Contributor
Mahinour Elsarky

AI-based Touristic Information Tool for LG (based on GEMMA model running locally)


Mentors
Andreuibanez, Claudia, Laura Morillo
Organization
Liquid Galaxy project
Technologies
python, docker, flutter, ai, dart, ml, LG core, Gemma AI model
Topics
virtualization, machine learning, mobile development, android
Seeking an escape from daily life chores, many turn to travel as a means of relaxation, renewal, and exploration. However, planning a trip can be overwhelming especially with the abundance of information available online that involves navigating multiple websites, guidebooks and apps to research. Furthermore, discovering the most touristic places in a foreign city or country, along with the best dining spots, shopping locales, and natural vistas, can be a daunting task. The main purpose of this application is to simplify the travel planning experience and inspire users to embark on memorable journeys and discover the most captivating point of interests (POIs) tailored to their preferences. The POIs would be generated by running GEMMA, one of the latest open-source generative text AI models, locally on the AI server in Lleida Lab utilizing Docker technology. Utilizing the liquid galaxy technology, the user would be able to visualize their entire trip on the LG rig with 3 or more screens where the tours along with their info would be sent through KML (keyhole markup language) that will help build beautiful and unique visualizations on the LG rig. In case of unavailability of an LG rig, the user would still have a captivating experience through the app and the integrated google maps. Deliverables - The model trained on the cloud and ready to be inferred in Lleida Lab AI local server. - The new App published on the Play Store under the Liquid Galaxy LAB account. - Full documentation and code on Liquid galaxy lab GitHub.