Deconstructing algorithms

This note features student work.

Deconstructing Algorithms 2020: Initial challenge brief given to the students

Algorithms are everywhere, from recommending which movie to watch next, automate grunt jobs, or accessing any information at fingertips. As described by Ed Finn in his book What Algorithms want (Finn 2017), he referred algorithms as culture machines, as complex assemblages of abstractions, processes, and people. As much as they might appear magical, behind them exist things, obscured from the end-user. Such information is potent of being used for or against the users in ways which can flip scenarios where our actions are informed by algorithms around us and not the other way around. Furthermore, this brings us to the questions: How much of our actions can be informed by algorithms? and How might we attempt to make sense of complex algorithms and their effects and make their information accessible to everyone?

Deconstructing Algorithms was offered as one of the thesis theme electives for the Undergraduate Design Program at Srishti Institute of Art, Design and Technology, Bangalore, India. Deconstructing Algorithms had invited aspiring practitioners to design experiments for sense-making and revealing algorithms by combining the contrasting viewpoints of understanding of the structured process of designing and programming algorithms, and the philosophical notion of deconstruction.

This note features student work.

Deconstructing Algorithms 2020: Initial challenge brief given to the students

Overview

Contributors

In clockwise direction: Daksha Dixit, Simran Singh, Apoorva Avadhana, Nupur Patny and Shreya Mishra

Daksha Dixit: Daksha is a design researcher passionate about education, healthcare, storytelling, and design thinking. Employing a variety of research techniques, she develops authentic solutions and uncovers relevant problems. Her experience includes projects in education, communication strategy, and curriculum design.

Simran Singh: Passionate about nature's systems and technology, Simran is a Human-Centered Design major. She works with emerging technologies to identify and design meaningful interactions, blending her curiosity and expertise to create innovative solutions.

Apoorva Avadhana: Driven by everyday inspirations, Apoorva combines her imaginative and critical thinking with user studies, technology, and business understanding. She focuses on transforming societal systems into usable and effective outcomes, ensuring delight and positive change.

Nupur Patny: Nupur specializes in both tangible and intangible aspects of design, crafting holistic design strategies and user experiences. Skilled in creating engaging, playful, and meaningful experiences, her designs transcend digital, physical, and service boundaries.

Shreya Mishra: Intrigued by complex natural and social systems, Shreya focuses on healthcare, ell-being, sustainability, behavioral economics, and information democratization. Her human-centric approach to design and research continually expands her ability to identify and address critical issues.

All rights belong to participants of the project.

Foreword by Gaurav

Algorithms and their implementation are often restricted to executives, designers, programmers, and mostly to a specific domain like computer science. This is a massive barrier for others to contribute to the making process or even make sense of it as the information comes from very discrete and structured forms (for, e.g., source code and system architecture). There is a need for creating passages into this field to allow more people and create a more inclusive environment. In one such effort, through this project, we envision to look at algorithms through the theories of deconstruction/post-structuralism, which can challenge the existing notions of algorithms in many ways — How might we attempt to understand algorithms beyond the existing structured ways? and What mediates between the algorithms and cultural reality? More such questions are worth engaging in the course of this project to bring forth this complex-verse of algorithms in front of everyone, especially those who could not participate due to restrictions posed by rigid compartmentalization of disciplines.

The first project, "Deciphering Algorithms" is an original work by Daksha Dixit and Simran Singh. They started off to simplify algorithms for the common, to make it so digestible that it enables everyone to understand the implications. Among many many things, in their final output, they take the inspiration from the code inspector in web browsers. Their final output imagines a similar inspection tool but for revealing all the mechanical clockwork of the underlying algorithmic system. Imagine knowing the decisions YouTube made before recommending you the dog video in your feed. Their concept explores possible ways of engaging with the information, in a manner that it does not overwhelm but scaffold them step by step to eventually help build self-efficacy. This report also documents the ups and downs in their journey to build their proposition into prototypes of their imagination.

The second project, "Greenbox Algorithms" is an original work by Apoorva Avadhana, Nupur Patny, and Shreya Mishra. It was nothing less than amazing to see the three coming together to push each other and the extent of what can be made possible through collaboration. What began on a piece of paper a few months ago is striving to come out in the real world. This is the first of the evidence of their hard work during this project and their concept details. Their project has imagined an alternate reality that amplifies rather more essential factors that should motivate people to invest in the financial market. The very premise of their research to imagine investing in financial instruments based on sustainability indices along their earning potential is radical and new. It is both complex and challenging to imagine new structures that are generally forbidden from non-experts from the domain of finance. I hope this report will highlight the relevance of considering environmental and sustainability identifiers during investing and possibly encourage young designers and practitioners to push the domain's status quo.

It is perhaps the most ambivalent I have felt in a while. It has undoubtedly been uneasy for me to pressurize students to submit this report during a pandemic, arguably one of the most challenging situations we have all seen in our lifetime. I am not happy doing so but limited by how each of our actions and responses is delegated through multiple levels and how it affects the students. However, it will be naive to acknowledge only in support of either facilitators, institutions, or students. My stance has been to only balance as much as I can and which affects students in least possibly manner.

It would also be unfair to authors and their efforts to call me as their project mentor for the work that is primarily initiated and managed by themselves. Maybe an occasional project mentor would be just. Finally, I congratulate them on all the hard work, all the nights, all the challenges they have gone through, as they could finally finish their thesis. But more importantly, I believe this will help them in becoming designers who will advocate for ethics and moral responsibilities over anything else.

Deciphering algorithms — Discovering Algorithmic Literacy through the lens of Explainable AI; by Daksha Dixit and Simran Singh

What if users could interact with elements they see on the user interface, inspect the digital systems they use, and understand the algorithm-driven processes behind them?

With back-end algorithmic processes becoming increasingly obscured and vastly complex, users are often unaware of the rationale employed behind the services they use. There is a call for a new kind of literacy. This project proposes 'Algorithmic literacy', which brings to the forefront information about algorithmic processes that automate and augment tasks users engage with on digital services. What one's agency is in algorithmic decision making and what actions they can control on a digital service are also parts of the proposed algorithmic literacy.

To achieve this vision, the project identifies 'Explainable AI' as a methodology which inspects and explains the steps and models involved at the back-end of artificial intelligence technology and makes it apparent and understandable for users.

While looking to engage users with this information, it is important to locate relevant factors that are understandable, and meaningful for them to decide the extent of their relationship with these services. When using Explainable AI methodologies as a medium to make obscured algorithmic processes apparent, this project finds it imperative to look at users as new learners, informing their decisions by ways of explainability that are borrowed from pedagogy and instructional design.

The project proposes to relook at users as learners, and their interactions with digital services as a 'site for learning'. We propose Algo-Lit plugin which helps users understand their preferences, available user controls on a digital service, understand what different User Interface elements mean and how they contribute in the decision making of the algorithmic processes and the entire system of decision making. The project is a speculation of explainability of AI features for end-users.

Greenbox algorithms — Financial indicators that motivate impact investment behavior; by Apoorva Avadhana, Nupur Patny and Shreya Mishra

Designing novel ways to motivate responsible investing behavior by creating an experimental stock indicator

As the global community aims to deliver on the United Nations' Sustainable Development Goals for 2030, there is increasing recognition that governments alone cannot facilitate the necessary change and funding. Citizens, governments, and investors are increasingly looking to companies to take responsibility and address societal challenges. The first step towards this is for companies to take accountability for their own sustainable practices and make that information transparently available for investors and consumers. The medium for this is ESG reporting (Environmental, Social, and Governance) which has only recently begun to have financial relevance for companies. Yet, ESG performance is still under-reported, and the onus for complete disclosure lies solely with the companies themselves. Impact investors use this report to analyze a company's sustainability performance and accountability and make responsible investments. The value that these investors possess is called 'Blended Value', where they consider environmental, social, and economic performance when making an investment.

We have created a platform to nudge investors towards impact investing by creating novel data representations and interactions with companies' sustainability and financial data by incorporating facets of 'ESG' and 'Sign Value' using natural language processing and fuzzy algorithms. We present to you GreenBox: An ESG Stock Indicator. GreenBox is a plugin that can be updated or linked to one' s portfolio. Instead of huge risks and portfolio transformations to start investing sustainably, the platform helps in setting comfortable value targets for a portion or percentage of investments as impact investments.

This project was also awarded first prize in the category of Social Impact and Bottom-up design by IAMAI Design Leadership Awards 2020, first place at the ADI Awards 2021 under the category of Design for Social Impact at Pune Design Festival 2021, and also created an opportunity for them to discuss their work with members at Zerodha, a leading stockbroker in India.

Epilogue by Vineeta Rath

The brief, to 'deconstruct algorithms' is not an easy one - because these 'intangible instructions' remain 'hidden' from average humans, who go about their daily lives in a more-or-less 'tangible' world. Irrespective of where they are deployed, functioning in the background, these algorithms remain unquestioningly sincere to the human minds that built them. And, more often than not, may also have embedded within them, innocently, the intelligences, interests and biases of their creators. Not to speak of the unexpected powers algorithms can acquire over time, as they evolve through learning the environments they are deployed in, both human and non-human.

This leads to the following few pertinent questions among many - Can we humans understand algorithms enough to be able to work with them as they evolve - and rather than beasts we cannot relate to after a point in time, can they continue to remain as powerful tools, as they would have originally been envisioned to be - and continue to operate as allies, towards more equitable and plural futures? Can we imagine more inclusive futures where the understanding of algorithms is as ubiquitous and 'normal' as their own proliferation and omnipresence? And therefore, curiously enough, can we truly 'deconstruct' these potent human creations called 'algorithms' - to then enable their 'more conscious constructions'?

No work of significance is accomplished overnight, and this is most certainly true for visionary endeavours such as this - where the need is to challenge, and eventually push the status quo. And, although there are more possibilities to explore, more worlds to address, more work left, always! - the two projects, Greenbox and Deciphering Algorithms are pertinent steps in the right direction - wherein the visions have been not to produce mere thesis projects, but pieces of work that act as symbols of possibility of what more humane and inclusive futures can be.

From breaking down words to understand the constructs of a language, to exploring Derrida's theories, to coming up with 'Oribaka', an experimental language to communicate origami instructions, what I have thoroughly enjoyed as an observer throughout as the project flowed its natural course has been the richness of the explorations. And, although these five bright, enthused minds along with their visionary mentor, Gaurav, might have envisaged the journey of this project differently - given the sudden challenges posed by the pandemic (which requires no elaborate mention), it is most certain that what they have finally arrived at together may have not been in their remotest collective imaginations.

Tenacious and determined as each one of them is, in her own way, what is really promising about Apoorva, Daksha, Nupur, Shreya and Simran is their genuine passion to not just prepare for the future, but actively build it, through their conscious, informed efforts. And their dreams and actions for a better world only give me more hope in the future - of these five spirited individuals of course! - but also of the world that tomorrow will be!

Shine your light on the world girls, your time has come!

(Vineeta Rath, a faculty member at Srishti-Manipal Institute, has been a mentor to the participants of this project throughout their bachelor's program at Srishti. She has had the unique opportunity to witness the growth of these students firsthand.)

Acknowledgement

Dr. Venkat Chilukuri: for his review and suggestion on project brief; and helping in facilitating the signup process; Vineeta Rath: for recommending students to join this project; and supporting this project to happen; for help in initial setup of the project space; for being constant support to students; Rustam Vania: for allowing and supporting this project to happen; Roćio Fatas: for her review and suggestion on project brief; Anand Vijayan: for his suggestions on finding people in Srishti who may help us with the notionn of Deconstruction; Hilda Cynthia: for being welcoming and helping in find a space to facilitate project in; Manjari Singh: for being welcoming and allowing to change the space as per the needs of the project; allocating more space near the screening room for putting up work; and guiding both Gaurav and students on every possible need; Dr.Naveen Bagalkot: for considering my request to offer this project; Karan Dudeja: for his review and suggestion on project brief; for help in initial setup of the project space; for giving feedback at Oribaka exhibition; Dr. Srijan Sandip Mandal: for his belief and support for the project; and for his review and suggestion on project brief; Kshama Nagaraja: for her review and suggestion on project brief; Dr. Geetha Narayanan: for listening to the idea of the project and her belief in the project; Team Rafiki - Arsh, Ananya, Sohaya, Janaki, Rashi: for participating in the pilot study of Oribaka; Zenab Bastawala: for visiting Oribaka exhibition with her students and encouragement; Pratyush Raman: for agreeing to swap spaces for the sake of project Whatsapp: for being the platform where students finally respond to my questions unlike email; Sajoo Bhaskaran: for listening our questions and giving us time to show us VITM; Arvind Venkatadri: for reviewing the project and his feedback; Tapasya Thapa: for reviewing the project and her feedback; Sanjukta Ghosh: for reviewing the project and guiding authors on topics related to data analysis; Paul Anthony: for reviewing the project and his feedback; Several others: Agrim Agarwal, Rohit Patny, Balchand Singhi, Debashish Banerjee, Shriniwas Mishra, Vijay Kumar, Niteen Pradhan, Ravi Mani, Ranjani Balasubramanian

Simran has a great sense of organisation. You will find her desk all neat and organised, composed with all her notes and sketches. The picture on the right is a tiny glimpse of her desk during the project. Daksha, on the other hand, is fearless when she is reading and thinking, she does not hesitate to use large brown sheets to map her notes. She reads, reads and reads until she really understands the topic. Her notes mirror her thinking process. Her only weakness might be skipping breakfast but who can beat her otherwise?

(Picture on left) Green team is forced to put down their plan for the next few weeks. I hoped to contribute in their thinking process, but they did an amazing job on their own. I failed to understand how they did it all but credit goes to them for the putting together the mountain of hard work.

Project room - 407, Srishti's N5 campus in Bangalore. The pandemic is here. Students have left to their home. "I want to fly" says the sticky notes on the wall next to Apoorva's desk. Apoorva may not listen every time but pulls through by the end of the day. But, Nupur is a great listener and very organised too. Shreya might be sitting under the table reading books, and what if she is running a bit late, her work reflects strong understanding of how things work. I wish I had a picture of them working together.

Evening before the exhibition. The creators are shooting a tutorial video.

Oribaka — A language to make Origami for the visually impaired. One of the activities pursued by all of the participants to build a first hand of understanding of "algorithmic thinking" - a systematic way of breaking down a task into smaller steps and then solving each of those steps in a logical manner. The activity was to create a language to make origami for the visually impaired. The language was created by the team and then showcased to several people to understand the effectiveness of the language.

Brainstorming while building the language

Trying out the language with Rafiki folks