Duration
4 weeks
Tools
Adobe Illustrator
Project brief
National call for women and non-binary students in STEM to 250-word abstract on any computer science topic, only 100 abstracts accepted
A1 poster on the submitted abstract
1. Research
. . . the challenge
To research a computer science topic and present an idea or product that aligns with my interests, skills, and knowledge. Portray that research into a poster, and present it at a national academic conference.
What I discovered:
- My knowledge gaps with G.A.I. and building web extensions led me to focus on building a 'proof of concept', rather than a fully working product
. . . the constraints
No technical knowledge on generative A.I.: No previous experience or knowledge building generative A.I. models, curating datasets, creating a suitable algorithm, etc.
No technical knowledge building web extensions: No previous experience or knowledge building generative A.I. models, curating datasets
Time: Only 4 weeks to bulid up my knowledge and create the poster, on top of my university work and part-time job
2. Building the Proof of Concept
Building an inaccessible website:
Coded an website with a multitude of standard WCAG issues.
An accessible website for the user, without the hassle:
Based off the user's profile and accessibility needs, website UI is adapted automatically using generative A.I. without the user needing to manually edit settings.
Coded an inaccessible website
Website UI is adapted for user's accessibility needs, maintains visual identity of the original website
3. Poster design
Drafting the poster layout:
Started by drafting out an initial, structured layout to organise the content.
Changed it based on the content I wanted to include
Drafting the layout
Finalising the layout after placing in content
Color scheme:
Goals: Playful, attention-grabbing, memorable
4. Presenting
Presenting my research:
- Presented my poster and discussed my research with an audience of 100+ attendees
Reflection:
I learned:
- About generative A.I. and the steps necessary to build a new high-quality G.A.I
- Mitigating bias in G.A.I. results
- Presenting 1-min high-level summaries of my research that captures people's attention and encourages further discussion
If I had more time:
- I'd research more about how to practically implement this idea as a web extension
- More research and attempting to build or adopt existing G.A.I. to use within my product
I plan to continue this research and explore this topic in more detail in my upcoming dissertation