103 research outputs found
A simple landscape design framework for biodiversity conservation
No full textLocal government planning agencies play an important role in conserving biodiversity in human-altered landscapes. Such agencies frequently have a limited knowledge of wildlife biology and few resources to carry out research, and therefore require simple, practical guidelines for biodiversity conservation. We propose a landscape design framework for biodiversity conservation that is sequential, prescriptive, and supported by current landscape ecological science. Unlike existing guidelines, our framework can be implemented in any given landscape using only land cover data and it explicitly considers constraints on land use planning. The steps of our framework, in the order in which they should be implemented are: (1) select land cover data and decide which land cover classes constitute unaltered or altered land covers; (2) list the constraints on land use planning (e.g., economic, social) that exist for the landscape; (3) maximize the total amount and diversity of unaltered land cover, especially near water; (4) minimize human disturbance within altered land cover, especially near water; and (5) aggregate altered land covers associated with high-intensity land uses, especially away from water. We illustrate the utility of our approach by applying it to a hypothetical landscape and comparing the outcome to those from the application of traditional ecological guidelines to inform land use planning
Neuro-symbolic learning of answer set programs from raw data
No full textArtificial Intelligence (AI) is becoming increasingly integrated in society. This raises serious questions of the robustness of the technology. Neuro-Symbolic AI is well-suited to help answer these questions. Many existing neuro-symbolic systems focus on training a neural network given manually engineered background knowledge. However, this can be laborious, and the complete knowledge may not be available. The question is, can we leverage recent advancements in the robustness and expressivity of symbolic machine learning to learn knowledge from raw data? To integrate both neural and symbolic learners, there are various challenges to address: (1) Can exact symbolic knowledge be learned from noisy neural network predictions? (2) Can symbolic knowledge be learned in an end-to-end fashion, since symbolic learners are not differentiable? (3) Can a neuro-symbolic learning framework scale to complex tasks? In this thesis, we propose three novel approaches, capable of learning highly expressive knowledge in the language of Answer Set Programming, defined in terms of primitive concepts extracted from raw data. First, we investigate the robustness of two state-of-the-art symbolic learners when learning complex knowledge, given noisy concept predictions from pre-trained neural networks. Empirical results demonstrate that symbolic knowledge can be learned accurately, even when large percentages of training data are subject to distributional shifts, which causes the networks to predict incorrectly with high confidence. Second, we develop an end-to-end architecture that iteratively trains both neural and symbolic components. We demonstrate that our approach outperforms a variety of baselines, achieving state-of-the-art results. Finally, we explore scalable solutions for neuro-symbolic learning, and develop an architecture that leverages the implicit knowledge embedded within large vision-language foundation models, to extract primitive concepts from raw data. Our evaluation demonstrates the scalability and data efficiency gained with respect to state-of-the-art neuro-symbolic AI methods, in cases where the symbolic knowledge is both given, and learned.Open Acces
Diagnosis of paediatric infectious diseases and the utility of host immune signatures
No full textBackground: It can be difficult to distinguish between bacterial and viral infections in children. Current diagnostics lack the sensitivity to confidently rule out serious bacterial infections. The aims of this thesis include discovery of novel host biomarkers to identify bacterial infections, and evaluation of the role of expanded molecular testing in detecting pathogens. Methods: The data in this thesis come from two prospective cohort studies recruiting children admitted to Patan Hospital, a tertiary-level hospital in the Kathmandu Valley, Nepal. The first cohort study recruited children with lower respiratory tract infections (LRTIs). Whole blood samples from these children were sent for RNA-sequencing and plasma samples were sent for mass spectrometry proteomics and cytokine analysis. The LRTI cases were grouped together based on aetiology; RNA and protein levels were compared between different LRTI groups. Data were divided into training and test datasets; using a partial least squares approach, signatures were identified to differentiate between bacterial and viral LRTIs. The data in the RNA and protein platforms were integrated to identify a multi-platform signature. The second cohort study recruited children with any signs of infection. Causes of infection were described using routine diagnostics; samples were tested using four additional research molecular panels. Results: In the LRTI study, samples from 258 LRTI cases were tested. The median age was 2 years and 38% were female. The most common pathogens detected were RSV, influenza and Streptococcus pneumoniae. Using a training subset of cases classified as bacterial, n = 47, or viral, n = 53, a three-gene RNA signature was identified; this signature identified bacterial infections, in the test dataset, with a sensitivity of 89% and a specificity of 100%. Integrating data from the RNA and protein platforms identified a signature with 100% sensitivity and 94% specificity for identifying bacterial LRTIs. In the second cohort study, 574 cases were enrolled, median age was 3 years and 35% were female, 498 cases had additional molecular testing. Using routine diagnostics only, 28.7% (143/498) of cases were classified as having a confirmed bacterial or viral aetiology. After additional molecular testing, 45.6% (227/498) of cases had a likely pathogen identified. Molecular testing for RSV infection, influenza virus infection, Neisseria meningitidis, enterovirus infection and dengue fever showed potential diagnostic utility. Conclusions: The signatures identified in this thesis have the potential to improve the identification of serious bacterial infections. A small number of molecular tests which had the potential to alter clinical management were identified
S1 Text -
No full textTable A: Names and affiliations of The Digital Diagnostics for Africa Network Contributors. Table B: Author contributions (CRediT Taxonomy). (DOCX)</p
The Common Frame of Reference in Europe
No full textDiscusses the origins of the Draft Common Frame of Reference and assesses the need for further work in particular areas, taking as an example the subject of restitutionary damages for non-performance of a contract. Also assesses the possible relevance of the DCFR in work on African legal unity
Changes in neural activation underlying attention processing of emotional stimuli following treatment with positive search training in anxious children
No full textPrior research indicates that positive search training (PST) may be a promising home-based computerised treatment for childhood anxiety disorders. It explicitly trains anxious individuals in adaptive, goal-directed attention-search strategies to search for positive and calm information and ignore goal-irrelevant negative cues. Although PST reduces anxiety symptoms, its neural effects are unknown. The main aim of this study was to examine changes in neural activation associated with changes in attention processing of positive and negative stimuli from pre- to post-treatment with PST in children with anxiety disorders. Children's neural activation was assessed with functional magnetic resonance imaging (fMRI) during a visual-probe task indexing attention allocation to threat-neutral and positive-neutral pairs. Results showed pre- to post-treatment reductions in anxiety symptoms and neural reactivity to emotional faces (angry and happy faces, relative to neutral faces) within a broad neural network linking frontal, temporal, parietal and occipital regions. Changes in neural reactivity were highly inter-correlated across regions. Neural reactivity to the threat-bias contrast reduced from pre- to post-treatment in the mid/posterior cingulate cortex. Results are considered in relation to prior research linking anxiety disorders and treatment effects with functioning of a broad limbic-cortical network involved in emotion reactivity and regulation, and integrative functions linking emotion, memory, sensory and motor processes and attention control
Dissociable functional roles of the human action-observation network (Commentary on E. S. Cross et al.)
No full textChromosome 9p21 SNPs associated with multiple disease phenotypes correlate with ANRIL expression
No full textAuthor Summary Genetic variants on chromosome 9p21 have been associated with several important diseases including coronary artery disease, diabetes, and multiple cancers. Most of the risk variants in this region do not alter any protein sequence and are therefore likely to act by influencing the expression of nearby genes. We investigated whether chromosome 9p21 variants are correlated with expression of the three nearest genes ( CDKN2A , CDKN2B , and ANRIL ) which might mediate the association with disease. Using two different techniques to study effects on expression in blood from two separate populations of healthy volunteers, we show that variants associated with disease are all correlated with ANRIL expression, but associations with the other two genes are weaker and less consistent. Multiple genetic variants are independently associated with expression of all three genes. Although total expression levels of CDKN2A , CDKN2B , and ANRIL are positively correlated, individual genetic variants influence ANRIL and CDKN2B expression in opposite directions, suggesting a possible role of ANRIL in CDKN2B regulation. Our study suggests that modulation of ANRIL expression mediates susceptibility to several important human diseases
- …
