▩▩▩▩▩▩▩ WEEK 11 ▩▩▩▩▩▩▩
Presentation Comments
After my presentation, Andreas provided some crucial insights. One significant point he raised was the need to consider how my prototype could have a lasting impact. Initially, I was a bit puzzled because, in my view, fostering algorithm literacy was already aimed at providing knowledge that would benefit my audience for a lifetime. Reflecting on his comment, I realized he might have been referring to the enduring presence of my project, extending beyond the physical space.
Another noteworthy comment from Andreas pertained to the term 'Speculative Prototyping' used in my methods. He pointed out that the sources referring to speculative prototyping lacked reliability, and there was a notable absence of a clear definition to use it effectively in my research. In a subsequent consultation related to my cohort presentation, he suggested using 'prototyping of speculative outcomes' instead to convey the intended message accurately.
The last comment Andreas made was regarding my mention of the 'lack of relevance of the prototype to research.' I am currently investigating this matter and require some time to fully address it. While I don't perceive it as a major issue, I believe I can work out the details as I progress further into the prototyping stage.
Wireframing Prototypes
After careful consideration, I've opted to divide my prototype into three components: Part A, Part B, and Part C, instead of creating three separate prototypes. I believe that this approach will provide me with more space to craft a compelling speculative narrative and make it more persuasive.
In Part A, I will focus on the interactive aspect, enabling people to engage with the algorithm and gain initial insights into its functionality and potential applications. During this phase, my objective is to capture the audience's interest and attention, prompting them to pose questions about its operation, accuracy, and significance. I aim to stimulate critical thinking and reflection, fostering an environment where questions naturally arise.
Part B will offer an opportunity for participants to observe how the algorithm reaches its decisions. At this stage, I hope they will begin to question the fairness of the algorithm's decision-making process. My primary goal is to lead them to recognize the unethical nature of my algorithm. While it may appear as a useful tool on the surface, delving into the algorithm's inner workings should evoke a sense of distrust. I am also exploring ways to draw parallels between my prototype and commonly encountered algorithms in daily life to help the audience establish a connection.
Transitioning to Part C, I will unveil how these algorithms rank data and content. Given an understanding of my algorithm's operations, I will intentionally manipulate the content to favor specific items, disregarding the primary purpose of measuring representative diversity. This stage serves to encourage the audience to contemplate the consequences of access to transparent algorithms and the potential for algorithm abuse. Additionally, I aim to draw links between my narrative and the actions of major corporations, which can manipulate algorithms for financial gain, providing further educational value to the audience.
The visual style of my prototypes is still awaiting finalization. Currently, my primary emphasis is on the individual components within each section of the prototype. I am deliberately allocating additional time to carefully consider the visual style, as I recognize its potential to significantly enhance the immersive quality and establish the desired tone for my speculative scenario.
Prototype A Wireframe
Prototype B Wireframe
Prototype C Wireframe (unconfirmed)
Cohort Presentation
In preparation for my cohort presentation, I arranged an additional consultation with Andreas to discuss my concerns about the presentation outline that Vikas had forwarded via email. I found the sequence of the outline a bit confusing, particularly the positioning of the 'design statement' after the 'body of work.' In hindsight, I believe this variation in structure could be attributed to differences in the organization of each atelier. Therefore, I am confident that I can reorganize the order to enhance its alignment with my presentation.
My background and context section was quite extensive, as I considered it fundamental for providing the audience with a comprehensive understanding of the significance of my research. Several layers needed to be unpacked before I could delve into the presentation of my prototype; otherwise, it would not have been coherent. I placed a particular emphasis on the technical aspects because I felt less confident in elucidating the psychological facets of speculative design at this stage. I believe more extensive reading is necessary to substantiate my perspective, as currently, it may appear as if I’m just making assumptions.
Cohort Presentation Slides
Dealing with an Impasse
For the past two weeks, I've grappled with a persistent challenge. I’m trying to save the detected faces in my TouchDesigner files as images, and I have the values of the bounding boxes, but the only issue is how I can apply them and save these faces as images.
I’ve been looking all around the Internet for solutions but I just couldn’t find one that works with the current type of value I have. There were a couple of options but not all of them work with multi-face tracking - which is an important aspect of my prototype.
I’ve asked Andreas, Joanne, and Federico for help, but none of them had any idea how I could do this. Furthermore, I’ve reached out to Youtubers who create videos surrounding OpenCV and HaarCascade, but none of them had reached out to me yet. Following Federico’s advice, I sought help in Reddit.
5 days after making the post, I have only gotten 1 reply. Despite the reply being pretty vague, I do understand the logic behind it. The only problem now is my software literacy, I have to go and understand how all these works. I’ve been procrastinating this for pretty long, but now, I guess no more shortcuts!
Asking for help on Reddit
Software Learning Day
I dedicated a full day to learning more about TouchDesigner. However, I found myself unsure of where to begin, so I decided to start by researching the Operators present in my patch to gain a better understanding of their functions.
I realized that being in the right mindset was crucial for an effective learning day. To prepare, I made some adjustments to my routine. The previous night, I opted for an early wind-down, going to bed three hours earlier than usual. The following morning, I kicked off my learning day at 7 a.m. with a cup of coffee in hand. This environment allowed me to fully immerse myself in the learning process.
Previously, I had attempted to watch tutorials during my commute on the train and bus to and from work. However, these attempts were often hampered by distractions and fatigue, making it challenging to absorb new information. On my dedicated learning day, I was able to maintain my focus without the constant need to rewind tutorials and backtrack to grasp missed or unclear concepts.
During this dedicated learning day, I unearthed numerous valuable tips to streamline my workload and enhance my creative process. One such discovery was the efficient use of the Cache and CacheSelect TOP, which significantly reduced the amount of manual work required while ensuring that my CPU wasn't overly burdened. Additionally, I learned the merits of employing the Info CHOP to extract and utilize all the values within a specific operator. Another valuable tool I grasped was the replicator tool, which proved to be an excellent alternative to my initial idea of exporting each cropped face individually from TouchDesigner and then bringing them back into a designated folder. The replicator operator efficiently stored all the detected faces, and with a few straightforward lines of code, allowed for their overlay on my face detection results.
While my prototype was still posing challenges, I sought reassurance regarding the feasibility of other functions I had planned. To achieve this, I delved into creating a GUI overlay, exploring how to incorporate buttons, sliders, and variable text to be displayed on top of my face detection results. The process was relatively straightforward, but I held some concerns about the processing power required, given the numerous interconnected nodes within the same patch file.
In sum, this learning session proved highly productive as it provided a respite from my research work, allowing me to focus exclusively on mastering the software. There were moments when I was tempted to pause the video tutorials and experiment directly with my working file, but I opted to keep the flow going and took diligent notes instead. I believe that a couple more days like this will equip me to complete my prototype effectively.
Some of the tutorials I saved to refer back to.