por medio de una investigación documental sobre las correspondencias y publicaciones de los eugenistas brasileños y argentinos, analizamos las vicisitudes del campo intelectual que permitieron el desarrollo del racismo científico latinoamericano a través de la institucionalización de la educación eugenésica. En diálogo con la bibliografía especializada, demostramos cómo estos documentos son fuentes fundamentales para comprender la conexión Brasil-Argentina prevista por el movimiento eugenésico, así como la permanencia de esta red de apoyo intelectual durante el siglo XX., A ciência do melhoramento racial ou eugenia foi criada no final do século XIX por Francis Galton. Baseada no darwinismo social, essa ciência popularizou-se entre a influente elite intelectual latino-americana, levando à criação de organizações comprometidas com o aperfeiçoamento racial da população através do sanitarismo e do combate às “ameaças degenerativas” representadas por negros, mestiços, estrangeiros e deficientes. Amparado pela perspectiva histórica, o artigo investiga o projeto utópico-racial do movimento eugenista na América Latina nas décadas de 1920-1930. Por meio de uma pesquisa documental sobre as correspondências e publicações dos eugenistas brasileiros e argentinos, analisamos as vicissitudes do campo intelectual que permitiram o desenvolvimento do racismo científico latino-americano através da institucionalização da educação eugenista. Dialogando com a bibliografia especializada, demonstramos de que maneira tais documentos consistem em fontes fundamentais na compreensão da conexão Brasil-Argentina vislumbrada pelo movimento eugenista, bem como a permanência dessa rede de apoio intelectual ao longo do século XX., The science of racial improvement or eugenics was created in the late 19th century by Francis Galton. Based on social Darwinism, this science became popular among the influential Latin American intellectual elite, leading to the creation of organizations committed to the racial improvement of the population through sanitary medicine and combating the “degenerative threats” represented by black, mestizos, foreigners and handicapped people. Supported by historical perspective, the article investigate the utopian-racial project of the eugenic movement in Latin America during the years 1920 to 1930. Through a documentary research of publications by Brazilian and Argentine eugenicists, the article seeks to understand the vicissitudes of the intellectual field that allowed the development of Latin American scientific racism through the institutionalization of eugenic education. In dialogue with the specialized bibliography, it is demonstrated how these documents are fundamental sources in understanding of Brazil-Argentina connection envisioned by the eugenic movement, as well as the permanency of this intellectual support network during the 20th century.

Human affect recognition is a well-established research area with numerous applications, e.g., in psychological care, but existing methods assume that all emotions-of-interest are given a priori as annotated training examples. However, the rising granularity and refinements of the human emotional spectrum through novel psychological theories and the increased consideration of emotions in context brings considerable pressure to data collection and labeling work. In this paper, we conceptualize one-shot recognition of emotions in context -- a new problem aimed at recognizing human affect states in finer particle level from a single support sample. To address this challenging task, we follow the deep metric learning paradigm and introduce a multi-modal emotion embedding approach which minimizes the distance of the same-emotion embeddings by leveraging complementary information of human appearance and the semantic scene context obtained through a semantic segmentation network. All streams of our context-aware model are optimized jointly using weighted triplet loss and weighted cross entropy loss. We conduct thorough experiments on both, categorical and numerical emotion recognition tasks of the Emotic dataset adapted to our one-shot recognition problem, revealing that categorizing human affect from a single example is a hard task. Still, all variants of our model clearly outperform the random baseline, while leveraging the semantic scene context consistently improves the learnt representations, setting state-of-the-art results in one-shot emotion recognition. To foster research of more universal representations of human affect states, we will make our benchmark and models publicly available to the community under https://github.com/KPeng9510/Affect-DML., Comment: Accepted to IEEE International Conference on Automatic Face and Gesture Recognition 2021 (FG2021). Benchmark, models, and code are at https://github.com/KPeng9510/Affect-DML

OBJECTIVE: Adjuvant chemotherapy (AC) improves survival of patients with resected non-small cell lung cancer (NSCLC). However, the cisplatin-vinorelbine regimen has been associated with a significant risk of clinically relevant toxicity. We sought to evaluate the effectiveness, safety, and feasibility of AC for NSCLC patients in a real-world setting. METHODS: This was a single-center, retrospective cohort study of patients with stage I-III NSCLC undergoing surgery with curative intent between 2009 and 2018. AC was administered at the discretion of physicians. The patients were divided into two groups: AC group and no AC (control) group. Study outcomes included overall survival (OS) and recurrence-free survival (RFS), as well as the safety profile and feasibility of the cisplatin-vinorelbine regimen in a real-world setting. RESULTS: The study involved 231 patients, 80 of whom received AC. Of those, 55 patients received the cisplatin-vinorelbine regimen. Survival analyses stratified by tumor stage showed that patients with stage II NSCLC in the AC group had better RFS (p = 0.036) and OS (p = 0.017) than did those in the no AC group. Among patients with stage III NSCLC in the AC group, RFS was better (p < 0.001) and there was a trend toward improved OS (p = 0.060) in comparison with controls. Of those who received the cisplatin-vinorelbine regimen, 29% had grade 3-4 febrile neutropenia, and 9% died of toxicity. CONCLUSIONS: These results support the benefit of AC for NSCLC patients in a real-world setting. However, because the cisplatin-vinorelbine regimen was associated with alarming rates of toxicity, more effective and less toxic alternatives should be investigated., SCOPUS: ar.j, DecretOANoAutActif, info:eu-repo/semantics/published

Autonomous vehicles clearly benefit from the expanded Field of View (FoV) of 360-degree sensors, but modern semantic segmentation approaches rely heavily on annotated training data which is rarely available for panoramic images. We look at this problem from the perspective of domain adaptation and bring panoramic semantic segmentation to a setting, where labelled training data originates from a different distribution of conventional pinhole camera images. To achieve this, we formalize the task of unsupervised domain adaptation for panoramic semantic segmentation and collect DensePASS - a novel densely annotated dataset for panoramic segmentation under cross-domain conditions, specifically built to study the Pinhole-to-Panoramic domain shift and accompanied with pinhole camera training examples obtained from Cityscapes. DensePASS covers both, labelled- and unlabelled 360-degree images, with the labelled data comprising 19 classes which explicitly fit the categories available in the source (i.e. pinhole) domain. Since data-driven models are especially susceptible to changes in data distribution, we introduce P2PDA - a generic framework for Pinhole-to-Panoramic semantic segmentation which addresses the challenge of domain divergence with different variants of attention-augmented domain adaptation modules, enabling the transfer in output-, feature-, and feature confidence spaces. P2PDA intertwines uncertainty-aware adaptation using confidence values regulated on-the-fly through attention heads with discrepant predictions. Our framework facilitates context exchange when learning domain correspondences and dramatically improves the adaptation performance of accuracy- and efficiency-focused models. Comprehensive experiments verify that our framework clearly surpasses unsupervised domain adaptation- and specialized panoramic segmentation approaches., Comment: Accepted to IEEE Transactions on Intelligent Transportation Systems (IEEE T-ITS). Dataset and code will be made publicly available at https://github.com/chma1024/DensePASS. arXiv admin note: substantial text overlap with arXiv:2108.06383

Intelligent vehicles clearly benefit from the expanded Field of View (FoV) of the 360-degree sensors, but the vast majority of available semantic segmentation training images are captured with pinhole cameras. In this work, we look at this problem through the lens of domain adaptation and bring panoramic semantic segmentation to a setting, where labelled training data originates from a different distribution of conventional pinhole camera images. First, we formalize the task of unsupervised domain adaptation for panoramic semantic segmentation, where a network trained on labelled examples from the source domain of pinhole camera data is deployed in a different target domain of panoramic images, for which no labels are available. To validate this idea, we collect and publicly release DensePASS - a novel densely annotated dataset for panoramic segmentation under cross-domain conditions, specifically built to study the Pinhole-to-Panoramic transfer and accompanied with pinhole camera training examples obtained from Cityscapes. DensePASS covers both, labelled- and unlabelled 360-degree images, with the labelled data comprising 19 classes which explicitly fit the categories available in the source domain (i.e. pinhole) data. To meet the challenge of domain shift, we leverage the current progress of attention-based mechanisms and build a generic framework for cross-domain panoramic semantic segmentation based on different variants of attention-augmented domain adaptation modules. Our framework facilitates information exchange at local- and global levels when learning the domain correspondences and improves the domain adaptation performance of two standard segmentation networks by 6.05% and 11.26% in Mean IoU., Comment: Accepted to IEEE ITSC 2021. Dataset and code will be made publicly available at https://github.com/chma1024/DensePASS

Recognizing Activities of Daily Living (ADL) is a vital process for intelligent assistive robots, but collecting large annotated datasets requires time-consuming temporal labeling and raises privacy concerns, e.g., if the data is collected in a real household. In this work, we explore the concept of constructing training examples for ADL recognition by playing life simulation video games and introduce the SIMS4ACTION dataset created with the popular commercial game THE SIMS 4. We build Sims4Action by specifically executing actions-of-interest in a "top-down" manner, while the gaming circumstances allow us to freely switch between environments, camera angles and subject appearances. While ADL recognition on gaming data is interesting from the theoretical perspective, the key challenge arises from transferring it to the real-world applications, such as smart-homes or assistive robotics. To meet this requirement, Sims4Action is accompanied with a GamingToReal benchmark, where the models are evaluated on real videos derived from an existing ADL dataset. We integrate two modern algorithms for video-based activity recognition in our framework, revealing the value of life simulation video games as an inexpensive and far less intrusive source of training data. However, our results also indicate that tasks involving a mixture of gaming and real data are challenging, opening a new research direction. We will make our dataset publicly available at https://github.com/aroitberg/sims4action.

At the heart of all automated driving systems is the ability to sense the surroundings, e.g., through semantic segmentation of LiDAR sequences, which experienced a remarkable progress due to the release of large datasets such as SemanticKITTI and nuScenes-LidarSeg. While most previous works focus on sparse segmentation of the LiDAR input, dense output masks provide self-driving cars with almost complete environment information. In this paper, we introduce MASS - a Multi-Attentional Semantic Segmentation model specifically built for dense top-view understanding of the driving scenes. Our framework operates on pillar- and occupancy features and comprises three attention-based building blocks: (1) a keypoint-driven graph attention, (2) an LSTM-based attention computed from a vector embedding of the spatial input, and (3) a pillar-based attention, resulting in a dense 360-degree segmentation mask. With extensive experiments on both, SemanticKITTI and nuScenes-LidarSeg, we quantitatively demonstrate the effectiveness of our model, outperforming the state of the art by 19.0% on SemanticKITTI and reaching 30.4% in mIoU on nuScenes-LidarSeg, where MASS is the first work addressing the dense segmentation task. Furthermore, our multi-attention model is shown to be very effective for 3D object detection validated on the KITTI-3D dataset, showcasing its high generalizability to other tasks related to 3D vision., Comment: Accepted to IEEE Transactions on Intelligent Transportation Systems (T-ITS). Code is publicly available at https://github.com/KPeng9510/MASS

Beyond assigning the correct class, an activity recognition model should also be able to determine, how certain it is in its predictions. We present the first study of how welthe confidence values of modern action recognition architectures indeed reflect the probability of the correct outcome and propose a learning-based approach for improving it. First, we extend two popular action recognition datasets with a reliability benchmark in form of the expected calibration error and reliability diagrams. Since our evaluation highlights that confidence values of standard action recognition architectures do not represent the uncertainty well, we introduce a new approach which learns to transform the model output into realistic confidence estimates through an additional calibration network. The main idea of our Calibrated Action Recognition with Input Guidance (CARING) model is to learn an optimal scaling parameter depending on the video representation. We compare our model with the native action recognition networks and the temperature scaling approach - a wide spread calibration method utilized in image classification. While temperature scaling alone drastically improves the reliability of the confidence values, our CARING method consistently leads to the best uncertainty estimates in all benchmark settings., Comment: Accepted as oral at ICPR 2021

BACKGROUND: The European Society for Medical Oncology-Magnitude of Clinical Benefit Scale (ESMO-MCBS) is a validated, widely used tool developed to score the clinical benefit from cancer medicines reported in clinical trials. ESMO-MCBS scores assume valid research methodologies and quality trial implementation. Studies incorporating flawed design, implementation, or data analysis may generate outcomes that exaggerate true benefit and are not generalisable. Failure to either indicate or penalise studies with bias undermines the intention and diminishes the integrity of ESMO-MCBS scores. This review aimed to evaluate the adequacy of the ESMO-MCBS to address bias generated by flawed design, implementation, or data analysis and identify shortcomings in need of amendment. METHODS: As part of a refinement of the ESMO-MCBS, we reviewed trial design, implementation, and data analysis issues that could bias the results. For each issue of concern, we reviewed the ESMO-MCBS v1.1 approach against standards derived from Helsinki guidelines for ethical human research and guidelines from the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use, the Food and Drugs Administration, the European Medicines Agency, and European Network for Health Technology Assessment. RESULTS: Six design, two implementation, and two data analysis and interpretation issues were evaluated and in three, the ESMO-MCBS provided adequate protections. Seven shortcomings in the ability of the ESMO-MCBS to identify and address bias were identified. These related to (i) evaluation of the control arm, (ii) crossover issues, (iii) criteria for non-inferiority, (iv) substandard post-progression treatment, (v) post hoc subgroup findings based on biomarkers, (vi) informative censoring, and (vii) publication bias against quality-of-life data. CONCLUSION: Interpretation of the ESMO-MCBS scores requires critical appraisal of trials to understand caveats in trial design, imp, SCOPUS: re.j, info:eu-repo/semantics/published

Reproductive mode, ancestry, and climate are hypothesized to determine body size variation in reptiles but their effects have rarely been estimated simultaneously, especially at the intraspecific level. The common lizard (Zootoca vivipara ) occupies almost the entire Northern Eurasia and includes viviparous and oviparous lineages, thus representing an excellent model for such studies. Using body length data for >10,000 individuals from 72 geographically distinct populations over the species' range, we analyzed how sex‐specific adult body size and sexual size dimorphism (SSD) is associated with reproductive mode, lineage identity, and several climatic variables. Variation in male size was low and poorly explained by our predictors. In contrast, female size and SSD varied considerably, demonstrating significant effects of reproductive mode and particularly seasonality. Populations of the western oviparous lineage (northern Spain, south‐western France) exhibited a smaller female size and less female‐biased SSD than those of the western viviparous (France to Eastern Europe) and the eastern viviparous (Eastern Europe to Far East) lineages; this pattern persisted even after controlling for climatic effects. The phenotypic response to seasonality was complex: across the lineages, as well as within the eastern viviparous lineage, female size and SSD increase with increasing seasonality, whereas the western viviparous lineage followed the opposing trends. Altogether, viviparous populations seem to follow a saw‐tooth geographic cline, which might reflect the nonmonotonic relationship of body size at maturity in females with the length of activity season. This relationship is predicted to arise in perennial ectotherms as a response to environmental constraints caused by seasonality of growth and reproduction. The SSD allometry followed the converse of Rensch's rule, a rare pattern for amniotes. Our results provide the first evidence of opposing body size— climate relati

A rapidly growing amount of content posted online, such as food recipes, opens doors to new exciting applications at the intersection of vision and language. In this work, we aim to estimate the calorie amount of a meal directly from an image by learning from recipes people have published on the Internet, thus skipping time-consuming manual data annotation. Since there are few large-scale publicly available datasets captured in unconstrained environments, we propose the pic2kcal benchmark comprising 308,000 images from over 70,000 recipes including photographs, ingredients and instructions. To obtain nutritional information of the ingredients and automatically determine the ground-truth calorie value, we match the items in the recipes with structured information from a food item database. We evaluate various neural networks for regression of the calorie quantity and extend them with the multi-task paradigm. Our learning procedure combines the calorie estimation with prediction of proteins, carbohydrates, and fat amounts as well as a multi-label ingredient classification. Our experiments demonstrate clear benefits of multi-task learning for calorie estimation, surpassing the single-task calorie regression by 9.9%. To encourage further research on this task, we make the code for generating the dataset and the models publicly available., Comment: Accepted to ICPR 2020

© 2019 Elsevier B.V. In this study, the performance of the combined-source variational data assimilation scheme (CS-VDA) is assessed in detail using in situ heat fluxes (i.e. sensible heat (H) and latent heat (LE)) collected at a Eucalypt forest savanna of Northern Australia (Howard Springs). The CS VDA scheme estimates surface turbulent heat fluxes via assimilation of sequences of land surface temperature (LST) and meteorological data into a surface energy balance model and a dynamic model. The main objectives of this paper were to extend previous studies to a semi-arid ecosystem and to evaluate the potential of using global meteorological forcing data (GMD) to drive the CS VDA model (rather than in-situ meteorological observations). In order to study the new errors associated with the use of GMD, the effects on LE of the uncertainty in air temperature and wind speed (the two key meteorological factors that controls the total estimation error) was quantitatively characterized. Using hourly in-situ measurements as inputs, the daily-averaged LE RMSE daily was 54 W/m 2 , which agrees with the errors previously reported in the literature. As expected, replacing local meteorological data with GMD reduces the performance of the LE estimation (GMA: RMSE daily = 82 W/m 2 , GLDAS: RMSE daily = 151 W/m 2 ). However, LE RMSE values at 8-day temporal scale for GMA are RMSE 8-days = 32 W/m 2 , similar to those reported in this area for other models (MODIS (MOD16A2) and Breathing Earth System Simulator (BESS)). The error propagation analysis indicate that the CS VDA model is very sensitive to uncertainties in wind speed measurements. Moreover, there are large discrepancies between in situ and GMD wind speed. These two factors combined can explain the degradation in LE estimations. In this context, our study is a first step towards the characterization of an operational daily LE estimation scheme using hourly LST observations.

Novelty detection is crucial for real-life applications. While it is common in activity recognition to assume a closed-set setting, i.e. test samples are always of training categories, this assumption is impractical in a real-world scenario. Test samples can be of various categories including those never seen before during training. Thus, being able to know what we know and what we do not know is decisive for the model to avoid what can be catastrophic consequences. We present in this work a novel approach for identifying samples of activity classes that are not previously seen by the classifier. Our model employs a voting-based scheme that leverages the estimated uncertainty of the individual classifiers in their predictions to measure the novelty of a new input sample. Furthermore, the voting is privileged to a subset of informed classifiers that can best estimate whether a sample is novel or not when it is classified to a certain known category. In a thorough evaluation on UCF-101 and HMDB-51, we show that our model consistently outperforms state-of-the-art in novelty detection. Additionally, by combining our model with off-the-shelf zero-shot learning (ZSL) approaches, our model leads to a significant improvement in action classification accuracy for the generalized ZSL setting., Comment: Published in BMVC 2018. First and second authors contributed equally to this work

The development of accurate and transferable machine learning (ML) potentials for predicting molecular energetics is a challenging task. The process of data generation to train such ML potentials is a task neither well understood nor researched in detail. In this work, we present a fully automated approach for the generation of datasets with the intent of training universal ML potentials. It is based on the concept of active learning (AL) via Query by Committee (QBC), which uses the disagreement between an ensemble of ML potentials to infer the reliability of the ensemble's prediction. QBC allows the presented AL algorithm to automatically sample regions of chemical space where the ML potential fails to accurately predict the potential energy. AL improves the overall fitness of ANAKIN-ME (ANI) deep learning potentials in rigorous test cases by mitigating human biases in deciding what new training data to use. AL also reduces the training set size to a fraction of the data required when using naive random sampling techniques. To provide validation of our AL approach we develop the COMP6 benchmark (publicly available on GitHub), which contains a diverse set of organic molecules. Through the AL process, it is shown that the AL-based potentials perform as well as the ANI-1 potential on COMP6 with only 10% of the data, and vastly outperforms ANI-1 with 25% the amount of data. Finally, we show that our proposed AL technique develops a universal ANI potential (ANI-1x) that provides accurate energy and force predictions on the entire COMP6 benchmark. This universal ML potential achieves a level of accuracy on par with the best ML potentials for single molecule or materials, while remaining applicable to the general class of organic molecules comprised of the elements CHNO., Comment: Accepted at J. Chem. Phys

This booklet contains 11 modules (290 multiple-choice items) designed for use in the first course of a three-course calculus sequence using the textbook "Calculus with Analytic Geometry" (Dennis G. Zill). In each module, relevant sections of the textbook are identified for users. It can, however, be used in conjunction with any calculus textbook. (Author/YP)