Waldo
Language translation and information management tool
Mile Two
Description
Intelligence analysts are often challenged with information overload. Waldo is a language translation and information retrieval tool that leverages open-source data models and modifies them in military domains to support intelligence analysts' needs to search through and make sense of large amounts of content. This project is Controlled Unclassified Information (CUI) which basically means that specific details are not sharable. What I learned through this process was how to be comfortable with a whole lot of ambiguity. The personas generated during this project were ever-changing and evolving, but we had to start somewhere. The designs and final representation that were created can not be shown, but I think the most interesting part of this project lives within the artifacts created to get an understanding of our users and their goals.
The current tool was built in a lab and the backend functionality is really impressive. Much like other tools that were built in research labs, desirability and design consistency was not top of mind. After better understanding what we could of our users, we utilized the artifacts that were generated to propose design enhancements and improvements that would help the users reach their goals.
By utilizing Human-Language technology, this project supports the translation, storage, organization, and retrieval of information from a variety of sources enabling users to understand numerous languages and cultures. This project was generated in the Air Force Research Lab at Wright Patterson Air Force Base.
The current tool was built in a lab and the backend functionality is really impressive. Much like other tools that were built in research labs, desirability and design consistency was not top of mind. After better understanding what we could of our users, we utilized the artifacts that were generated to propose design enhancements and improvements that would help the users reach their goals.
By utilizing Human-Language technology, this project supports the translation, storage, organization, and retrieval of information from a variety of sources enabling users to understand numerous languages and cultures. This project was generated in the Air Force Research Lab at Wright Patterson Air Force Base.
My Role: As lead UX researcher, my role was to gain alignment with the stakeholders and team on the direction of efforts, outcomes.
Approach
My design approach is to gather a deeper understanding of the problem space by creating artifacts that help us gain a greater connection with our users and their goals. This included creating artifacts and representations to align with stakeholders and define the direction of our work.
Knowledge Mapping
This exercise was really important for various stakeholders in the project to participate in because it helped us call out our assumptions and generate research questions.
This exercise was really important for various stakeholders in the project to participate in because it helped us call out our assumptions and generate research questions.
Competitive Analysis
Creating this artifact was pivotal in our understanding of the project goals from a stakeholder's perspective. We were able to better understand the goals from a language translation perspective, but the conversation that came out of this presentation helped us redirect our focus and lead down another path.
Creating this artifact was pivotal in our understanding of the project goals from a stakeholder's perspective. We were able to better understand the goals from a language translation perspective, but the conversation that came out of this presentation helped us redirect our focus and lead down another path.
Service Blueprint
This artifact helped align stakeholders with the ideal workflow we were designing for.
This artifact helped align stakeholders with the ideal workflow we were designing for.
Solution
After asking a lot of questions I went down many paths to explore problems that analysts and linguists face. The result of this exploration is a concept rooted in the understanding of an analyst and language-enabled analysts' workflow. Their goals and the goals of those on their team.
Utilizing our user personas, the design team presented a concept as a low-fidelity prototype that utilized an open-source Large Language Model into an already structured NLP pipeline. (not pictured here due to classification restrictions)
Utilizing our user personas, the design team presented a concept as a low-fidelity prototype that utilized an open-source Large Language Model into an already structured NLP pipeline. (not pictured here due to classification restrictions)
AI/ML Concept Ideation
Below are some initial, low-fidelity versions of the LLM/ML concepts. What is not shown here is a prototype of this proof of concept in action. It was super fun to work on.
Below are some initial, low-fidelity versions of the LLM/ML concepts. What is not shown here is a prototype of this proof of concept in action. It was super fun to work on.
