First Artificial Intelligence (AI) LMS and Online Training System for USCG & Treasury Department
ALICE Functionality-Innovations:
Given a Cold Fusion server, and a designated training area, ALICE is first programmed to find ‘new’ training documents. If new documents are found, ALICE logs its contents as a hierarchy of knowledge, using a cache of the system’s runtime ‘brain’. To do this, ALICE must identify all parts of speech, using a SQL database of this model. Once all parts of speech are mapped, a deeper learning model is applied to David Merrill’s content performance database, (map nouns to facts, concepts, processes and procedures to verbs etc). ALICE creates new values to DM’s SQL database (given any new fact, concept, process and procedure), and then translates this into artificial intelligence markup language (AIML), using a combination of server scripts and algorithms. Given a new AIML entry, so the server can read and display interaction in a manner we understand, ALICE is triggered to re-process this, this time updating and saving all runtime files as ‘new’. When a perpetual script on the cold fusion server ‘sees’ a given file is triggered as new, a search of all user profiles is then conducted. Once a user profile match is found to all related procedures to be taught, another query is then run to match each user session for any gaps in prerequisite knowledge. If any gaps in prerequisite knowledge are found, ALICE triggers an email script, alerting a designated training manager. If the training manager has created ALICE’s online training plan, another query is then run that maps all existing user lesson plans to existing knowledge, and training that’s known to exist, including SOPS and MPCs. If the manager has also taken the time to discuss training plans with their employee, and has completed a job task analysis form, and work breakdown structure form, that data is also scanned. Given any change in an existing work breakdown structures; however, the system then creates a change entry into a case-based reasoning (CBR) repository of the database model, for further analysis over time. If for instance a ticket to IT is created one day, which shares a relational table with ALICE, the system is triggered to create a log of potential training to be explored in this history as well. In the end, when any user queries ALICE, this queries the ‘brain’ to then begin conversing meaningful answers, on demand, while providing relevant links to training media. Whenever ALICE provides any answer, this is recorded as a session unique to each user and time, and as a metric including any unique IP address as a new training candidate, as an administrator report, which also aids in updating the ALICE system administrator to training opportunities which can be explored further.
— The following are endorsements of these accomplishments which supports these skills: “…I have worked with Eric Sandstrom at the U.S. Coast Guard Finance Center for about 3 years in the areas of training and performance improvement. Eric has taken many of our training processes to a higher level. He is extremely talented and innovative and has greatly improved our training events by applying his many artistic talents and innovative ideas to the development and creation of cutting-edge training concepts. I am also extremely impressed with Eric’s grasp of the “people” side of how organizations work. Eric has outstanding project management skills. I am always confident that when Eric takes on a project, it will be accomplished on time and with extremely high levels of quality. Eric is an extremely valuable part of our organization. Robert Bowles Operations Chief, OMB Branch U.S. Coast Guard Finance Center
