Community Research and Development Information Service - CORDIS

H2020

Lingvist Report Summary

Project reference: 666367

Periodic Reporting for period 1 - Lingvist (Lingvist - the next generation language-learning software)

Summary of the context and overall objectives of the project

The objective of the Lingvist project is to develop Lingvist, the language-learning software, to make language learning 5–10 times faster.
Learning acceleration is achieved by combining three innovative working principles to language-learning methodology:

• statistical language corpus analysis;
• measurements of individual memory processes and abilities;
• mathematical optimisation of the individual learning time.


The implementation of mathematics and computer science makes language-learning software highly adaptive to every individual. The learning process is tailored individually under requirement that it would take minimal amount of time for every user.
The innovation project is based on the early version of the Lingvist software and has four main goals:
• increase the efficiency of the software by additional functionalities;
• scale up the software to be used by millions of users;
• add new language pairs and languages for specific purposes;
• test the performance of the software extensively together with universities.

Lingvist addresses EU-wide and even global challenges. The lack of language skills creates walls between countries and cultures even now, in the 21st century. The biggest problem with language learning is that it takes a huge amount of time. People in Europe spend ca 40 billion hours annually on learning languages. The same amount of working hours would cost 400 billion EUR every year! The progress in language learning is so slow that 95 % of people give up before they reach proficiency. Lingvist concentrates on adult language learners on the European market initially, with a view to expand to other customer segments and the US and Asian markets later.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

Work Package 1: Prototyping and design
WP1 comprises of 8 deliverables, of which 3 were delivered on time during Period 1 of the project.
Lingvist has delivered A/B testing framework, a first iteration of Motivation Design and the preparatory work for New User Interface to be tested. A/B testing framework is an enabler-type task which helps us make better decisions about the effect and desirability of feature iterations. Motivation Design and New User Interface aim to increase retention and consequently increase study progress.
We have continuously worked on improving Product Design (T1.8) based on Testing activities (WP4) and utilizing the A/B testing framework capacities (D1.1). We also put considerable effort into T1.2 that adjusted scope from various Additional Functionalities into improving the language course quality of French (T3.5) and English (T3.6) as target languages. We also tackled preparatory work for T1.3 – Content Sharing / Publications.

Work Package 2: Scaling up – Software
WP2 comprises of 4 deliverables, of which 2 were delivered on time during Period 1 of the project.
We finished D2.1 – changing Lingvist’s database structure from MySQL to NoSQL in order to serve millions of language learning on the programme at once. We also built a User Interface for Philologists under D2.4 which has lead to faster, and easier, language course creation.

Work Package 3: Scaling up - Languages
For T3.1-T3.4 we made the decision to postpone the development of Languages for Specific Purposes. The back-end structure of the initial software design did not let us to separate different language courses very easily. We had to re-architecture some parts of the back-end infrastructure to make it easier to add new courses and the development took extra time and effort.
For T3.5-T3.9 our amendment no 1 (to be submitted on the 01.04.2015) replaces 5 language modules with different 5 language modules that have a significantly higher market potential. After an extensive study which took into account what is the market needs, customer acquisition cost and expected life time value of users we calculated a priority list which language pairs should be produced in which order. The results of this study revealed that the biggest need is to learn English as a target language. Furthermore, as we have signed a strategic investment with Rakuten, which is the biggest e-commerce corporation in Asia we added Japanese-to-English, Chinese-to-English, Arabic-to-English, German-to-English and Portuguese-to-English language modules. To fit those additions to into the budget we had to delete German-to-French, French-to-German, Spanish-to-German, French-to-Russian and Spanish-to-Russian language modules. The total number of language modules remains the same as well the total person months with no affection to the foreseen costs.

Work Package 4: Testing
During the reporting period, we have undertaken both usability testing as well as efficiency testing. Usability testing serves the purpose of optimizing the product to meet user needs and raise the commercial viability of the end product. Efficiency testing validates our unique methodology to internal and external stakeholders, as well as helps us optimize the product from a philological/pedagogical aspect.
All planned testing streams have been undertaken, but only most viable activities will continue until the end of Project term.

Work Package 5: Demonstration, exploitation and communication
Efforts under WP5 during the reporting period include various bigger and smaller demonstration, exploitation and communication activities. These have been targeted mostly at the general consumer public (end user of the Lingvist product) via online media channel, but also at industry (edutech, startups) stakeholders and policy makers, publishers and other interested parties.

Work Package 6: Project management
Continuous project management and reporting activities, as conscious and reflexive risk management hav

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

By combining three novel approaches to language study (statistical prioritiy of vocabulary, memory measurement, algorithm calculations in real time), and developing the deliverables set up in this project, we have taken significant steps towards accelerating language learning for the general public.
Although official speed tests will be conducted in the next half of the innovation project, early results indicate that study time, study volume (vocabulary units) and study frequency have already risen significantly since the start of the project. The specific results on how each deliverable has pushed learning metrics higher, has been brought out in Deliverable reports.
Approximately 75% of users on Lingvist are 18-44 years old, matching well with the set focus on adult learners. However, we have received unexpectedly positive feedback on the product fit to also high school students and senior learners (60+).We will continue monitoring and comparing the results of the senior learners against other age groups, and other relevant demographic parameters to see if this is a viable marketing target group in the future.
Secondly, teacher feedback shows that many use Lingvist as a learning supplement in their classes without us requesting it. Teachers value that Lingvist is adaptive to individual progress (reduces classroom boredom for pupils) and helps teachers focus on particular problems that students keep having (optimizing shared classroom time, attention from teacher etc.). Using ready-made, extra curricular technology also creates variety in the classroom that otherwise requires extra effort from the teachers.

Related information

Record Number: 186678 / Last updated on: 2016-07-14