2iC (United Kingdom)

In machine learning applications, the reliability of predictions is significant for assisted decision and risk control. As an effective framework to quantify the prediction reliability, conformal prediction (CP) was commonly developed with <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> -nearest neighbor, known as CPKNN. However, the conventional CPKNN suffers from high variance and bias, and long computational time as the feature dimensionality increases. To address these limitations, a new CP framework-CP with shrunken centroids (CPSC) is proposed. It regularizes the class centroids to attenuate the irrelevant features and shrink the sample space for predictions and reliability quantification. We have also compared the CPSC with CP based on SVM, LightGBM, RF on the herbal medicine dataset, with the electronic nose as a case and assessed them in two tasks: 1) off-line prediction: the training set was fixed, and the accuracy on the testing set was evaluated and 2) online prediction with data augmentation: CPKNN and CPSC filtered unlabeled data to augment the training data based on the prediction reliability, and the final accuracy of the testing set was compared. The result shows that CPSC significantly outperformed CPKNN in both two tasks: 1) CPSC reached a significantly higher accuracy with lower computation cost, and with the same credibility output, CPSC generally achieved higher accuracy and 2) the data augmentation process with CPSC robustly manifested a statistically significant improvement in prediction accuracy with different reliability thresholds, and the augmented data were more balanced in classes. This novel CPSC provides higher prediction accuracy and better reliability quantification, which can assist in reliable decision support.

To design an individualised questionnaire to measure the impact of cancer and its treatments on quality of life (QoL). MATERIALS& METHODS: Design of the Cancer-Dependent Quality of Life (CancerDQoL) questionnaire was based on the Audit of Diabetes Dependent QoL (ADDQoL) questionnaire and related -DQoLs for other conditions. Item selection, face validity and content validity were established through clinician and patient ratings of the importance and relevance of 60 domains from the -DQoL Item Library, and semi-structured interviews with 25 English-speaking participants with a range of cancers attending a cancer centre in Zimbabwe (age range: 25-78 years; 16 women, 9 men). Ten interviews were subsequently conducted with UK English-speaking participants with a range of cancers attending Maggie's Centres in London and Dundee (age range: 40-76; 5 women, 5 men) to adapt the CancerDQoL for UK use. The first draft of the CancerDQoL contained 25 domain-specific items from the -DQoL Item Library plus four overview items. Zimbabwean participants indicated that cancer negatively impacted on all life domains included, except 'having children'. Weighted impact (impact ratings multiplied by importance) was most negative for 'sex life', 'depend on others' and 'physical capability'. The least negative weighted impact was found for 'having children', 'spiritual/religious life' and 'past medical/self-care'. UK interviews confirmed no new items were required. Face and content validity of the CancerDQoL is established for an adult sample of English-speaking cancer patients in Zimbabwe and confirmed in an adaptation following UK interviews.

International audience

Abstract In 1795, the private library of William Macintosh, author of Travels in Europe, Asia, and Africa (1782), was seized from his home in Avignon by the revolutionary authorities. Using the inventory of this seizure as a guide, this note presents a reconstruction of Macintosh’s library and identifies those books that exist today in the collections of the Bibliothèque municipale d’Avignon.

We present a novel and generalised notion of doping cleanness for cyber-physical systems that allows for perturbing the inputs and observing the perturbed outputs both in the time– and value–domains. We instantiate our definition using existing notions of conformance for cyber-physical systems. We show that our generalised definitions are essential in a data-driven method for doping detection and apply our definitions to a case study concerning diesel emission tests.

This chapter looks at the case of running batch processes—different from steady web servers—using the Job controller. Kubernetes’ parallelization capabilities, which help reduce the total processing time of large, computational-expensive, multiple-item batch jobs are also given special attention.