LTfLL > overview > 5.2 – The LTfLL Online Synthesis Advisor PenSum

5.2 – The LTfLL Online Synthesis Advisor PenSum

Download: LTfLL-Service-Description

Problem: In numerous educational contexts, learners produce textual reports (e.g. summaries, essays, syntheses) about the learnt notions, and feedback is offered about their results. The problems they encounter are long waiting time for feedback (stagnating them in the writing process) and the limited feedback opportunities that do not stimulate explorative approaches (“what if-trials”), but force them to hand in mainly completed versions. During writing, it is difficult to self-assess ongoing work and to identify possible misunderstandings. The teacher has only a limited overview of the learners’ progress, may find out specific problems too late to use them during the current course, and may be unsure about the consistency of feedback given by tutors.

 

Solution: Service 5.2 will support learners in the automatic assessment of their essays (summaries, syntheses) in order to let teachers focus on higher-level activities (e.g., individual learner guidance or course design). Pensum analyses how well learners understand course texts as shown by their textual productions; it provides frequent just-in-time feedback on the ongoing writing activities (relevance of written sentences, inter-sentence coherence of the synthesis, résumé of each course sentence).

 

Story: Ulysses launches Pensum as a Web service. He selects the course domain Natural Language Processing and starts to express the main questions, problems and notions he wants to tackle in this course in a dedicated notepad. Then he starts to write a synthesis of the most important ideas of the course, according to his understanding. Whenever Ulysses is uncertain about whether he grasps the most important notions of a text, he asks support from Pensum. The system gives Ulysses feedback on his written synthesis, e.g. the relevancy of sentences, or the inter-sentence coherence of his synthesis. Ulysses is in control of his own learning process, he requests feedback whenever he wants and he can update his notepad according to the main points he understands and can continue writing on the same synthesis or one on another topic.

 

 

Key functionalities

 

Preparation

Define course domain

Add course texts

Train to system

Assign learners to a course domain

 

Learner work with Pensum

After selecting the course domain to synthesize, either the student writes a new synthesis or the student reads/revises a already-written synthesis.

 

Then students can ask a feedback about their synthesis and write learning questions in a notepad. These possibilities are available all the time and they can freely switch to one another.

 

Then feedback is prompted in a textual (in tooltips) and graphic form (underlining of some synthesis/course sentences).

 

Course texts, synthesis and content of notepad are stored in a database.

 

The feedback is based on cognitive and computational models of writing assessment by using LSA.

Techniques and data:

Semantic similarity computing (LSA)

General language corpus

Course corpus

Database for the course texts, synthesis, learning questions (notepad) and students’ tracks.

Cognitive and computational models.

Feedback generator.

Validation

Our validation goals are to investigate to which extent the:

·         students’ understanding is fostered through free exploration of the course content;

·         students can get feedback on their productions as often as they want;

·         teacher’s activity is more directed on higher levels of students’ activity (guidance, collaborative learning management, production assessment on style).

·         users find the software easy to use

·         users’ cognitive load is reduced.

·         students can get correct feedback (i.e., performing tests beforehand to measure whether the validity of cognitive models is satisfying)

 

We will validate the software in the domain of educational technologies in French. Both quantitative and qualitative analysis will be undertaken. We will run a pilot with 20 students in educational sciences (year 4) and 1 teacher. Instruments used are: semi-structured interviews with teacher to collect judgments and comments on pedagogical soundness; Questionnaires to students to collect data on their appreciation of the service, its usability and learning aspects. In addition we collect data on cognitive workload. We track students’ activity (number of loggings, synthesis, learning questions, feedback), and compare their grades of synthesis.

 
tl_files/Ltfll_pics/Gallerie/SUM 5.2.jpg

Link: Get in Touch