New work developed by the lab looking at how matrix methods can be applied to identifying properties and clusters within office networks.
This approach takes both social data (emails, directed messages) and spatial data (office floor plans) and develop a rationale for placing users into space to support their social networks and potentially change their behavior beneficially.
A mini-site for this work can be found here: link
Meta Lab's Sam Joyce is pleased to be involved in Smart Geometry 7-12th May 2018 in Toronto.
The cluster is entitled 'behavioural enviro[NN]ments' and undertaken in collaboration with Kate Jeffery a Professor of Behavioural Neuroscience and Wellcome Trust Investigator at UCL and Jonathan Irawan a Computational Designer at Hassell Studio.
A synopsis of the ctuster aims are below:
Autonomous, hyper adaptive and responsive buildings and environments have been a topic of obsession for many designers and architects since the 1960s. Nicholas Negroponte envisioned a system that enables an architect to create an intelligence to design the building themselves. Within the Architecture Machine Group at MIT, they created The ‘Seek’, a computer-controlled landscape environment full of small blocks which were inhabited by gerbils. It was the task of the machine to analyse the variation between reality and virtual models with the responsibility of keeping the original plan intact if possible, which was then used to predict the future behaviour of the gerbils. Unfortunately however, the system was not successful, due to an unclear criteria and high number of permutations in variables of disruptions and behaviour. The cluster will revisit the project, incorporating modern advancements in Machine learning algorithms and techniques to transform our own version of the dynamic environment itself into a neural network. The environment, in an attempt to predict the occupancy behaviour and usage pattern preferences of gerbils, will utilise reinforcement functions based on movement(navigational) patterns and spatial usage. The final aim is to create a landscape that would best suit the daily behaviour or activities of the small animal subjects.
A link to the cluster can be found here behavioural enviro[NN]ments
Due to the award of a new SUTD-MIT joint research grant, we are now looking for a key post-doc or experienced researcher to help us.
The project aims to research, develop and test data-visualisation for design, but taking a 'decision-first' approach looking at how data can be used to support good design decisions rather than simply make good visualisations.
A job spec and contact information can be found here: link
Work presented at ACADIA 2017 in MIT, Massachusetts.
The research investigates the effect of applying genetic programming applied to parametric models.
We look into measuring the geometric and parametric complexity of models which are derived through evolutionary optimisation.
Parametric models are defined as genomes using Cartesian Genetic Programming a method of developing logical structures which are consistent with parametric definitions.
These enable machine learning processes to manipulate and define autonomously parametric models.
This work has a complimentary website, where you are invited to explore the process. Using basic parametric components and automatically generated designs, which are optimised both based on hard objectives and also user interaction.
The website can be explored here: link
The paper can be found here: link
New work implemented by Verina and presented at IASS 2017 in Hamburg.
Concept based on experimenting with using servers as project repositories for design visioning.
Taking inspiration from Git type code systems but allowing them to work for design modelling.
Current work is integrated with the Grasshopper parametric associative software.
More information can be found here: here
The lab has been investigating the application of big-data to understand how people experience cities.
Visually exploring a dataset of 12 million Tweets geo-tagged in Kyoto and sent between 2011 to 2015.
Using this data to obtain insights into areas of activity.
This work is ongoing but Some initial work can be seen here
We have some initial work by Nazim on exploring on the web how we can see the development of models which have been evolved.
Here we look at 'Meta-parametric' models a higher level computational representation of parametric model which could open the doors to A.I. based creativity in architectural design.
The aim is to understand the effects of application of Cartesian Genetic Programming (CGP) [Miller, Julian F. 2011] as a mechanism to construct and control parametric models generatively using an evolvable digital genome.
Initial results are exciting as the point to effects on the performance and complexity of the parametric model and resultant output geometry.
This tool allows one to see the effects of a chosen number of iterations of evolution on a chosen population size, using a selectable 'palette' of components under different mating and selection criteria.
The research is ongoing and being developed for publication.
The site can be explored here
Verina will be presenting a big-data paper "Connecting the dots: using Twitter for design insights into social behaviour around metro transport nodes in Singapore
and beyond" at University of Southern Australia, Adelaide.
This is research in collaboration with social geographer Ate Poorthuis looking at how Twitter can be used to understand what activities happen around metro stations and comparing 5 Asian cities; Singapore - Hongkong - Kuala Lumpur - Bangkok - Osaka.
More info here
We present here some early work looking at how we can use Twitter to measure happiness by using specialised twitter data.
This work uses Emoji (picture characters originally developed in 1999 for mobile phones) as a language unspecific way of capturing the feelings of locals and visitors in relation to specific places. The data is shown using d3 visualisation allowing users to explore the data and find paces of activity, sentiment and emotion in the current Jurong East.
The site can be explored here
Sam will be presenting a paper entitled "Ecological and Economic Concepts for a Mixed Initiative Networked Design Infrastructure" at Georgia Tech, Atlanta.
This work focuses on the required collaboration infrastructure required to enable a parity between human and machine participants in the processes of creative design, design analysis and optimisation and the and design review. This paper looks at how financial markets as well as ecological and evolutionary environments can provide effective feedback on individual actors to enable a balanced design system.
The paper can be found here
Sam will be presenting a paper entitled "Cloud Based Design Option Control Systems a Discussion and Implementation".
It considers the definition of needs for networked systems to support creative engineering and architectural design.
It focuses on the plurality of the design process, how design teams use multiple alternate and hierarchical model representations of a design at any one time.
The paper compares this process with technology current used by software developers to version code, looking at how these concepts can be applied using cloud computing to AEC design contexts.
The paper can be found here
Verina has been successfully awarded the prestigious Singapore University of Technology and Design President’s Graduate Fellowship to pursue a PhD in SUTD under Professor Sam Joyce. Her work-in-progress aim is to explore applications of computation to enable wider public interaction and engagement with the architectural design process. Specifically focusing on methods which enable grass-roots design initiatives in fast developing cities to leap-fog existing planning and public consultation processes.
Verina has joined us from FCL. She comes from a computer science background, especially the application of computer graphics and game engines to enable wider designer and audience understanding of complex issues. She will be working on the application of web-services in early stage conceptual design, under the 'Automated Design Space Exploration' IDC project.
Nazim Has joined us from NUS. His previous work looked and generative design and Houdini for working on energy models and building optimisation. He will now be working on the 'Automated Design Space Exploration' project. Focusing on approaches to enable more flexibility in Parametric modelling tools by investigating their use on both the server and client side web. As well as investigating machine learning approaches to generating parametric models, using techniques such as Genetic Cartesian Programming.
Sam Joyce will give a presentation at the Future Cities Laboratory (FCL) on the NUS Create campus as part of their Lunch Talks series. The talk is titled "Interactive Design Computers and Users" and will cover how machine learning, the web and data visualisation is combining to develop new design environments towards enabling multiple stakeholders and designers to collaborate in new and exciting ways.
Andre has joined us from an urban studies and big data project at SUTD, after extensive experience in design, research, teaching and consulting in Europe and the Americas. At META Design Lab he is focused on developing new thinking around existing computational methods, looking at their limits and how new approaches can extend or exceed these to improve design as a result. His main areas of inquiry here include: generative design, machine learning and "mixed-initiative" systems, encompassing topics such as parametric-associative modelling, genetic programming, multi-criterion optimisation, design space exploration, problem-solution co-evolution and preference learning.
Our group has developed this site as a window into our current activities. We hope you like it and welcome any feedback or communication.
Today marked the official start of our pilot project "Automated Design Space Exploration" within the Sustainable Built Environment grand challenge at SUTD-MIT/IDC .
There is increasing interest in understanding the use of urban space around heavily trafficked areas like transport interchanges; both in terms of their cultural social commercial use (Cha, 2001) but also in the context of ‘last mile’ journeys. However, this type of use is often not captured in traditional census, survey or transportation data. This research focuses on the use of social media data to develop insight into how spaces around public transportation stations are used in a comparative study of metro stations in several Asian cities. It focuses on finding similarities and differences between such spaces, including how different master planning and key urban amenities affect use. Specifically, this research aids in devising strategies to promote wider use and interconnectedness of area around stations and to understand the impact of large dense agglomerated building typologies, which are becoming a feature especially in Asian cities (Frampton et al., 2012).
Full article here
The paper discusses a newly developed cloud based design-model version control system for supporting manual and semi-automated generative design via a web-based platform which enables exploration and retrieval of designs and the solicitation of users' evaluations of them to enable interactive genetic algorithms to produce preference optimised design options. The current needs of a commercial building engineering-design context is considered and relevant state-of-the-art computational design software is appraised, the review also covers software development based version-control-systems, identifying unmet needs in the former which could be solved by the latter which is the motivation for this work. A description of the new system's rational and current developmental architecture and technical components derived in response to these needs follows. Principal features include an online repository of design options, a parity between of users (human and machine) creating and acting upon the design options which align with existing tasks and processes undertaken by a design team. An example structural frame design problem is explored, discussing how the system functions to support performance based creative design via exploration, including visualisation and data dashboard of a design session. We discuss further work that could be undertaken to improve and widen the impact of this repository and collaborative human machine approach. We conclude that such systems addresses pressing software infrastructure requirements for the AEC industry especially for expanding the range of design options considered and managed particularly for early-stage design activities.
Full article here
This paper describes the conceptual underpinnings of a infrastructure currently in development which aims to provide a computational environment for productive and creative human and computer co-creation of architectural design artefacts in an open, networked collaborative system. It identifies some key collaborative activities in design and looks to combine these within a system geared towards the productive utilisation of various agents involved towards developing a range of good design options. It takes as a conceptual foundation the application and hybridisation of evolutionary, ecological and economic principles. This approach is outlined with special emphasis on aligning with, and maximising the benefit of, an internet/network based infrastructure, connecting the design agents, supporting human/machine parity. The contribution for this paper is to assess the creative design context of the AEC industry, especially at early conceptual and ideation stages, and compare this against existing tools and paradigms used during creative design, to propose a new technological intervention to support these activities.
The paper can be found here
In response to the call for papers of the special edition of the International Journal of Architectural Computing entitled Design with Freedom we have written an article considering the flexibilities and limitations afforded by 'manual' design and prevalent computational design paradigms.
Generative processes and generative design approaches are topics of continuing interest and debate within the realms of architectural design and related fields. While they are often held up as giving designers the opportunity (the freedom) to explore far greater numbers of options/alternatives than would otherwise be possible, questions also arise regarding the limitations of such approaches on the design spaces explored, in comparison with more conventional, human-centric design processes. This article addresses the controversy with a specific focus on parametric-associative modelling and genetic programming methods of generative design. These represent two established contenders within the pool of procedural design approaches gaining increasingly wide acceptance in architectural computational research, education and practice. The two methods are compared and contrasted to highlight important differences in freedoms and limitations they afford, with respect to each other and to ‘manual’ design. We conclude that these methods may be combined with an appropriate balance of automation and human intervention to obtain ‘optimal’ design freedom, and we suggest steps towards finding that balance.
Research surveying the history of data visualisation from a design and engineering point of view. Followed by investigation of the application of modern web based approaches as used in design decision making for a number of commercial case study projects.Full article here
Research into the use of parametric modelling in the creative and optimisation process. Limits are identified with respect to topological differentiation on the parametric models and thus the variation of the resultant models. Novel methods of increasing the topological variability are discussed in the cortex of early stage design along with it's implication.Full article here
Paper co authored with Robert Aish of Autodesk Research, Andrew Marsh of Ecotect and Al Fisher of Buro Happold. Paper documents the thinking and implementation of design evaluation and optimisation with a parametric associative programming language based design interface 'Design Script', which is now integrated into Autodesk Revit as the engine behind Dynamo.Full article here
Paper for the Institution of Civil Engineers special edition journal investigating the use of optimisation to assist design and engineering decision making. Evolutionary algorithms used to find performance trade-offs by automating parametric models and linked structural and environmental analysis.Full article here
Back to top
© META Design Lab
Singapore University of Technology and Design, 8 Somapah Rd, Singapore, 487372
T: +65 6499 7454 E: sam_joyce [at] sutd.edu.sg