Geoffrey Broadbent'in 'Design in Architecture: architecture and the human sciences' adlı epiy eski bir kitabı var. [London : John Wiley & Sons, 1973] giriş kısmındaki (sayfa xi) şu iki paragrafa bakınız:
More and more over the past two hundred years, the various functions of building design have been separated out. As the theory of structures developed, so the engineer became a separate practitioner whose work, on the whole, could be quantified. He believed himself, for this reason, to be doing a tougher job than the architect, ans so, as they too emerged, did the quantity surveyor, the heating, ventilating and electrical engineers. the architect's task gradually shrank; he was left with the 'soft-edged' aspects of building design, those things which could not be quantified but relied instead on his skill and judgment as an aesthete and a gentleman. He was concerned, above all, with matters of form, proportion, colour and texture -the raw material, as it were, of visual delight.
Increasingly, however, it seemed probable that these too could be quantified. colour, for instance, could be specified in terms of hue, value and chroma; experiments could be set up in which psychological response to colour could be quantified, and thus a further aspect of the architect's remaining task would be eroded. Form, proportion, texture and so on could follow so that the architect might be left, literally, with nothing to do.
doktoramda yola çıkış hedefim tam olarak buydu. renk ve biçimlerle ilgili tarzlar tariflenebilir ve bu ilkel bir yapay tasarlama sisteminin üretilmesi için yeterli bir çıkış noktası olabilir diye düşünmüştüm. hala öyle düşünüyorum. ama yapay tasarımın önündeki engeller bu noktada değil aslında. broadbent takip eden sayfalarda tasarımcının bir kültürel iklim içinde hareket ettiğini ifade ediyor ve iddia ediyor: [s xiii]
It is clear, therefore, that the design of architecture finally can never be a matter of completely automated decision. ... [proponents of this view] offer powerful tools for decision -making --when one has data in quantifiable form- ...
hm. kültürel iklimin bir yapay zeka sisteminin becerilerinin sonsuza kadar dışında kalacağını düşünüyor gibi görünüyor. aksi yöndeki görüşe-arayışa "strong AI hypothesis" diyeler. insan tekinin yürüttüğü herhangi bir entelektüel görevi yürütebilecek zeki-ajanın üretilmesi. gündelik bilginin günümüz yapay zeka yaklaşımları için kullanılabilir kılınmasını hedefleyen projeler yürütülmekte bugün: cyc, opencyc, dbpedia ve freebase projeleri bir gugıllanabilir mesela. günümüzün hakim yapay zeka paradigması da gittikçe katı ve basit* mantıksal hesaplama algoritmalarından daha olasılıkçı bir yöne doğru kayıyor. yol uzun olmakla beraber zekanın çeşitli veçheleriyle ilgili çalışmalar sürüyor, yani doğal dil işleme-kullanma, robot görüşü ve hareket, öğrenen sistemler...
bu konudaki kanımı daha önce yazmıştım. tam anlamıyla yapay tasarım da güçlü-AI programı gerçekleştiğinde mümkün olacak. ama bu tasarlamanın bazı zahmetli alanlarının daha önce otomasyonun konusu kılınamayacağı anlamına gelmiyor. bazı biçimsel hususlarda tarzı tarifleyip yapay tasarım sürecine eklemlediğimizde büyük bir iş yapmış olacağımızı, tasarım alanlarını sadece işleve ve performansa indirgeme üzerinden otomatize etme düşüncesinin konunun önemli özelliklerini gözden kaçırdığını... (nerdeyse aynı satırları schumacher'in bir metninde de okudum temin.)
eğer bu inanç 1973'te kalmış ve ortadan kalkmış olsaydı tabii bu eski ders kitabına bakıp böyle bir entry girmeyi düşünmezdim. lawson ve dorst'un 2009'da çıkan kitabında da aynı inanç benzer bir kestirmelikte ortaya konmuş. s16:
The first significant international conference to explore the nature of design expertise took place in Sydney in 2003 (Cross and Edmonds, 2003). this built on attempts to understand the whole idea of expertise that in turn had been driven by attempts to develop artificial expert systems and the consequent need to capture human expertise in symbolically coded digital environments. Design however remains a human activity beyond the capability of artificial intelligence [1] and therefore poses some interesting challenges to the computational theory of mind that lies behind such work. [2] While some argued that it is merely a matter of time before computers will be able to design, others including the authors of this book argue that there is something essentially human about this highly creative activity. [3]
[1] nolu iddianın doğru olduğu açık. ve [2] nolu cümle bu alanda çalışmak için en önemli motivasyon kaynağını ifade ediyor. ama [3] biraz kestirme bir açıklama gibi. 'özsel olarak insan olan', dolayısıyla başka hiçbir sistem tarafından yürütülemeyecek olan... ruh falan? allah'ın üflediği şeyler? belki de uzman sistemleri yapay zekanın geçmiş gelecek tüm alanı ile özdeşleştiriyorlar. tabii böyle iddialar yapay zeka alanının sembolik sistemlere kitlenmişliği karşısında meşru kalıyorlar. [meşhur simon da sembolik yaklaşımın savunucularından]. sembolik sistemler atomik öğeler üzerinden yapılan matematiksel-mantıksal işlemler ile kararları ve kavramları hesaplıyorlar. gerçeklikte bu atomik öğelere tam olarak karşılık gelebilecek bir 'şey' yok, ne de o atomik öğelerin oluşturduğu önermelere karşılık gelecek bir hakikat türü. ama mesela sub-sembolik (nöral ağlar, evrimsel algoritmalar...) ya da hibrit yaklaşımları (mesela act-R), ya da bayesçi karar-verme sistemlerindeki gelişmeleri gözardı etmemek gerek. çok mu ilerlenmiş bu hususlarda? yok. daha başlangıç aşamasında belki. bir gün bu boğazı sıkılmış von neumann mimarisi de, formel mantığa dayanan programlama ve yapay zeka yaklaşımları da aşılacaktır diye düşünüyorum. ve aşılmalıdır da, çünkü aşağıdaki paragrafı bir eleştiri olarak, biraz yüzeysel olmakla beraber, anlaşılır buluyorum:
s104 (lawson ve dorst, 2009):
Designers, however, clearly use another form of cognition which is generally described as 'visual thinking'. They think directly by manipulating graphical information which is extremely difficult to encode entirely in conventional symbolic systems as required by AI. This is yet another reason why it seems unlikely that computers can ever be enabled to design in the sense that humans do. ['bugünkü bilgisayarlar' diye de okunabilir]
*[katı" zira mantıksal önermelere dayanan bir bilgi türü varsayıyor ama dünya ile ilgili olarak bu tip bir bilgiye sahip değiliz ve pratikte her zaman eksik bilgi üzerinden hareket etmek zorundayız, basitliği de bu algoritmaların karmaşık ve belirsiz bilgi alanlarında hareket etmeyi sağlayamamasından kaynaklanıyor, oysa insan faaliyetlerinin büyük kısmı belirsiz bilgi alanlarında geçiyor. eğer bir alanda bir seri kural ile ifade edilebilecek türden bir bilgi birikimi varsa diyelim ki bir uzman sistem üretebiliyorsunuz, ama ya kitapta olmayan bir sorun karşınıza çıkarsa? uzman sisteminizden mimarlık yapmasını beklemeyin. 'uzman sistem ne?' derseniz, yapay zeka çalışmalarının ticari açıdan başarılı olan ilk ürünüymüş, 80'lerin başında çok popüler bir konuymuş, bugün pek çok alanda popüler sayılmaz.]
yapay zeka etiketine sahip kayıtlar gösteriliyor. Tüm kayıtları göster
yapay zeka etiketine sahip kayıtlar gösteriliyor. Tüm kayıtları göster
19 Nisan 2011 Salı
28 Aralık 2009 Pazartesi
why automation | niyeotomasyonyaniniyeyani
the question continues to linger (thus asks my temporary supervisor) :
but, why automation?
i usually refrain from giving a direct answer. i give some political explanations reminiscent of the situationists which hold some portion of truth (yea automation might be good, my argument through dishwasher). and point towards growing academic and economic interests (more and more intelligent-design-software will flood the market).
but indeed, though vaguely, i can discern why; why this fascination with artificial intelligence, computation, perception, cognitive science, consciousness; why this eagerness to imitate a portion of human behavior which is allegedly a peculiarity of the human being (i.e design activity).
it is just an assault towards the last fortress defending against complete disenchantment. nothing is special about human mind. indeed there is nothing special about anything. this is the researcher as positivist and nihilist. (two in one? don't go too far researcher!)
but, why automation?
i usually refrain from giving a direct answer. i give some political explanations reminiscent of the situationists which hold some portion of truth (yea automation might be good, my argument through dishwasher). and point towards growing academic and economic interests (more and more intelligent-design-software will flood the market).
but indeed, though vaguely, i can discern why; why this fascination with artificial intelligence, computation, perception, cognitive science, consciousness; why this eagerness to imitate a portion of human behavior which is allegedly a peculiarity of the human being (i.e design activity).
it is just an assault towards the last fortress defending against complete disenchantment. nothing is special about human mind. indeed there is nothing special about anything. this is the researcher as positivist and nihilist. (two in one? don't go too far researcher!)
neyle ilgili?
artificial intelligence,
design automation,
tasarım otomasyonu,
yapay zeka
11 Aralık 2009 Cuma
artificial design and strong AI | yapay tasarım ve yapay zeka için zor problemler
mary lou maher writes in a kind of short introductory text somewhere about AI in the large and AI in the small. she seems to say, it all started with huge ideas in an all-encompassing manner (which in a way, mostly failed), but now it turned into a series of partial and modest patches here and there... if there is an AI in the small, there's an automation-in-the-small as well. and design games is just one of the modest, transient and sometimes not-so-meaningful patches. and i have a new phrase as a substitution to 'design automation', it's 'artificial design'.
[0404010:] i also want to add a note about this phrase: "AI complete"
[wikipedia:]
"In the field of artificial intelligence, the most difficult problems are informally known as AI-complete or AI-hard, implying that the difficulty of these computational problems is equivalent to solving the central artificial intelligence problem—making computers as intelligent as people, or strong AI."
i have been thinking about "task-decomposition" since i've started this project four years ago. design games model was an abstract thinking about sub-tasks of architecture [and "the star-map of architecture" which we had produced as "sanki-dikdörtgen" was strongly related]
now, some sub-tasks in a particular architectural design process are rather mechanical (some even trivial) and can be solved through some calculation (if necessary use well known AI techniques and algorithms), and there are some others, which include a huge set of parameters, objectives, and constraints, that renders them practically impossible to solve; i mean, if you approach them through Good Old Fashioned Artificial Intelligence paradigm (i.e. GOFAI, compare with "sub-symbolic" approaches, which includes evolutionary computation).
these latter types of subtasks are still being researched (as a few examples: distribution of the masses, building orientation, floor plan generation, envelope shape and aperture formation, or complete conceptual formal design in an integrated fashion). but this line of research is still rather underdeveloped, from a designer's point of view. nevertheless there's great hope, 'cause existing techniques seem to be offering several possibilities.
however there is also a third type of sub-tasks, which have to be tackled for a convincing design solution to appear. these include the conceptual (or say, mostly verbal) brief generation and problem setting stages, where the designers need extensive understanding of every aspect of culture as well as a complete ability to understand and manipulate shapes and objects (at least as well as human beings do). a creative and sensitive design proposal which is in dialogue with it's cultural, political, economic, and historical context, doesn't seem to be attainable with mere adaptation or pre-defined rules or constraints or as a result of a mostly random evolution.
are we obliged to wait for a machine, at least as clever as human mind (i.e. for strong-AI program to be accomplished) before entrusting design activity altogether to the machines?
[0404010:] i also want to add a note about this phrase: "AI complete"
[wikipedia:]
"In the field of artificial intelligence, the most difficult problems are informally known as AI-complete or AI-hard, implying that the difficulty of these computational problems is equivalent to solving the central artificial intelligence problem—making computers as intelligent as people, or strong AI."
i have been thinking about "task-decomposition" since i've started this project four years ago. design games model was an abstract thinking about sub-tasks of architecture [and "the star-map of architecture" which we had produced as "sanki-dikdörtgen" was strongly related]
now, some sub-tasks in a particular architectural design process are rather mechanical (some even trivial) and can be solved through some calculation (if necessary use well known AI techniques and algorithms), and there are some others, which include a huge set of parameters, objectives, and constraints, that renders them practically impossible to solve; i mean, if you approach them through Good Old Fashioned Artificial Intelligence paradigm (i.e. GOFAI, compare with "sub-symbolic" approaches, which includes evolutionary computation).
these latter types of subtasks are still being researched (as a few examples: distribution of the masses, building orientation, floor plan generation, envelope shape and aperture formation, or complete conceptual formal design in an integrated fashion). but this line of research is still rather underdeveloped, from a designer's point of view. nevertheless there's great hope, 'cause existing techniques seem to be offering several possibilities.
however there is also a third type of sub-tasks, which have to be tackled for a convincing design solution to appear. these include the conceptual (or say, mostly verbal) brief generation and problem setting stages, where the designers need extensive understanding of every aspect of culture as well as a complete ability to understand and manipulate shapes and objects (at least as well as human beings do). a creative and sensitive design proposal which is in dialogue with it's cultural, political, economic, and historical context, doesn't seem to be attainable with mere adaptation or pre-defined rules or constraints or as a result of a mostly random evolution.
are we obliged to wait for a machine, at least as clever as human mind (i.e. for strong-AI program to be accomplished) before entrusting design activity altogether to the machines?
Kaydol:
Kayıtlar (Atom)