Annotation (informatique) | EPFL Graph Search

En programmation, une annotation est un élément permettant d'ajouter des méta-données à un code source. Selon le langage de programmation et ce qu'a choisi le programmeur, elles peuvent être accessibles uniquement lors de la compilation, présentes uniquement dans le fichier compilé, voire accessibles à l'exécution. Cette technique est une alternative aux fichiers de configuration, souvent écrits dans des formats tels que le XML ou le YAML. Selon le langage, les annotations peuvent être ajoutées aux classes, méthodes, attributs, paramètres, variables locales et paquetages ou même aux annotations elles-mêmes. Les annotations ou leurs équivalents existent pour Microsoft .NET (attributs .NET), Java (depuis la version 1.5), PHP (depuis la version 5) et Python (decorators). Des équivalents existent pour le langage C aux travers des attributs, qui sont des extensions non standard au langage C. Ils sont moins puissants que les annotations car spécifiques au compilateur utilisé et ils ne so

Expanding Eye-Tracking Methods to Explain the Socio-Cognitive Effects of Shared Annotations

Andrea Mazzei

Social technologies are leading to transformations in education, by empowering the way learners connect to each other, by introducing new means for teaching and learning and by reshaping the way knowledge is delivered. Annotating texts is a learning strategy that is spontaneously adopted by learners during lectures or while reading text books. Students still individually take annotations on paper, that tend to be thrown away when the learning goal is achieved. Students also engage occasionally in the spontaneous practice of note sharing. Our work explores experimentally the benefits of note-taking and note-sharing behaviour. First, we study how sharing student-annotated instructional texts can improve learning, enriching an experimental approach with a new eye tracking method. Second, we conduct experiments on computer-mediated note sharing in the classroom. Our results demonstrate the virtuous circle of note-taking: both annotating while reading and reading annotated documents lead to a better learning achievement. In the first experimental study, we measure if the presence of annotations, on an instruc- tional text, can influence the reading pattern, and how visual features of annotations can elicit the reader’s attention. To complement the results concerning learning and reading comprehension, we look into the readers’ gaze patterns, to explain how the students’ learning outcome relates to reading annotated texts. For this purpose, we design a novel eye tracker, that can be used to study reading and note taking in unconstrained experimental settings. This eye tracking system is based on a systematic error correction procedure that exploits the appearance similarity between the annotated texts and the spatial distribution of the fixation points. We show that this method can extract accurate gaze measures without introducing experimental constraints, that could disturb the note taking process and affect the readers’ comprehension. In the second study, we move from a controlled experimental setting to the classroom. We discuss how the use of technology can facilitate a spontaneous transition from personal to shared annotations and support students in the learning process. We complement the analysis by reporting a friendship bias in browsing the shared annotated material. We further speculate on the potential of shared annotations in triggering adaptations of the instructional material and teaching workflow. These two studies provided an insightful understanding of the effects of student-generated annotations on reading comprehension, and the underlying impact on a community of learners. The obtained findings should inspire further experimentation on social learning environments, meant to facilitate knowledge sharing through shared annotations and their diffusion within educational institutions.

EPFL2013

Learning to Reduce Annotation Load

Ksenia Konyushkova

Modern machine learning methods and their applications in computer vision are known to crave for large amounts of training data to reach their full potential. Because training data is mostly obtained through humans who manually label samples, it induces a significant cost. Therefore, the problem of reducing the annotation load is of great importance for the success of machine learning methods. We study the problem of reducing the annotation load from two viewpoints, by answering the questions âWhat to annotate?â and âHow to annotate?â. The question âWhat?â addresses the selection of a small portion of the data that would be sufficient to train an accurate model. The question âHow? focuses on minimising the effort of labelling each datapoint. The question âWhat to annotate?â becomes particularly compelling if we can select data to be annotated in an iterative and adaptive way, a setting known as active learning (AL). The key challenge in AL is to identify the datapoints that are the most informative for the model at a given stage. We propose several techniques to address this challenge. Firstly, we consider the problem of segmenting natural images and image volumes. We take advantage of image priors, such as smoothness of objects of interest, and use them in a novel form of geometric uncertainty. Using this, we design an AL technique to efficiently annotate data that is tailored to segmentation applications. Next, we notice that no single manually-designed strategy outperforms others in every application and that often the burden of designing new strategies outweighs the benefits of AL. To overcome this problem we suggest learning an AL strategy from data by formulating the AL problem as a regression task that predicts the reduction in the generalisation error achieved by labelling each datapoint. This enables us to learn AL strategies from simulated data and to transfer them to new datasets. Finally, we turn towards non-myopic data-driven AL strategies. To this end, we formulate the AL problem as a Markov decision process and find the best selection policy using reinforcement learning. We design the decision process such that the policy can be learnt for any ML model and transferred to diverse application domains. Effectively addressing the question âHow to annotate?â is of no less importance as large cost savings can be achieved by labelling each datapoint more efficiently. This can be done with intelligent interfaces that interact with a human annotator. We make two contributions towards answering the question âHow?â. Firstly, we propose an efficient technique to annotate 3D image volumes for image segmentation. Annotating data in 3D is cumbersome and an obvious way to facilitate it is to select a subset of the data lying on a 2D plane. To find the optimal plane (i.e. the one containing the most informative datapoints) we design a branch-and-bound algorithm that quickly eliminates hypotheses about the optimal projection. Secondly, we propose an intelligent data annotation method to train object detectors. Instead of always asking the human annotator to draw bounding boxes in images, we detect automatically in which cases we can rely on the current detector and verify its proposal.

EPFL2019

Attentional Processes in Natural Reading: the Effect of Margin Annotations on Reading Behaviour and Comprehension

Pierre Dillenbourg, Frédéric Kaplan, Tabea Koll, Andrea Mazzei

We present an eye tracking study to investigate how natural reading behavior and reading comprehension are influenced by in-context annotations. In a lab experiment, three groups of participants were asked to read a text and answer comprehension questions: a control group without taking annotations, a second group reading and taking annotations, and a third group reading a peer-annotated version of the same text. A self-made head-mounted eye tracking system was specifically designed for this experiment, in order to study how learners read and quickly re-read annotated paper texts, in low constrained experimental conditions. In the analysis, we measured the phenomenon of annotation-induced overt attention shifts in reading, and found that: (1) the reader's attention shifts toward a margin annotation more often when the annotation lies in the early peripheral vision, and (2) the number of attention shifts, between two different types of information units, is positively related to comprehension performance in quick re-reading. These results can be translated into potential criteria for knowledge assessment systems.

2014