Generalized Intelligent Framework for Tutoring (GIFT)¶
GIFT is an empirically-based, service-oriented framework of tools, methods and standards to make it easier to author adaptive instructional systems (AIS), manage instruction and assess the effect of AIS, components and methodologies. GIFT is being developed under the Adaptive and Intelligent Training Systems (AITS) Team at the Learning in Intelligent Tutoring Environments (LITE) Laboratory, part of the U.S. Army DEVCOM Soldier Center - Soldier Effectiveness Directorate (SED).
The technology gap for a reusable AIS framework to support individual and small team tutoring was identified through a review of the intelligent tutoring systems (ITS) literature in 2009-2010 and the Training and Doctrine Command (TRADOC) formalized the Army requirement for GIFT in their Army Learning Concept (2011). GIFT was brought to practice in 2011 by the LITE Lab team. The first public demonstration of GIFT was conducted at the Interservice/Industry Training Systems and Education Conference (I/ITSEC) in December 2011. The first release of GIFT was completed in May 2012.
While GIFT is being developed to facilitate the use of AIS by the U.S. Army, the intent is to collaboratively develop GIFT and have it function as a “nexus” for AIS research being conducted within government, industry and academia.
Research Goals¶Adaptive Tutoring research goals that are driving future GIFT development include:
- an ontology to represent the knowledge and concepts of AIS, their relationships and their interactions
- automated tools and processes to author AIS, their models, and components
- automated delivery methods for tailored instruction to individuals and small teams
- a set of reusable, domain-independent modules that include standard processes and structures
- a testbed methodology to support the systematic comparison and assessment of AIS , their components, and instructional methods
Obtaining GIFT Software¶
After registering on this site, the GIFT software, including source code, is available for download on the Downloads tab
GIFT consists of several core modules which interact with each other to perform the functions within an AIS. The Sensor Module has interfaces to support commercial sensors (e.g., Affectiva Q-Sensor) and its function is to format, process and store sensor data. The Domain Module provides domain content to support training, assesses trainee performance against standards, and provides domain-specific feedback to the trainee when the Pedagogical Module identifies the need for feedback based on trainee performance. The Trainee Module uses trainee performance, historical data (e.g., past performance) and sensor data to determine the trainee’s cognitive and affective state.
Each of these Modules is interchangeable through the virtue of interfacing standards. This allows each Module designer to select the type of approach that they believe is suited towards instruction. For instance, a sample configuration may have a webcam sensor that interprets Facial Action Units (FACs), a rule-based performance assessment, a Feedback Generation Engine that generates varying levels of hints upon request, a finite state machine of trainee assessment, and pedagogy that gives hints on failed problems.This release features functional elements of each of the modules in a game-based tutoring domain. It does this on top of the training content created by the Games for Training (http://www.peostri.army.mil/PRODUCTS/USAGFTP/).
- Sensor Module - includes the Affectiva Q sensor to measure electrodermal activity (EDA); a custom mouse sensor to measure EDA; and a software-based "Self Assessment" sensor that can be manipulated for testing
- Learner Module - functions as a prototype pass-through module, collating information from VBS2 and the respective sensors
- Pedagogical Module - acts to give hints on concepts which are 'underperforming'
- Domain Module - listens to VBS2 traffic, assesses performance on 4 different tasks, supports varying levels of feedback as requested
In addition to functional Modules, several additional components are included for researchers, including:
- The ability to author your research questions with the Survey Authoring Tool
- The functionality to keep track of users and historical performance inside of a mySQL database
- The ability to analyze sensor/performance/assessed data after an experiment in a collated .csv file, with your choice of software
A longer description of the current project and goals is available under the Documents Tab. This paper is the recommended paper for citation.
If you are looking to develop with GIFT, we recommend the "Unwrapping GIFT" series of documents, found here (http://ceur-ws.org/Vol-1009/0702.pdf) and in the GIFTSym proceedings each year.
The software documentation for GIFT is extensive, updated, and complete. It is available with each GIFT download, in the Downloads section.
Other information such as answers to the "I'm wondering where to get started...?" and "How can I automatically conduct an experiment with GIFT?" questions can be found in the GIFT Forums (https://gifttutoring.org/boards/6/topics/205) and (https://gifttutoring.org/boards/6/topics/342), respectively.
An online version of the GIFT release documents can be found here.
|2023-1||Nov 2023||Current Official Release(Release Notes), also released online at https://cloud.gifttutoring.org|
|2022-1||Nov 2022||(Release Notes), also released online at https://cloud.gifttutoring.org|
|2021-2||Nov 2021||(Release Notes)|
|2021-1||April 2021||(Release Notes)|
|2020-1||May 2020||(Release Notes)|
|2019-1||April 2019||(Release Notes)|
|2018-12-22||December 2018||Interim release (Release Notes)|
|2018-10-05||October 2018||Interim release (Release Notes)|
|2018-1||July 2018||(Release Notes)|
|2017-12-22||Dec 2017||Interim release (Release Notes)|
|2017-1||July 2017||(Release Notes)|
|2017-02-22||February 2017||Interim release (Release Notes)|
|1.0||May 2012||Initial Release|