Intelligent Design and Machine Learning
I have been harvesting seeds of my thoughts about Machine Learning and Intelligent Design. Noting them in my diary since days. On one usual day, when I hit the TV remote Netflix button, the recommendation engine (AI) showed ‘Human: The World Within’ with 92% match. While watching the first episode named ‘React’, my brain started to flourish the thoughts of Machine Learning and Intelligent Design. My visionary neurons taken the command, started to connect synapses and visualising how we humans, our ecosystems and infinite natural processes on earth are the well connected intelligent designs of the super intelligent design system of this universe, where trillions of various AI system (ML NLP Vision Recognition…) cell engines are only running throughout our own body itself.
In this article, I am sharing my high-level design views and visions about Machine Learning (ML) and how this one AI technology process is a superb addition in the design world. Also, how designers can think around this technology to design intelligent systems (iSD), products (iPD) and experiences (iUX).
Since my right brain neurons are more passionate and a bigger portion of my thinking goes in visuals, so for efficient communication it would be intuitive to draw out some of my views and visions to simple visualisations. In this article I have created a couple of such visualisations about the ML function complexities, its AI relationships and its usage in the intelligent design experience.
I recommend while you go through this article pay close attention to visualisation images as well, as I communicated some additional thoughts in them apart from what I have expressed in texts.
Machine Learning in Simple Terms
Machine learnings are packs of algorithm models. These models are designed to ‘Collect, Learn and React’.
Collect = Inputs
Learn = Analyse Inputs, Develop logics and rules
React = Output/s based on developed logics and rules
ML system, AI family tree and Intelligent Design
Based on the current state of evolving ML technology, I have visualised its functions, relationships and SOTA in the AI family tree to simplify its understanding towards the design world.
Machine Learning (ML) is a one type of learning model in the AI family.
Reinforcement Learning (RL) is a one type of process in the ML family.
Deep Learning (DL) and Deep RL are the types of advanced processes in the ML family.
Neural Network is a one type of intelligent process in the DL and Deep RL family.
These AI and ML family algorithm models, their processes and sub-processes are advancing. Many more such process models are under research by AI researchers and scientists around the globe. Once these innovative tech models are shaped in a meaningful problem-solving technology, they are named based on function, patented on approach/methods and introduced to the world.
Neural Network Simple Terms
Inspired by how brain neural network functions, ML neural network allows acquired input information to pass through multiple signal processing layers (like neuron) and shares thousands of input data signals between layers (like synapses) as a one analysis function of Machine Learning for required output.
Similar article : AI impact on UX Design
Deep Learning in Simple Terms
More like our brain processes the input signals received by our sense organs to identify, learn, develop muscle memory and have an intelligent output based on recognition, experience, memories and learnings — The deep learning process is designed to collect thousand inputs, analyse them, develop logics, learn, get trained and provide an intelligent output.
If I define smart design based on my first sense of experience then —
“The Smart Design provides a fixed set of intuitive response, assistance, supervision or guided process etc. to user, subject and system through visible elements of experience.”
Which is happening by designing WHAT to REACT by the trainable products, systems and processes.
But if I go a level deep in my thoughts and define intelligent design with extended sense of experience then —
“The Intelligent Design is not only about the intuitive response, assistance, supervision or guided processes but much beyond, which is to provide the ‘MOST NEEDED REACTIONS’ in varying situations and develop trust with users on subjects in systems through visible and invisible elements of experiences.”
Which would be possible by syncing AI models with product design, designing WHAT, HOW and WHEN to REACT by the TRAINABLE, TEACHABLE, ADAPTABLE products, systems and processes.
In a few years we will reach a remarkable level of intelligent design through many advancing AI models with sync of product designs and UX designs. Machine Learning is one such imperative advancing AI branch towards intelligent design journey.
ML for Intelligent Design
ML experience has been increasing since a couple of years in our daily digital interactions like Netflix watching experience, Amazon shopping experience, Google’s email, map and search experiences and many such in ML enabled products.
All of these ML enabled recommendation engines are a baby step towards Intelligent Design (The Most Needed Reaction) and experience (IDX).
Current State of ML
Do we need Netflix ‘More like this’ feature shown as a fixed part of design?
Does the Amazon home page need to show us ‘Inspired by your shopping’ or ‘Top picks for you’ as a fixed part of landing page design?
Or let’s say you shopped on Amazon to gift an item of a guest choice, who is visiting you at your home and the recommendation engine (ML) learned it as part of ‘your need or choice’, which you never use or like and shows similar products at home page under various categories.
Are these Intelligent Designs? Are these ‘Most Needed Reactions’ in varying situations? These are the sections of smart designs. These are going through a fixed set of recommendation output design, dev and business strategy under ML rules and remain surfaced as a fixed set of UI designs to users.
How to align ML systems towards Intelligent Design?
ML development sync with product design and UX is the one key start towards Intelligent Design. For example, in the design process I also consider ML as a smart research usability tool and sync it with UX research and product design strategies to identify the ‘most needed reaction’ in design.
To keep design towards the path of 'the most needed reaction', I try to,
Design the product, architecture and UX around ML and other AI developments.
Define and create design and development channels to boost ML and other AI indexes.
Intermittently, enhance the design development strategies and DDT models (deep design thinking).
Apart from training these models to LEARN, tune them with smart input-output analysis channels of user interactions and train them to learn WHAT, HOW and WHEN to REACT.
Please note, all the above should not be looked only through the lens of digital applications UI (mobile, web) but it also applies well on their systems, processes, all AI ML enabled products, robotics, next generation holographic projection interfaces and beyond!